A Comprehensive Study of Passive Wake-up Radio in Wireless

Delft University of TechnologyMasterrsquos Thesis in Electrical Engineering

A Comprehensive Study of Passive Wake-upRadio in Wireless Sensor Networks

Liang Huo

A Comprehensive Study of Passive Wake-up Radio

in Wireless Sensor Networks

Masterrsquos Thesis in Electrical Engineering

Embedded Software SectionFaculty of Electrical Engineering Mathematics and Computer Science

Delft University of TechnologyMekelweg 4 2628 CD Delft The Netherlands

Liang HuoLHuostudenttudelftnl

10072014

AuthorLiang Huo (LHuostudenttudelftnl)

TitleA Comprehensive Study of Passive Wake-up Radio in Wireless Sensor Networks

MSc presentation17072014

Graduation Committeeprof dr Koen Langendoen(chair) Delft University of Technologydr Przemyslaw Pawelczak Delft University of Technologydr Hong Li NXP Semicondutorsdr Wouter Serdijn Delft University of Technology

Abstract

Wake-up radio as a secondary radio transceiver is implemented to monitorthe channel condition so that the main radio can be turned off when there isno communication activity This thesis work focuses on the analysis of pass-ive wake-up radio (PWUR) where the wake-up radio is entirely powered bythe wake-up signal and does not need any additional battery supply Thework covers PWURrsquos both hardware implementation and performance ana-lysis In particular we analyze some typical issues when PUWR is adoptedinto the ZigBee networks Our results show that implementing PWUR inZigBee networks improves latency and energy consumption compared to tra-ditional duty cycling approaches except at the highest packet rates Alsothe addition of PWUR improves the ZigBee network throughput by intro-ducing the wake-up delay to the CSMACA process

Preface

This master thesis is a cooperation project between Embedded Softwaregroup of TU Delft and NXP Semiconductors In passive wake-up radio wire-less power is transferred in the format of wake-up signals from the wake-uptransmitter to the wake-up receiver Embedded Software group in TU Delfthas been working on the topic of wireless power transfer very successfully inrecent years Meanwhile NXP Semiconductors has their focus on analyzingthe overall ZigBee network performance and exploring new technologies andmethods to improve the network reliability This thesis topic is formulatedin order to combine both sidesrsquo best interests

I would like to thank my supervisors Dr Przemyslaw Pawelczak at TUDelft and Dr Hong Li at NXP Semiconductors It is their support thatguides me through my thesis work over the past 9 months I came to theNetherlands two years ago and never expected to have such an amazingjourney Thank all my friends in Holland who make this country my ldquohomeaway from homerdquo Thank my mom She is always there when I need herIt is her unconditional love and support that make me who I am today

Liang Huo

Delft The Netherlands10th July 2014

Contents

Preface v

1 Introduction 1

11 Overview of WSN 1

12 WSN Supporting standards 2

13 Wake-up radio 3

14 Problem statement 6

15 Structure of the report 7

2 Background 9

21 Hardware design 9

22 MAC layer performance 11

3 Theoretical Model 15

31 Wake-up process 15

32 Analytical model for duty cycling and PWUR 17

321 Duty Cycling 18

322 PWUR 23

33 Medium access control 25

331 TDMA 26

332 CSMACA 27

333 Combination 29

4 Simulations 33

41 Simulation setup 33

411 ZigBee model in OPNET 33

412 Wake-up radio implementation 36

413 Simulation assumptions 38

42 Results and analysis 39

421 Duty cycling VS PWUR 39

422 PWUR on CSMACA 44

5 Conclusions and Future Work 5751 Conclusions 5752 Future Work 58

Chapter 1

Introduction

11 Overview of WSN

The emerging field of wireless sensor network (WSN) has attracted signific-ant research interests in the recent years [10][11] WSN is considered to bethe most essential component of the Internet of Things (IoT) [27] where itis widely used in systems of smart parking smart grid smart home etc Inaddition WSN supports numerous traditional applications such as environ-ment and habitat monitoring health monitoring machine surveillance andtraffic control [12]

A wireless sensor network is built from a few to hundreds or even thou-sands of nodes Each node contains at least one sensor to monitor physicalor environmental conditions such as temperature sound pressure etc andto cooperatively pass their data through the network to a main locationApart from the sensor each node has typically several parts a radio trans-ceiver with an internal antenna or connection to an external antenna amicrocontroller (MCU) an electronic circuit for interfacing with the sensorsand an energy source usually a battery or an on-board form of energyharvesting Multiple nodes make up a sensor network following a specificnetwork topology such as star network or an advanced multi-hop wirelessmesh network[17][33]

Among all the issues that exist in WSN there are three of them standingout the most namely the energy consumption performance of the sensornode latency performance for data transmission and the overall reliabilityperformance of the system

Energy

First and foremost is the energy consumption issue The sensor nodes areusually powered by batteries thus have very limited lifetime if no powermanagement is performed Plus replacing sensors or batteries in manyapplications is not cost-efficient or sometimes even impossible It is crucial

to minimize energy consumption of wireless sensor network and extend itslifespan

Latency

Latency is a measure of time delay which is defined as the time it takes fromwhen a data packet is transmitted from the originating sensor to reachingits final destination It is usually application-dependent For example mes-sages to actuators should be communicated in real-time in our industrialautomation system Applications such as room temperature to detect a firealarm often has a higher requirement than other more delay-tolerant sys-tems In WSN information is exchanged between nodes in the format ofdata packet Thus packet delay is another important issue

Reliability

Last but not least since wireless sensor network is often deployed in a verydynamic environment Lots of factors can lead to performance varianceCollisions and data retransmissions may happen too often due to large num-ber of nodes in the network Parameters like received power or Signal NoiseRatio (SNR) vary from time to time due to the nature of wireless chan-nel To design a robust WSN system we need to provide a certain level ofreliability to handle all the uncertainties

12 WSN Supporting standards

ZigBee

There are quite a few competing WSN standards on the market ZigBee[1]Bluetooth LE[2] WirelessHART[3] ISA100[4] are some of the most popularones

Compared to WirelessHART and ISA100 ZigBee is much more widely ad-opted in industry Up until now ZigBee products have covered the domainsof building automation health care smart energy and so on About 400manufacturers have adopted the ZigBee specification and joined the mem-bership of ZigBee Alliance where NXP semiconductors is a member of theBoard of Directors[1] Compared to Bluetooth LE another popular stand-ard ZigBee provides better wireless meshing capability to cover a larger areawhile Blueetooth LE is more suitable for Personal Area Networks (PAN) orBody Area Networks (BAN) Our work applies to an indoor environmentoffice network with a size of 100 to 200 square meters So we choose ZigBeeas the WSN standard to analyze

ZigBee is a specification for the higher protocol layers upon the phys-ical (PHY) and medium-access control (MAC) layers defined in the IEEE

802154 specification ZigBee operates in the industrial scientific and med-ical (ISM) radio bands 868 MHz in Europe 915 MHz in the USA and 24GHz in most jurisdictions worldwide Data transmission rates vary from 20kilobitssecond in the 868 MHz frequency band to 250 kilobitssecond in the24 GHz frequency band There are two versions of the CSMACA mechan-isms at the MAC layer slotted (beacon-enabled) and unslotted CSMA (non-beacon enabled) The ZigBee protocol is based on the ad-hoc on-demanddistance vector (AODV) algorithm This means routing discovery andpeer-to-peer communication is possible through this routing protocol Meshnetworking topologies are supported All nodes share the same channel andfrequency hopping is not available There are two classes of network devicesin ZigBee standards such as Full-Function Devices (FFD) and Reduced-Function Devices (RFD) FFD can form networks of any desired type suchas mesh star and hybrid whereas RFD can only connect to a full functionnode [29]

RFID (radio frequency identification) is a means of storing and retrievingdata through electromagnetic transmission using a radio frequency (RF)-compatible integrated circuit Today it is applied widely in supply-chaintracking retail stock management parking access control library booktracking marathon races airline luggage tracking electronic security keystoll collection theft prevention and healthcare [25]

Efforts have been put into merging the RFID technology into WSN [25][35] The main idea is to combine the RFIDrsquos properties (identifying andpositioning) into WSN (sensing and multihop communications) WSN hasthe advantage of more capabilities and offers more potential At the sametime RFID products are now more readily available on the current marketDue to the fact that RFID often provides very limited functionalities itis often cheaper than a wireless sensor device Therefore combining thestrengths of WSN and RFID is both practical and promising

In the following chapters we will introduce how we integrate RFID tech-nology into our ZigBee system to realize our passive wake-up radio imple-mentation

13 Wake-up radio

Mentioned in earlier section energy and latency are two crucial factors inthe design of WSNs When a node is active and waiting to receive data itwastes energy on idle listening Since traffic loads are usually low in WSNssuch idle listening can waste enormous amounts of energy unless efficientcommunication mechanisms are employed

To extend the lifetime of a sensor node its radio component is turned offits microcontroller (MCU) is set into a sleep mode and a timer is used toturn the node active periodically This scheduled approach is called dutycycling Duty cycling has been widely used in WSN However it suffers anenergy-latency trade-off issue (Figure 11)

(a) shorter delay higher energy consumption

(b) longer delay lower energy consumption

Figure 11 Energy-latency trade-off

Figure 11 describes the process of a sensor node receiving a packet Thenode tries to manage the trade-off between energy consumption and datalatency by setting the duty cycle accordingly With a higher duty cyclelike Figure 11a the packet delay is shortened because even if packet wasgenerated when the receive node is asleep it does not take too long untilthe next active period However in this case more energy is consumedsince the node is turned to active more frequently which leads to more idle-listenings On the other hand with a lower duty cycle like Figure 11bthe node will consume less energy at the cost of higher latency for datadelivery since sleeping period in each cycle is now longer There is no wayto improve the energy and latency performance at the same time by theduty cycling approach The best we can do is to try to synchronize thewhole system so that receiver is exactly at its active period when the packetis generated at the transmitter side But there are two major disadvantagesof synchronization

First usually system events are not so periodical or predictable Thismakes the synchronization process more difficult Second the synchroniz-ation process itself introduces overhead and complexities to the MAC pro-tocol

Wake-up radio provides a solution to break the energy-latency tradeoffIt is an on-demand approach where the node is woken up by a radio sig-nal namely wake-up signal (WuS) A WuS triggers a node to wake up fromthe sleep mode and start receptiontransmission activities The wake-upsignal is sent or received by a secondary radio transceiver which consumesextremely low power The energy benefit of using radio wake-up in compar-ison with duty-cycling is that nodes do not waste energy on idle listeningof the main radio since they are only awakened when there is a request forcommunication The latency benefit is that since wake-up radio is a purelyon-demand approach the only delay is the one introduced by the wake-upprocess itself

Figure 12 summarized multiple ways to categorize different wake-up re-ceiver schemes [18] The power source of a wake-up radio has a big impact onthe energy efficiency of WSN Recent hardware developments have providedrealistic approaches to power the wake-up radio passively [31][15] Howeverthe topic of passive wake-up radio is still lacking thorough theoretical ana-lysis That is why in our work we follow the classification of what is thepower source of the Wake-up Receiver (WuRx)

Figure 12 Classification of wake-up receiver schemes [18]

Based on their energy sources wake-up receiver proposals can be classifiedas passive wake-up receivers in which the wake-up circuitry is triggered by anexternal energy source and as active wake-up receivers in which the internalbattery is used such that the wake-up receiver monitors for the possiblewake-up signal In our work we are more interested in the performanceof Passive Wake-up Radio (PWUR) which harvests energy from the WuStransmitted by the sender to power themselves To be more specific weachieve the wake-up radio using a passive RFID tag as the wake-up signalreceiver and an RFID reader as the wake-up radio transmitter More details

on the hardware implementation will be discussed later

14 Problem statement

So far we have discussed WSN (ZigBee in particular) and PWUR Our mainfocus is to investigate

How can PWUR improve the overall performance includingenergy latency and reliability in a ZigBee network

To answer this research question we first propose a PWUR-enabled Zig-Bee network architecture in Figure 13 To our knowledge this is the first-ever effort to combine PWUR with ZigBee network specifically

Figure 13 PWUR-enabled ZigBee network

Figure 13 shows a cluster-tree topology ZigBee network Due to thepower cost and size of a RFID reader it is not realistic to equip an RFIDreader onto every ZigBee end-device Thus ZigBee end-device cannot func-tion as a wake-up transmitter As an alternative we equip the RFID readeronly on the ZigBee routers In a start network it would be on the ZigBeecoordinator A ZigBee end-device is equipped with an RFID tag to harvestenergy from WuS as a WuRx Noticing that in this network architectureend-devices do not have the capability of waking up each other Any com-munication between two end-devices has to go through routerscoordinator

Having our network architecture defined we focus on two specific prob-lems to answer the research question

Problem (1) The routercoordinator needs to send a data packetto an end-device It first sends a WuS to wake up the end-device

After being woken up the end-device sends a RTR (Request ToReceive) message to require the reception of data packet from co-ordinator After receiving the RTR the coordinator sends the datato end-device What would the energy and latency performance becompared to the traditional duty cycling approach

Problem (2) The routercoordinator needs to receive a data packetfrom an end-device It sends a WuS to wake up the end-deviceAfter being woken up the end-device starts to transmit the datapacket to coordinator What would the overall MAC layer per-formance be when such wake-up mechanism is introduced

The wake-up signal itself does not tell if the end-device must receive orsend End-device sends the coordinator the data or RTR packet to distin-guish these two traffic types

These two approaches cover two sorts of WSN applications Scenario (1)is that the coordinator is going to distribute information to the end-deviceFor instance a central control in a smart home network informs the sensor onthe air-conditioner what is the correct temperature to be set at the momentScenario (2) is that the coordinator is going to collect information from theend-device Typical use case is the health monitoring where coordinatorreceives all the data from sensors on different parts of human body

We do have to point out that in the real WSN Problem (2) is the one weencountered more often because mostly sensors send their sensed informationto the central coordinator However for the integrity of our work we analyzeboth scenarios

15 Structure of the report

The structure of the report is as follows Chapter 2 gives a backgroundintroduction of PWUR study including how we refer to the previous worksand improve them Chapter 3 introduces the theoretical analysis on PWURcompared to duty cycling and PWURrsquos influence on CSMACA protocolIn Chapter 4 simulations of the two scenarios mentioned above includingthe simulation setup simulation results and analysis are given Finally inChapter 5 we conclude our work and give a suggestion of the future on thistopic

Chapter 2

Background

The related work of passive wake-up radio can be categorized into twogroups PWUR hardware design and its MAC layer performance analysisIn this chapter we will introduce the background including how we refer toand improve them in our work

21 Hardware design

While there has been a lot of research on active wake-up radio [23][28][19]there has not been many on PWUR Based on the wake-up capabilityPWUR can be categorized into range-based PUWR and ID-based PWURUsing a range-based wake-up radio all sensors within the range of the trans-mitted wakeup signal will be woken up while an ID-based wake-up radiotransmits a wake-up signal that contains the intended destinationrsquos addressand thus only wakes up the node with a matching address ID-based wake-upprovides the identifying capability therefore its hardware is more sophistic-ated A demodulation component should be added in order to obtain theWuSrsquos destination address

Range-based

One of the earliest and most cited work [20] laid the groundwork for PWURstudy The circuit of its passive WuRx is shown in Figure 21 Althoughsome more recent works have more sophisticated circuit design the basicconcept is still the same

In Figure 21 the antenna reacts to the EM wave of the WuS and gen-erates an input voltage The transformer increases the peak voltage of thesinusoidal signal curve to a certain level Then a diode is used as a rectifierto produce the output voltage It uses a capacitor to store energy

When the antenna receives the radio signal the flow of current accumu-lates energy on Cse and the voltage across the capacitor increases accord-ingly This process is described by equation Equation 2122

Figure 21 Passive WuRx circuit [23]

E= Pr times t times ef (21)

V out =

radic2 times E

Cse(22)

where E is the total amount of RF energy the circuit collected Pr is thereceive power of WuS t is the time required to accumulate energy ef is theefficiency of energy harvesting Cse is the capacitance of the capacitor Csewhich is responsible of storing the energy and generating the output voltageWhen sufficient energy has been stored on Cse Vout becomes higher thanthe threshold of interrupt voltage of sensorrsquos MCU A successful wake-up iscompleted [20] claims that its PWUR circuit could operate at the range of10 ft with 5 ms latency If a comparator and an amplifier are added whichrespectively consume negligible currents of 350 nA and 880 nA the radiocould theoretically reach up to 100 ft with 55 ms latency However there isno hardware implementation in [20]

Later [16] provides their PWUR circuit design According to [16] theyachieve a sensitivity of -293 dBm to obtain an output voltage of more than07 V at the frequency of 868 MHz Figure 22 is their wake-up circuitdiagram Unlike [20] their CMOS IC is fabricated Our later analysis isbased on this hardware implementation

ID-based

ID-based wake-up requires the identifying capability on top of basic range-based wake-up RFID technology is a natural candidate to meet the need Infact WISP (Wireless Identification and Sensing Platform) [31] developedby Intel Research is exactly such a hardware designed to explore sensor-enhanced RFID applications (Figure 23)

WISP is initially designed to conduct research which combines RFID andWSN However because it has the demodulation module shown in Fig-ure 24 it can be directly applied to realize ID-based PWUR with an externalsensor attached to WISP Related works are published in [12] Work of [15]

Figure 22 Block diagram of the wake-up circuit [16]

Figure 23 WISP 41 DL

has extended WISPrsquos wake-up range from 13 ft to 17 ft This is howeverstill noticeably shorter than the range-based wake up since the addition ofthe demodulation component lower the sensitivity level even more Detailedanalysis of wake-up range calculation will be provided in Section 31

22 MAC layer performance

Apart from the hardware implementation previous work also tried to answerhow the PWUR can improve WSNrsquos MAC layer performance and what newissues and requirements will be introduced when PWUR is applied

The authors of [32] [24] [34] [22] proposed their PWUR MAC protocolsSome of those also compared their MAC protocolsrsquo performance with otherexisting protocols The protocol in [32] called PRFW-scheme is extremelysimplistic and only applies to point-to-point communication system Shanet al gave the equations for energy and latency calculation [32] but nevercarried out any simulation To break the energy-latency trade-off authorof [24] proposed a radio wave sensor called RF watchdog that is dedicated

Figure 24 Block diagram of the WISP [31]

to sense the presence of an RF signal The RF watchdog is essentially awake-up radio The sensorrsquos RF module can maintain a sleep state until itsRF watchdog detects a communication signal Their PWUR-based MACprotocol is called ZeroMAC which can deliver data packets at least 187times faster by eliminating both idle listening and sleep delay while it con-sumes only 3 of the energy compared to X-MAC and A-MAC HoweverZeroMAC is designed based on IEEE 80211 DCF which is not compatiblewith ZigBee PWUR-based MAC protocol in [34] is called RTM (Radio-Triggered-Sensor MAC) Energy and latency performance of RTM are ana-lyzed in [34] However the MAC protocol ignores wake-up radiorsquos hardwarefeasibilities In particular it assumes the optimal wake-up capability forevery sensor node which is not realistic Khatibi et al proves in [22] that ifthe packet rate exceeds a critical threshold the wake-up radio actually con-sumes more power than the traditional ZigBee MAC protocol Therefore itproposed an adjustable MAC protocol that switches between the traditionalapproach and wake-up radio approach based on the packet rate Howeverthe performance of this MAC protocol is not given

Authors of [26] [36] focused more on the analytical model of PWUR MACBased on the analytical model they proposed in [26] Lont et al com-pared their WURx MAC protocol with two other low power MAC protocolsnamely the asynchronous X-MAC [13] and TDMA protocol While[36] ana-lyzed the energy efficiency of one-channel and two-channel wake-up radioZhang et al found out the applicable domain of the two wakeup schemesfor energy efficiency maximization with regards to the latency requirementunder different network settings

In terms of using RFID to realize PUWR [32] proposed their so-calledldquoPRFWrdquo (passive RF wake up) scheme and conducted a very simple energyconsumption analysis [21] proposed a MAC protocol called RFIDImpulseand compared its energy efficiency and transmission rate with the BMACprotocol and the IEEE 802154 standard

However with all the previous work that has been done there are stillsome limitations that have not been addressed

1) When designing MAC protocols hardware feasibility is often ignoredFactors like sensitivity of circuit components antenna gain influence ofwireless channel property etc should be considered into the MAC layerperformance analysis Although these are all PHY layer factors MAC layerperformance analysis should rely on valid PHY layer assumptions

2) Most previous RFID-based PWUR analyses assume that both a com-mercial RFID reader and a passive RFID tag are attached to each sensornode The analysis is based on an important assumption that all nodeshave the capability to wake up their neighbors This is unrealistic due tothe considerate amount of energy required by RFID reader and its large size

3) No previous work has conducted an in-depth PWUR energy and latencyperformance analysis in a ZigBee network ZigBee using the IEEE 802154standard to support its PHY and MAC has its own characteristics Forexample unslotted CSMACA is used in its MAC layer where an initialbackoff takes place before the first clear channel assessment (CCA) Thisand other factors should all be taken into consideration

In our work we have considered the limitations listed above Improvedupon the previous work we believe that we provide more all-around morein-depth and more realistic results and analysis

Chapter 3

Theoretical Model

In this chapter we discuss the theoretical models which are related to ouranalysis Firstly we introduce the wake-up process of the PWUR-enabledsensor node that we proposed Then in order to compare the energy andlatency performance of duty cycling and PWUR approach we describe ouranalytical models for these two approaches Lastly in order to make ourmodel more applicable to the real-world multiple end-nodes scenario twomedium access control methods that are widely used in WSN are intro-duced namely TDMA and CSMACA We also propose a possible solutionto combine TDMA and CSMACA in our analytical model

31 Wake-up process

We choose a 2-channel wake-up radio because it helps to avoid the collisionsof WuS and data packets in a one-channel wake-up scheme Also at lowerfrequency (868 MHz) path loss is weaker than that on 24 GHz which helpsto extent the wake-up radiorsquos operating range

In a 2-channel wake-up radio enabled sensor node wake-up radio andthe main radio are deployed at different frequencies Therefore two sets oftransceivers are needed The block diagram below shows the work flow ofthe wake-up process

As shown in Figure 31 wake-up radio is deployed on the 868 MHz bandwhile main radio operates on the frequency of 24 GHz [9] When a wake-up signal from Wake-up Transmitter (WuTx) is successfully received by theRFID WuRx WuRx collects the RF energy from the wake-up signal until itis able to generate a voltage Vout that is higher than the threshold voltageto produce an interrupt signal to MCU We have discussed the hardwareimplementation of this process in Chapter 2

Once the MCU is interrupted by WuRx it checks whether there is aneed to transmit data or an RTR (Request To Receive) packet If so itturns on the main radio module for corresponding activities The main

Figure 31 Block diagram of a WuRx

radio module and the MCU exchange the necessary data during transmissionandor reception

It can be seen that during the whole wake-up process the battery moduleis only needed to support the main radio module The wake-up moduleoperates in an entirely passive way Without any power supply from thesensor node itself it solely relies on the RF energy that WuRx collectedfrom the wake-up signal This is similar to the working mechanism of passiveRFID RFID tags are powered by RF signal from the RFID reader In factthe RFID technology has been considered as one of the best solutions toachieve PWUR due to its off-the-shelf availability

However as we have discussed in Section 21 while providing the ad-vantage of purely passive working manner PWUR suffers from a shorterworking range compared to active wake-up radio There are mainly 2 limit-ation factors

1) Regulation constraints on the maximum transmission power allowedAccording to UHF and EPC global standards [8] at the frequency range of865 sim 8676 MHz a maximum of 2 W ERP (38 W EIRP) of transmissionpower is allowed in European region (Table 31)

2) Hardware limitation which leads to the lower sensitivity of WuRxNormally sensitivity of the off-the-shelf main radio receiver eg a JN5148node is between -92 sim -95 dBm [5] However for PWUR circuit in [16]the sensitivity can only reach the level of -293 dBm If Pr lt -293dBmthe WuRx is not able to harvest RF energy from wake up signal since Pr ltPloss

Taken the 2 factors above into consideration the expected operating dis-tance for our range-based PWUR can be calculated with the following log-arithm form of the Friis equation of pathloss with a term for polarization

Table 31 Regulation on maximum transmission power [8]

loss included

Pr = Pt minus 20log(4πd

λ) +Gt+Gr minus Lp (31)

In Equation 31 EIRP of the transmitter is Pt + Gt = 20log38W =35798 dBm At the center frequency of 868 MHz λ = 0345m Thereceive antenna gain is Gr = 0 dBd = 215 dBi and the polarizationloss is Lp = 3 dB Lp occurs because only half of the power transmittedfrom the circularly polarized transmit antenna is received by the linearlypolarized receive dipole antenna The reason for using a circularly polarizedtransmitting antenna and a linearly polarized receive antenna is that theposition of wireless sensor is often quite unpredictable The circular-linearantennas pair guarantees that the receive power is the same regardless of theend-nodersquos position Using the operating threshold of -293 dBm from [16]Equation 31 predicts a maximum operational range of 35m For ID-basedwake up component (WISP) wake-up range is derived similarly [31] Butit goes without saying that ID-based wake-up has a even shorter wake-uprange

However let us not forget this operational range is derived with the freespace pathloss model When a more realistic pathloss model is appliedthe operational range of PWUR may also vary We will discuss this moreintensively in Chapter 4

32 Analytical model for duty cycling and PWUR

One of the main reasons that wake-up radio is proposed is that the tradi-tional duty cycling mechanism in WSN faces an energy-latency trade-off InSection 13 we have explained how the energy-latency trade-off happensIn order to prove that PWUR is able to break this trade-off and providea better performance on both ends we need to first propose our analyticalmodels for duty cycling and PWUR approach respectively

321 Duty Cycling

In this part we first introduce the analytical model of duty cycling whichis shown in Figure 32 The analytical model describes a communicationprocess between the ZigBee coordinator and ZigBee end-device

Figure 32 Analytical model of duty cycling

Latency

In Figure 32 x-axis describes the latency performance and y-axis representsthe power consumption of each phase The end-device undergoes the s2a(sleep to active) phase to shift from the sleep mode to active Once it turnsactive it sends out a CM (Control Message) to its coordinator The CMmay contain 2 sorts of information the updated real-time sensed informa-tion (temperature humidity etc) and the address of the sender end-deviceCM is sent periodically so that the coordinator can be aware of both thecondition and the presence of all the end-nodes in the network

Once the CM is successfully received coordinator sends an ACK (Ac-knowledgment) to end-device to inform the successful transmission Duringthis interval end-device has gone through the sft (shift) phase from Tx modeto Rx mode in order to receive the ACK In [9] the default setting of max-imum waiting time for ACK is 50 ms End-device remains in the listeningmode for 50 ms for the ACK message 50 ms is redundant enough to handlethe multiple packets transmissions in one cycle introduced by multiple nodesscenario in Section 33 or higher packet rate After the listening phase thenode goes back to the sleep mode

One cycle length of the end-device is composed of active period Tactive andsleeping period Tsleep In the process we discussed above the active period

is the sum of s2a CM sft and listen phase If the duty cycle is representedas η we have Equation 32 33 34 below

T = Tactive + Tsleep (32)

Tactive = Ts2a + TCM + Tsft + Tlisten (33)

T = Tactiveη (34)

We set the CM to be 32 bytes (256 bits) This is an example valueThe lenght of CM should be long enough to carry both the sensorrsquos real-time condition and address information But also the length should notbe comparable to the length of data packet in order to lower the channelload And from [6] we can also get the exact value for Ts2a TCM TsftTherefore Tactive is considered to be a fixed value The cycle length T onlydepends on the duty cycle value For example if the duty cycle is 10 from Equation 34 we can derive that T is 10 times of Tactive When η is1 T is 100 times of Tactive

The ldquoeventrdquoin Figure 32 represents that coordinator has generated apacket for the end-device However it is obvious that the packet will not bereceived if the end-device is in the sleeping mode As a result the packetwill only be sent if the CM from the destination end-device is received atthe coordinator side In Figure 32 an event occurred during the sleepingperiod of the second cycle The coordinator has to wait until it receivesthe CM from the third cycle to start the transmission The packet delay isthus the time between the real transmission and the event occurrence Inthis case ACK will not be transmitted anymore because the data packethas a higher priority Also if the end-device can receive the data packetsuccessfully then it also proves that its CM is received so that a separateACK message becomes redundant

As we investigate deeper into the model we find out that an event couldhappened in 3 different time domains (T1 T2 T3) which will results indifferent delay values as shown in Figure 33 Here we have

T1 = Ts2a + TCM + Tsft

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Suppose the event occurrence time is t When t ε T1 end-device has towait for the reception of CM to be completed So delay = T1 - t Whent ε T2 the end-device is already listening so transmission can start right

Figure 33 Packet delay calculation

away delay = 0 When t ε T3 end-device is in sleep mode The coordinatorhas to wait until the next CM from the same end-device so here we havedelay = T + T1 - t Because T3 which is the sleeping time constitutes thelargest part of a cycle unit we can expect that most packets arrive duringT3 hence the average packet delay will be close to T32 asymp T2

Energy Consumption

Unlike latency performance which is only related to the packet arrival timeEnergy consumption performance is also related to the packet arrival rateTo be more specific the relation between packet interval TI and cycle lengthT affects the energy consumption per packet Let n = TI T n gt 1 meansthat it takes more than one cycle to successfully receive a packet Forexample n =100 Between the two consecutive packet arrivals the end-device has gone through 100 cycles On the other hand n lt 1 means thatmore than one packet arrive within a single cycle This relation is illustratedin Figure 34

Figure 34 Idle cycle and rx cycle

In Figure 34 event occurred during the sleeping period of the secondcycle The first two cycles do not receive any data packets from the co-ordinator Idle listening happens in these two cycles which is why we callthem idle cycle However when the listening period starts at the third cycleend-device receives a data packet from coordinator We call this cycle anrx cycle The time to receive a data packet Tdata is different than Tlisten

so energy consumption of an idle cycle and rx cycle is different as shown inEquation 35 and 36

Eidle cycle = Es2a + ECM + Esft + Prx times Tlisten + Psleep times Tsleep (35)

Erx cycle = Es2a + ECM + Esft + Prx times Tdata + Psleep times Tsleep (36)

When n gt 1 it takes n cycles to receive a packet Only one of them isrx cycle So average energy consumption per packet can be calculated byE = (n-1)timesEidle cycle + Erx cycle When n lt 1 we have multiple eventsoccurred at one cycle The average number of packets received at each cycleis 1n So we can make sure that no idle cycle exists because the end-devicereceives at least one packet during every listening period Then the energyconsumption per packet is again related to which time domain (T1 T2 andT3 in Figure 33) the first packet falls into the cycle Figure 35) shows allthe 3 possible scenarios for that matter We made an assumption that therewill not be two consecutive packets arriving within T1 and T2 meaningthat except for the first packet all the following packets are all in T3 Wewill prove later that this is a valid assumption

(a) First packet arrives during T1

(b) First packet arrives during T2

(c) First packet arrives during T3

Figure 35 Energy consumption calculation

In Figure 35 if tp1 isin T1 it will be received in T2 of the current cycle Iftp1 isin T2 node has to keep listening in T2 before it catches the packet If tp1isin T3 packet can only be received in T2 of the next cycle Different amountof energy are thus consumed according to the arrival time of the first packetin the current cycle

Packet arrival process

For previous work analyzing duty cycling performance [12][32] packets ar-rives in a deterministic way The packet interval TI is a fixed value whichis decided by the data packet rate λ (pktmin) In our analysis we wantto simulate a more dynamic random traffic model In queueing theory aPoisson process is often used to model the packet arrival process[14] There-fore instead of uniform distribution in our analytical model packet arrivesfollowing the Poisson distribution Let k equal to the number of packetsthat arrive in one minute we have the probability mass function of eventsnumber below

Pr (X = k) =λkeminusλ

The mean packet rate is λ the mean value for packet interval time is thus1λ Probability density function f(xλ) of the packet interval time thereforefollows the exponential distribution as in Equation 38

f(xλ) = λeminusλx x isin [0infin) (38)

One important assumption we made is that whatever the packet rate isthere will not be 2 consecutive packets arrived at the T1 or T2 (Figure 32)of the same cycle Now let us verify whether this assumption is valid or not

We choose 4 typical duty cycle values in our analysis namely 01 051 10 But for each duty cycle value the length of T1 and T2 is fixedTactive = Ts2a + TCM + Tsft = 00028 s T2 = 005 s Suppose that a packetarrives at the beginning of T2 Here we choose the highest rate in ouranalysis λ= 60 pktmin = 1 pkts The possibility that interval between 2successive packet arrivals TI lt 005 is shown in Equation 39

Pr(x lt 005) = 1minus eminusλx = 1minus eminus005 = 0049 (39)

Equation 39 shows the probability for 2 successive packets in the sameT2 is quite low even for the highest packet rate We can imagine for a lowerpacket rate like λ= 3 pktmin = 005 pkts the probability is even lower(Pr = 25e-3)

Thus it is safe to make our assumption when the packet arrives followingthe Poisson distribution

Algorithm

Now we have introduced the latency energy and packet arrival process inour analytical model We now list all the parameter values that we use inour model (Table 32 33)

Symbol Explanation Typical Value

Ts2a startup delay from sleep mode to active mode 163 ms

TCM time consumed to transmit a CM packet (32 bytes) 1024 us

Tsft settling time due to mode shift between transmission mode and receiving mode

130 us

Tlisten time for listening to incoming packets 50 ms

Tdata time to txrx a packet (1296 bits) 5184 ms

Tactive active period of one cycle unit 52784 ms

duty cycle of the sensor node 01 05 1 10

120582

number of packets transmitted to per minutes (pktmin)

low rate 05 1 15 2 25 3 high rate 10 20 30 40 50 60

Table 32 Duty cycling parameters table I[6]

Our algorithm of calculating average latency and energy consumption perpacket of duty cycling is shown below Algorithm 1 It is implemented inMATLAB

322 PWUR

The analytical model for passive wake up radio is much simpler compared toduty cycling The reason is that the end-device can stay in the sleep modeunless it receives a wake up signal from the coordinator (WuTx) We nolonger need to care about duty cycles Packet rate does not affect latencybut does have an impact on the average energy consumption per packetThe PWUR analytical model is shown in Figure 36

Figure 36 only shows the end-device (WuRx) side of the system since itsenergy consumption requirements are more challenging compared to WuTxWhen end-device receives a wake-up signal it needs a certain time to ac-cumulate enough RF energy in order to generate an output voltage thatexceeds the threshold The WU part describes this phase And since thewake-up process is entirely passive the power consumed is equal to PsleepWe have discussed in Section 21 that the wake-up delay is related to wake-up signal strength capacitance of the storing capacitor and the thresholdvoltage

After the node is woken up it follows the same process as in the dutycycling model Equation 310 to 312 describe the latency and energy per-formance of PWUR

Parameters Power

power consumption in sleep mode 27 uw

average power consumption during sleep mode to standby mode

855 uw

power to transmit one CM packet 0dBm output power

339 mw

average power consumption during transition between transmission mode and receiving mode

246 mw

power consumption in receiving mode 1Mbps

354 mw

Table 33 Duty cycling parameters table II[6]

Tdelay = TWU + Ts2a + TCM + Tsft (310)

Etotal = Psleep times [Ttotal minusM times (Ts2a + Trx + TCM + Tsft)]

+M times (Es2a + Erx + ECM + Esft)(311)

Epacket =EtotalM

Taking the parameter values from Table 31 and 32 energy consumptionE = P times t for each phase Equation 310 shows that latency is independ-ent from packet rate The only exception is when the following packet hasalready arrived before the wake-up process of current packet is done How-ever this scenario requires an extremely high packet rate since wake-upprocess often takes very short time Since Ts2a TCM Tsft are all fixedvalues the wake-up delay is the only variable here It can be calculatedfrom Equation 21and 22

Equation 311 shows the total energy PWUR consumes during Ttotal thetotal length of time that the node works In our simulation it is set to be oneday (24 hours) M is the total number of packets the node received in oneday Erx is the energy consumed to receive a data packet We do not haveto worry about the lower duty cycling and longer sleeping period will bringus longer latency Also for the energy consumption since the node can stayasleep unless an event occurred it consumes Psleep for most of the time Aspacket rates get higher M in Equation 311 increases meaning we receivemore packets in one day and the sleeping time becomes shorter Average

Algorithm 1 Latency and energy consumption per packet calculation

energy consumption per packet in Equation 312 is thus slightly increasedThere are no idle cycles like in the duty cycle approach any more Thereforethe on-demand wake-up effectively prevent the idle listening problem

33 Medium access control

So far the models are not including the multi-nodes overhead In WSNmedium access control is a big issue since there are often tens or hundredsof end-devices in one network competing for the medium at the same timeThat is why in this section we add the multi-nodes scenario into our ana-lytical model MAC protocols are mostly based on TDMA and CSMACAIn this section we first introduce how they are used in our analysis Thenwe propose an approach to combine these two schemes

Figure 36 Analytical model of PWUR

331 TDMA

Given there are N nodes connected to the same central coordinator Asthey all need to communicate with the coordinator channel access controlis needed in the system

In duty cycling TDMA scheme assigns one time slot of the whole cycleunit to each end-node to handle the multiple access control The length ofone time slot should be at least longer than necessary time for one successfuldata packet transmission Otherwise before one node finishing its transmis-sion the next slot for another node has already started In our case theslot length is the length of the active period within a cycle

However TDMA has its own limitation Suppose we have duty cyclingvalue n = 10 for each end-device The maximum number of end-nodes Nthe system can support without collision is 110 = 10 (Equation 313)

N = 1n (313)

It is obvious that when n = 10 10 time of an end-nodersquos cycle isactive So during the 90 of sleeping time other nodes may turn into theiractive periods The time resource is effective spreaded to all the end-nodesThe details are shown in Figure 37

In Figure 37 following the TDMA scheme coordinator sends data toend-node that is active at the time When there are 10 end-nodes in a 10duty cycling system the coordinator reaches its full capacity If there aremore than 10 end-nodes the coordinator cannot find another slots for theextra nodes

From Table 32 we can derive that 10 duty cycling means end-nodeupdates its information around every 05s This is a very high rate in thereal WSN Most systems that follows a lower duty cycling 1 for example

Figure 37 TDMA in duty cycling

(cycle length is around 5s thus end-node updates around every 5s) is ableto handle more end-nodes (100 nodes for 1 duty cycling) according toEquation 313

So we come to the conclusion that if duty cycle is high and we have largenumber of end-nodes to the same coordinator TDMA cannot handle thesystem needs anymore We need to think of another way to increase thecapacity

332 CSMACA

There are few versions of CSMACA for different communication stand-ards In ZigBee we use the unslotted CSMACA defined in [9] Instead ofscheduling each nodersquos transmission asynchronously like in TDMA nodescan start the CSMACA at the same time and compete for the medium

When the CSMACA starts the node first does an initial random back-off with back-off length varying from 0 to 2BE minus 1 back-off units Thenit performs the clear channel assessment (CCA) to check if the channel isfree at the moment If so then the node successfully obtains the mediumand starts the transmission If not the node starts another back-off Untilthe number of back-offs exceeded the pre-defined limits the node starts theretransmission When maximum number of retransmission is reached andthe node still fails to obtain the medium and start the transmission the nodeclaims the failure of competing for the medium and discard the packet

In PWUR-enabled ZigBee network the CSMACA process is invoked bythe wake-up process So wake up delay is added before the CSMACA initialback-off

As shown in Figure 38 due to the difference of each nodersquos wake-up delaylength their CSMACA starting time differs differently In the later chapterwe will see that this variance can sometimes improve the overall network

Figure 38 Communication process in PWUR-enabled WSN

performance significantly since the absolute simultaneous transmissions inlarge network lead to massive amount of collisions

In IEEE 802154 one back-off unit is set to be 320 us CCA (clear chan-nel assessment) takes 128 us And the backoff length can only be an integermultiple (random number between 0 and 2BEminus1) of 320 us Suppose theinitial backoff exponent is 3 There are in total 8 backoff lengths availableWhen more than 8 nodes start CSMACA simultaneously at least one col-lision is bound to happen Now we shift one nodersquos starting time by 1 ms(around 3 slots) These 8 nodes still generate random backoff lengths butthe probability of collisions will be reduced because the shifted node joinsthe CSMACA at a different time when a few nodes have already completedthe contention If the shifted node takes a random backoff of more than 5units collisions are not bound to happen any more In a large network withup to 100 nodes by shifting the CSMACA starting times of multiple nodesdifferently such collisions can be avoided to a great extent

If in the end we have Twakeminusup + Tcsmaprime lt Tcsma (where Twakeminusup isthe wake-up delay Tcsmaprime is the new CSMACA duration and Tcsma isthe original CSMACA duration with simultaneous transmissions) or thenumber of successful delivered packets increases given the same maximumretransmissions allowed we can draw the conclusion that PWUR helps toimprove the network performance

Compared to TDMA performing CSMACA does not need the coordin-ator reserve a time slot for a specific end-device Especially when theslot length can be much longer than the actual data transmission timeCSMACA provides a more dynamic efficient way to make use of the chan-nel However when the number of nodes increases especially when too manynodes start CSMACA process simultaneously to compete for the mediumCSMACArsquos performance drops as collisions are not perfectly avoided as inTDMA

We have discussed how TDMA alone (Figure 37) cannot handle the mul-tiple nodes scenario If we use CSMACA alone in order to reduce collisions

we can either increase the back-off exponent or maximum allowed back-offtimes But these two parameters are defined in IEEE 802154 with an upperbound of 8 and 5 respectively [9] We can allow a large maximum numberof retransmissions to maintain the packet delivery ratio For example wesimulate CSMACA process of 10 and 100 nodes If we want to guaranteeless than 10 packet loss the maximum number of retransmissions allowedand corresponding CSMACA duration are shown in Table 34

number of nodes retransmissions allowed CSMACA duration

10 4 006 s

100 75 065 s

Table 34 CSMA in multiple nodes scenario

When N = 100 in order to maintain less than 10 packet loss we haveto permit each node to have a maximum of 75 retransmissions The latencydoes not drop too much since CSMACA duration almost increases linearlywith the number of nodes But allowing 100 nodes to join the CSMACAprocess consumes much more energy because the node needs to spend energyto monitor the channel condition for as much as 065s when the actualtransmission time is only few milliseconds Apparently CSMACA alone isalso not good enough to handle the multiple nodes scenario effectively

333 Combination

Now we have explained neither TDMA nor CSMA can handle the multiplenodes scenario by itself We now propose a combination scheme to handlethe medium access control effectively

Overhead of TDMA is introduced by the design of time-slot length Over-head of CSMACA is introduced by CSMACA back-off and collisions Wehave pointed out that in the duty cycling TDMA is sometimes not enoughto handle large number of end-nodes in the system while CSMArsquos perform-ance drops if too many nodes start channel contention simultaneously Wenow combine TDMA with CSMACA to solve this issue The basic idea isto first use TDMA to divide a large network into smaller clusters Withineach cluster we apply CSMACA to solve the multiple end-nodes problem

For example we have discussed that as Equation 313 shows 10 dutycycling system can only handle 10 nodes because there are only 10 slots eachcycle in total to be assigned However by applying our approach each slotis no longer assigned to a node but a cluster with multiple nodes We nowhave 110 = 10 clusters When N = 30 number of nodes in one cluster Nrsquois N10 = 3 When N = 50 Nrsquo = 5 etc Nodes within the same cluster sharethe same timeslot Coordinator only deals with the cluster which is activeat the time Within one cluster nodes send out CM following CSMACAFigure 39 shows the process

Figure 39 CSMACA and TDMA Combination

End-node 1 sim 5 share the same timeslot In the first CM phase theycontent for sending out its CM to the coordinator Coordinator receivesthe CM knowing both the condition and presence of the sender node Thecoordinator also checks whether it has a packet for the sender node Thenafter receiving the Control Message from every node of the cluster in theListen phase coordinator starts to send data packets (if any) to the CMsender node in the order of how it receives the CMs After the data packetshave been successfully sent to the end-node this cluster of end-nodes goesto its sleeping phase And coordinator also moves on to deal with the nextcluster

The approach in Figure 39 will bring several changes in our existinganalytical model

1) T1 in Figure 33 becomes longer due to CSMACA process of CMpackets Instead of just sending out one CM like in Figure 33 the CM periodcontinues until the last CM within a cluster has been successfully receivedby the coordinator As the number of nodes in each cluster increases thelength of CM period also increases Table 41 shows the empirical values ofCM period length corresponding to different cluster sizes

2) Buffering time in T2 needs to be considered if the coordinator haspackets for multiple nodes within the same cluster After the CM periodthe coordinator checks if it has packets to nodes in this cluster If thecoordinator has packets for more than one node in the cluster it sendsthem at the order of how CM is received When the data packet is still inthe buffer its destination node needs to remain in the listening mode Nowif a node receives the packet on T2 its energy consumption is

ET2 = Erx + Prx timesNbuffer times Tdata

where Nbuffer is the number of packets in the coordinatorrsquos buffer It is a

random number related to the packet generation process of the coordinatorIts expected mean value is Tlisten(2times Tdata)

3) A longer CM period leads to higher energy consumption and latencyvalues A node has to spend more energy on monitoring the channel condi-tion (to see if it is free) before sending its CM packet Energy spent on T1is as follows

ET1 = Es2a + ECM + Esft + Prx times (T1minus Ts2a minus TCM minus Tsft)

4) A longer listening time may be needed according to the number of nodeswithin one cluster If the listening time is too short before the coordinatorfinishes sending out all the packets in the buffer the end-node will go tosleep

Chapter 4

Simulations

In this chapter we first introduce how our simulations are set up Then theresults and analysis on passive wake up radio performance are shown

We want to implement the architecture that is shown in Figure 31 ina simulator From our understanding there is no previous work that hasdone this before What we need in our design is a WuTx and a WuRx modelwithin each contains the main radio module and the wake-up radio moduleWake-up radio module is able to control the communication activities of themain radio module based on the wake-up signal

We implement our model in OPNET [7] since it already has a stableZigBee model This is a big advantage compared to some other popularsimulators like ns-3 Also its 3-tier (network model node model processmodel) hierarchical designing framework makes it easy and straightforwardto implement our new architecture

41 Simulation setup

411 ZigBee model in OPNET

OPNET has its default ZigBee model To build a ZigBee network in OPNETWe can first select the network topology to be either star mesh or cluster-tree topology Three kinds of ZigBee devices are defined to support differentnetwork topologies namely ZigBee coordinator ZigBee router and ZigBeeend-device Figure 41 shows an example cluster-tree ZigBee network whichcontains one coordinator 5 routers and 28 end-devices Coordinators con-trol the formation and security of networks Routers extend the range ofnetworks End devices perform specific sensing or control functions In asmart home network the coordinator may be a home theater control sys-tem with advanced support for lighting and security Devices such as lightfixtures thermostats and air conditioners could be configured as routingdevices Simple devices such as light switches and security sensors could beend devices

Figure 41 Cluster-tree ZigBee network

After defining the network model we need to define the node model asour next step Figure 42 illustrates the ZigBee node model in OPNET

Figure 42 ZigBee node model

As shown in Figure 42 each node is made up of different modules whichare responsible for the tasks of a specific layer as in OSI model Each moduleworks based on its process model which determines the modulersquos tasks andcharacteristics We see the wireless transmitter and receiver which make thephysical layer of the ZigBee node model Three other modules represent theMAC network and application layers of the node The physical layer con-tains the transmission and reception components which are responsible fordata exchange The MAC module provides the functional and proceduralmeans to transfer data and to detect and correct errors occurred in the phys-

ical layer The network layer is responsible for transferring variable lengthdata and for routing packets delivery including routing through interme-diate routers Finally the application layer is responsible for identifyingcommunication partners determining resource availability and synchroniz-ing communication In our work we mostly focus on the MAC layer moduleIn fact in the next section you will see that we modified the default IEEE802154 MAC process model shown in Figure 43

Figure 43 ZigBee MAC process model

Process model contains states transitions and conditions A green (forced)state is a state that the process goes to another state when the functions ofthis state are done (eg int state in Figure 43) In contrast the processremains in the red (unforced) state after finishing their tasks until an in-struction comes or an interrupt occurs indicating that the process should goto another state (eg idle state in Figure 43) The transitions that connectthe states define the conditions for going from one state to another Theseconditions and the tasks that should be done when the conditions are metare defined in the process modelrsquos function block

The MAC of the ZigBee node works based on CSMACA (Carrier SenseMultiple Access with Collision Avoidance) ZigBee MAC module is madeof a main process model which invokes the CSMACA as a child processwhenever it intends to transmit data CSMACA protocol in OPNET isdefined by five states (Figure 44) initialization idle state channel checkingback-off state and transmission

So now we have introduced the ZigBee model in OPNET Figure 45 isthe architectural view of the model structure

Figure 44 CSMACA process model

Figure 45 ZigBee OPNET model in a nutshell

412 Wake-up radio implementation

Our wake-up radio model was built on top of the models we discussed inSection 411 The difference is our new PWUR-enabled node model con-tains a wake-up module along with the main radio module (Figure 46)The wake-up module is able to control the communication of the main radiomodule

The wake-up module and main radio module operate on seperate channelsA wake-up process starts from WuS (wake-up signal) generator at WuTx(coordinator) side generating a WuS If the signal strength is higher thanWuRxrsquos sensitivity WuRx(end-node) is able to harvest energy from WuSWake-up delay is the time consumed until WuRx harvest enough energy Itis calculated by

WU delay = E(Pr minus Ploss) (41)

where E is the energy needed to generate the interrupt voltage (Equa-tion 2122)

The WuS processor calculates the WU delay and returns this value tothe MAC module of main radio The main radio does not send its bufferedpackets until it receives the WU delay input from wake-up radio modulesuggesting a successful wake-up

Once a node is woken up it transmits all the packets in the current bufferWe modified the 802154 CSMACA module so that the first transmitted

Figure 46 Wake-up radio node model (leftWuTx rightWuRx)

packetrsquos fist backoff length equals to the sum of wake-up delay and the ori-ginal random CSMACA backoff For a retransmission attempt CSMACAstill follows its normal back-off mechanism since the wake-up delay has beentaken into consideration before its first transmission attempt Also for allthe other buffered packets other than the first transmitted one wake-up pro-cess does not affect their CSMACA backoff They can start immediatelyafter the previous packet in the buffer is transmitted

On the other hand if Pr of WuS is lower than Ploss meaning that WuRxis not able to harvest RF energy thus it is a failed wake-up Main radiorsquosMAC module buffer the packets from its application network module untilthere is a successful wake-up

So we made two major changes to the original OPNET ZigBee model

1) Originally when ZigBeersquos MAC module receives a packet from its up-per layer (network layer) it only checks the availability of MAC and thenimmediately invokes the CSMACA process to transmit the data packet onPHY Now invoking CSMACA is also controlled by WuS processor

2) The initial CSMACA backoff of the first packet transmission takesplace after WU delay

The introduction of PWUR affect the sensor nodersquos communication pro-cess which is implemented as the Algorithm 2

Initializationfor i = 1 -gt N do

if i has bufferd data thenif i is woken up by WuTx then

after WU delay i start its CSMACA processtransmit all the packets in the buffer

elsebuffer the data

elsegenerate new data packetwait for new wake-up

endAlgorithm 2 WuRxrsquos communication process

413 Simulation assumptions

To simplify our simulations we made the following assumptions

bullWe only consider the energy consumption of a sensor nodersquos radio com-ponent Energy costs for sensing activities are ignored as they will notimpact the performance evaluation

bull Propagation delay is ignored

bull The link is ideal meaning that a packet is correctly received if the receivepower is higher than the nodersquos sensitivity level The effect of interferenceis not considered

bull The length of one data packet is 128 bytes which is the maximumpacket size defined in ZigBee As we are not focused on any specific WSNapplications the maximum packet size is set to leave a decent redundancyTogether with all the overhead the total packet length is 162 bytes

bull Maximum allowed number of retransmissions is set to 5 If after 5 trialsthe packet is still not successfully transmitted we discard the packet

bull The length of a wake-up signal is long enough to charge the RFID tagon WuRx Also receiving power stays constant within the same signal

bull Wake-signalrsquos transmission power takes the maximum value allowed byTable 31 which is 38 W EIRP at the frequency of 868 MHz

As for other more specific simulation assumptions and requirements wewill introduce them in the correspond section if necessary

42 Results and analysis

421 Duty cycling VS PWUR

This section answers Problem 1 in Section 14 end-nodes receive packetsafter being woken up what would be the energy and latency performancecompared to the traditional duty cycling approach

In Chapter 3 we have discussed about our analytical model of duty cyc-ling and PWUR in WSN Duty cycling and PWUR represent two differentapproaches to improve the overall performance of WSN pre-scheduled andon-demand approach We compare them in terms of both latency and energyconsumption Finally to make our analysis more applicable to the multipleusers scenario we analyze the performance of the TDMA amp CSMACAcombination approach that we proposed in Section 333 In this set of sim-ulations focused on latency performance we first generate a series of packetsfollowing the Poisson distribution Based on the average packet rates sim-ulations are divided into the high-rate (10 sim 60 pktmin) and low-rate (05sim 3 pktmin) group in order to cover a wide range of applications

4211 Latency

In duty cycling the average packet delay is related to the packet arrival timeIn our simulation packet arrives following Poisson distribution When thepacket arrival time is generated latency is calculated as in Figure 33

Given a specific duty cycling value which decides the cycle length accord-ing to Equation 34 and a specific packet rate there can be 0 1 or morethan one packet arrives during one cycle length To calculate packet delaywe can neglect the scenario of 0 packet in one cycle And when there aremore than one packet in one cycle the average packet delay is the meanvalue for all packets which arrives to the specific cycle

To prove our theoretical analysis is valid we hereby list the delay distri-bution in Figure 47 for the case when there is only one packet in the cycleCycle length is 52784s 10557s 5278s and 0525s for 01 05 1 and10 duty cycling respectively (Equation 34) We take the 10 duty cyc-ling scheme as an example where from Equation 32 to 34 and Table 32we have

T1 = 00028s T2 = 005s T3 = 04751s

We run our simulations 10000 times and plot the CDF of delay distribu-tion of this 10000 iterations When packets arrive in T1 T2 or T3 packetdelay turns out to be within different ranges Figure 47 proves that theprobability distribution of packet delay is related to the proportion of T1T2 and T3 compared to the entire cycle length For example for the 10 duty cycling scheme probability of packet delay is 0 is P(d=0) = T2T =00947 Our simulation result 937 complies with the result

0 01 02 03 04 050

Packet delay (s)

CDF of delay distribution

Figure 47 Packet delay probability distribution

As for the latency of PWUR TWU in Equation 310 is calculated basedon the parameters of the circuit in [16] given the wake-up distance is 5 mUsing the parameters listed in Table 31 we obtained the average packetdelay of both duty cycling and PWUR in Figure 48

05 1 15 2 25 3

Packet rate (pktmin)

Latency performance low rate

10 20 30 40 50 60

30Latency performance high packet rate

01 duty cycling05 duty cycling1 duty cycling10 duty cylingPWUR

Figure 48 Latency performance comparison

Figure 48 shows the latency performance comparison of duty cycling andPWUR It is obvious that the lower the duty cycling is the longer theaverage packet delay will be This is due to the longer sleeping period in eachcycle However the latency performance of PWUR surpasses the one withhighest duty cycling (10) It proves that the on-demand PWUR approachis indeed superior to handle the delay-sensitive real-time applications

We notice that for 01 duty cycling at the packet rate of 1 pktminthe average packet delay drops slightly For 01 duty cycling the cyclelength is 52784s The average packet interval 60s is quite close to 52784sIn our simulations there are some sequences of packets which arrive at the

beginning of the cycle (T1 or T2) or at the end of the cycle (end of T3)Packet delay for these packets is quite short When cycle length and packetinterval are close this kind of packets sequences appears more frequently Sothe system is coincidentally synchronized with the packet arrival time Overa large number of iterations a slightly shorter packet delay is expected

4212 Energy Consumption

In terms of energy consumption PWUR is supposed to have a better per-formance since it avoids idle listening in duty cycling which consumes con-siderate amount of energy However this is not always the case When thepacket rate becomes higher too frequent wake-ups may lead to even higherenergy consumption Our simulation result in Figure 49 also proves thepoint

05 1 15 2 25 310

minus1

Energy performance low packet rate

10 20 30 40 50 6010

minus1

Energy performance high packet rate

01 duty cycling05 duty cycling1 duty cycling10 duty cyclingPWUR

Figure 49 Energy consumption performance comparison

In the low packet rate group as we expected higher duty cycling willlead to higher energy consumption PWUR scheme consumes less energyper packet since it only consumes extremely low energy in the sleep modewhen there is no packet arriving However as shown in Figure 49 whenpacket rate exceeds 15 packets per minute while still maintaining to bemore energy efficient than high duty cycling the advantage of PWUR overlow duty cycling scheme starts to vanish Equation 311 and 312 explainthat the energy consumption per pkt of PWUR varies very slightly withthe packet rate However when packet rate gets higher the possibility thatno packet arrives during one cycle decreases thus number of idle listeningsdecreases which eventually leads to lower average energy consumption perpkt As the packet rate continues to increase it could be even more energy-efficient to use low duty cycling scheme rather than PWUR Because everysingle packet received by PWUR requires a separate wake-up process whichconsumes energy However multiple packets arrived during one cycle timeonly need a single ldquoactiverdquo period in duty cycling

The energy consumption per packet does not vary too much among 01

05 and 1 duty cycling schemes This is because their duty length T(52784s 10557s 5278s) is comparable or significantly longer than averagepacket interval 1λ (1 to 6s at high rate) Idle-listening is unlikely to happenThat is why energy consumption performance for these three duty cyclingschemes is quite desirable which even surpasses PWUR slightly

4213 Multiple nodes scenario

The latency and energy consumption performance analysis above are allbased on a per-link manner to investigate the difference between duty cyc-ling and PWUR protocols themselves However this is not sufficient if wewant to apply our results for analyzing a large WSN where the central co-ordinator has to communicate with multiple end-nodes In Section 333we proposed a TDMA amp CSMACA combination approach on top of ouranalytical model to handle the multiple nodes scenario In this section weanalyze its performance

The packet rates range from 10 pktmin to 60 pktmin Duty cyclingis 10 Therefore end-nodes are divided into 10 clusters We simulatethe scenarios of 10 30 50 100 150 200 250 end-nodes in the networkrespectively So each cluster has 1 3 5 10 15 20 25 nodes Table 41illustrates the time consumed by CSMACA process within a cluster Werun our test program in OPNET and generate the average time (100 itera-tions) when the last node in the cluster has successfully transmitted its CMto the central coordinator Results show that as the number of nodes withineach cluster increases CM period in Figure 39 becomes longer since theCSMACA process for the control message takes longer time Notify herethat the channel load for transmitting CM remains to be around 10 to15 If its channel load is higher meaning that we have more nodes withinone cluster CSMACA can not function properly due to massive amount ofcollisions at the higher channel load

Figure 410 shows the latency performance of multiple nodes scenariosTDMA first divides all the nodes into 10 clusters We assume this processdoes not contribute to the whole delay Then the packet delay within acluster increases following the number of nodes The average packet delaywhen network size is 250 nodes is about 6 times than there are only 10 nodesin the system

Figure 411 shows the energy consumption performance of multiple nodesscenario In fact unlike the latency performance the number of nodes doesnot affect too much on the average energy consumption per packet Thereason for that is in WSN extra energy consumption is mainly introducedby idle listening When number of nodes increase it does not bring extraidle-listenings to each node The node only needs to spend extra energyon monitoring the channel in the longer CM periods for the purpose ofCSMACA This part introduces the slight increment on energy consump-

Number of nodes within each cluster

Time consumed inCSMACA process (ms)

10 888

15 1387

20 162

25 1976

Table 41 CM period length

10 20 30 40 50 600

Latency performance for multiple nodes scenario

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

Figure 410 Latency performance of multiple nodes scenario

tion which is why in a network with 250 nodes energy consumption perpacket is slightly higher than the one with 10 nodes For energy consump-tion packet rate is the more decisive factor since it is directly related to thenumber of idle-listening periods a node has to endure in order to receive onepacket

In summary we introduce a combination approach of TDMA and CSMAon top of our analytical model which brings different influence on the latencyand energy performance To be more specific as the number of nodes in-creases latency performance starts to drop meanwhile the energy perform-ance only suffers very limited impact

10 20 30 40 50 600

20Energy performance for multiple nodes scenario

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

Figure 411 Energy consumption performance of multiple nodes scenario

422 PWUR on CSMACA

This section answers Problem 2 in Section 14 end-nodes start to send pack-ets after being woken up What would the overall MAC layer performancebe when such wake-up mechanism is introduced

We will be discussing 3 approaches on the MAC layer TDMA (to senddata asynchronously) CSMACA (to send data synchronously) and PWURMAC (to send data semi-synchronously) Figure 412 illustrates the differ-ence among these three which is mainly the starting time of packet trans-mission The irregular forms of back-off in the PWUR MAC is introducedby the wake-up delay which is related to the received power of the WuS(Wake-up Signal)

In the simulation the association process between coordinator to end-nodes is done from 0s to 30s And the packet transmission starts from 30suntil 5 minutes In this section the packet rate is 1 pkts So for each runof simulation 270 packets transmissions per node take place

CSMACA

When we introduced in Section 332 about CSMACA process in IEEE802154 we mentioned that if multiple nodes start CSMACA process sim-ultaneously system performance drops as the number of nodes increasesThis is a more typical issue in a range-based wake-up ZigBee network Co-ordinator sends wake up signal (WuS) which could wake up the end-nodesin a certain range After being woken up end-nodes transmit packets to thecoordinator following the CSMACA before the actual transmission of datapacket Given the path loss of WuS follows the Friis model nodes withinthe same range may have very slight difference in terms of their received

Node 1

Node 4

Node 3

Node 2

T0(wake-up) T0(wake-up)T0(wake-up)

CSMACA TDMA PWUR MAC

Figure 412 Three MAC approaches

WuS signal strength Therefore their wake-up delay is relatively the samethey will start the CSMACA simultaneously This scenario will bring ustwo consequences longer delay due to more back-offs and higher collisionrate Figure 413 and 414 illustrates delay and collision rate respectivelygiven the number of nodes woken up simultaneously are from 10 to 100 Wesimulate the process for 5 minutes and the packet rate is 1 pkts

0 50 100 150 200 250 300 3500

Simulation Time (seconds)

ZigBee Application Endminustominusend Delay (seconds)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

[smoothed] ZigBee Application Endminustominusend Delay (seconds)

N = 10N = 30N = 50N = 100

Figure 413 Data packet delay of CSMACA (left real-time data rightsmoothed data)

Figure 413 shows the growth of packet end-to-end delay as number ofnodes increases As more nodes compete for the medium more collisionstake place The number of back-offs and retransmissions increases Whenretransmission attempts reach the limit the node has to discard the packetFigure 414 shows that packet delivery percentage drops as more packets

0 50 100 150 200 250 300 3500

Packet Delivery Percentage

N = 10N = 30N = 50N = 100

Figure 414 Successful packet delivery percentage of CSMACA

are discarded due to collisions When there are only 10 end-nodes in thenetwork almost 100 of packets are successfully delivered However as N is100 the percentage drops to under 20 meaning that during our 5-minutesimulation period only less than 20 of packets are received successfully bythe coordinator

If we multiply the number of users by the packet delivery percentagewe find out the outcome is always around 20 This suggests that the fullchannel load that CSMACA can handle is 20 nodes sending packets at thesame time (In Figure 414 when N = 10 lt 20 delivery percentage is almostalways 100) If the node number exceeds 20 packets are dropped becausethe channel has been over-loaded

Instead of starting the CSMACA process immediately after the nodes arewoken up we can also use TDMA to coordinate the transmissions of allthe nodes in the system While CSMACA is driven by the WuS of theRFID reader (WuTx) TDMA is a tag-driven approach where each WuRxhas a pre-set sequence number m which ranges from 1 to N (total numberof nodes) Suppose one time slot length in TDMA is t upon woken up eachnode backs off until the time mt To avoid any collision the value of t shouldbe longer than the maximum possible value of the whole transmission timeof one data packet so that slots do not overlap each other In our simulationwe have

t = Tprocesing + Tmac access + T(txamprx Data) + T(txamprx ACK) ge 00075s (42)

Figure 415 Data packet delay (TDMA)

The transmission of each node starts in an asynchronous mode Its packetdelivery percentage is 1 since collisions are all avoided Figure 415 showsthe packet delay of this asynchronous approach The value of packet delaydoes not vary too much from time due to the absence of random back offdelay in CSMACA

PWUR on CSMA

The previous analysis in the CSMACA part is based on the assumptionthat nodes wake up simultaneously However this is not always the case Iftwo WuRxrsquos distances to WuTx is different such that it takes them differenttime to harvest enough energy from the WuS This wake-up delay is appliedbefore the CSMACA process starts So we now have a semi-synchronousscheme for data transmission

We place all the nodes randomly within the coordinatorrsquos wake-up rangeEach node has its own distance to the coordinator Their individual wake-updelay can be calculated by Equation 412 For example Cse in Figure 21 is02uF interrupt voltage is 07V for a node which is 10 meters away from thecoordinator its wake-up delay is 46ms For a node with 5 meters distanceto WuTx its wake-up delay is 11ms Thus the CSMACA starting timefor this two nodes has a 35ms shift

Packet delay and successfully delivery rate are shown in Figure 416

4221 Effective Packet Rate (EPR)

We have analyzed both packet delay and delivery percentage for all threeMAC layer approaches CSMACA (synchronous) TDMA (asynchronous)and PWUR on CSMACA (semi-synchronous) If we only look at the delay

0 50 100 150 200 250 300 3500

[Smoothed] ZigBee Application Endminustominusend Delay (seconds)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

Figure 416 Performance of PWUR (left Smoothed data packet delayright packet delivery percentage)

performance we would find out that TDMArsquos performance is worse thanthe other CSMACA-based approaches Especially when there are a largenumber of nodes (N = 100) its end-to-end packet delay is more than twice ofthe other two Between CSMACA and PWUR MAC there is no significantdifference (Table 42 )

NodeNumber CSMACA TDMA PWUR MAC

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

Table 42 Average packet delay (s)

However latency alone does not represent the network overall reliabilityIn fact the parameter of packet delay can only measure those successfuldelivered packets It cannot reflect all the discarded packets Therefore tocompare the overall performance we need a new parameter that describesthe network better

Effective Packet Rate (EPR) of a network is the number of successfully de-livered packets per second by all the nodes It is calculated in Equation 43

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

Figure 417 EPR comparison (real time value)

EPR = ptimes 1

D+ (1minus p)times 1

infin=

p is the packet delivery percentage therefore 1 - p is the probability that

a packet is discarded D is the packet end-to-end delay and1

Depresents

the packet rate of the successful transmissions (in pkts) When a packetis not successfully delivered its end-to-end delay is infin therefore the packet

rate is 0 The final result EPR =p

Dis the mean value of packet rate It

shows that the Effective Packet Rate is dependent on both packet delay anddelivery percentage

For example in a network successfully transmitting one packet costs 10ms on average And only 70 transmissions are successful By Equation 43we have EPR equals to 70 pkts 70 packets are successfully transmittedin 1 second The concept of EPR is close to network throughput (in bits)However using EPR with the unit of pkts is more straightforward to com-prehend the packet transmission

Figure 417 shows the EPR comparison of three MAC layer approachescorresponds to different nodes densities It shows the general trend of EPRvalue EPR of CSMACA and PWUR MAC have a higher variance thanTDMA The average EPR are shown in Figure 418

We can draw the following observations from Figure 418

N=10 N=30 N=50 N=1000

30EPR of 3 MAC Protocols

Node density

CSMACA

PWUR MAC

Figure 418 EPR comparison (average value)

1) Both TDMA and PWUR MAC outperform CSMACA of 802154 inEPR

2) The maximum value of EPR is achieved when transmissions are con-ducted continuously and seamlessly Given the minimum duration of a suc-cessful transmission is 526 ms the optimal EPR when N = 10 is 346 pktsThe EPR of TDMA and PWUR MAC at N = 10 are both 251 pkts whileEPR of CSMA is 202 pkts The margin to the optimal EPR in TDMA isintroduced by the length of time slot We set it to be the maximum possibletime for one data transmission which brings overhead The margin to theoptimal EPR in PWUR MAC is introduced by the back-offs and retrans-missions in CSMACA Although it has been improved massively comparedto the complete synchronous CSMACA overhead still exists

3) As node number increases PWUR MAC starts to have higher EPRthan asynchronous mode Because it does not wait for theoretical max-imum interval between 2 transmissions At N = 30 and N = 50 the EPRof PWUR MAC is 77 and 128 higher than CSMACA 21 and 22higher than TDMA However when node density becomes too high PWURMAC cannot handle the collisions effectively anymore Transmitting 100packets takes 05376s The packet rate is 1 pkts Therefore the channelload has reached 5376 Our previous study shows that when the channelload exceeds 40 CSMACA can not function properly While the packetdelivery percentage of TDMA is always guaranteed to be 100 That is whyat N = 100 TDMA has a slightly higher EPR than PWUR MAC AlthoughPWUR MACrsquos EPR is still 67 higher than CSMACA

In conclusion the jitter of the wake-up delay at each end-device leads tothe improvement of PWUR MAC over CSMACA In the following sections

we will explain that this network performance improvement is related tohardware parameters end-nodesrsquo location etc

4222 Influence of wireless channel performance

In the analysis above propagation loss of WuS follows the Friis model inEquation 44

Pr = Pt +Gt +Gr + 20 log 10(λ

4πd) (44)

where the distance between WuRx and WuTx determines the WuS signalstrength It is good enough if we simply want to test the MAC layer per-formance of our PUWR model The wireless channel performance on thePHY layer is simplified However in the real-case scenario the channelperformance is a massively important factor in any wireless networks Es-pecially in our PWUR-based WSN wake-up delay is directly related to Pr ofWuS It is worthwhile to use a separate section to investigate when a morerealistic propagation loss model is applied what the influence of wirelesschannel performance will be

We chose the log-distance path loss model in Equation 45

PL = Pt minus Pr = PL0 + 10γ log10d

d0+Xg (45)

where PL is the total path loss PL0 is the path loss at the reference distanced0 γ is the path loss exponent Xg is a Gaussian random variable with zeromean reflecting the attenuation caused by flat fading In case of no fadingthis variable is 0 In case of only shadow fading or slow fading this randomvariable may have Gaussian distribution with σ standard deviation in dBresulting in log-normal distribution of the received power in Watt In anindoor environment like our case γ is set to be 276 and Xg is 6 dB [30]

Given the distance between WuRx and WuTx the signal strength of WuSis not fixed anymore Nodes can be woken up when Pr of the WuS ishigher than the WuRx circuits sensitivity level (-293 dBm) Figure 419shows the theoretical and our simulation result on wake-up probability atdifferent ranges Theoretical result is obtained by calculating the CDF ofthe receive power of WuS in MATLAB The wake-up probability is theaccumulative probability when the signal strength is higher than the WuRxrsquossensitivity value Each result is obtained by 100 OPNET simulation runsThe successful wake-up times divided by 100 is the simulational wake-upprobability

It can be seen that wake-up probability drops as distance increases Between0sim4 meters wake-up probability is almost 1 When the distance becomes 20meters the wake-up probability is around 02 meaning that when WuRx is20 meters away from the coordinator it almost takes 5 WuSs to wake up the

0 5 10 15 200

Wakeminusup probability VS Distance

Distance between WuTx amp WuRx (m)

eminusup

Analytical result

Simulation result

Figure 419 Relation between wake-up probability and distance (log-distance path loss model)

end-node once The end-node continues to generate data packet at the rateof 1 pkts It buffers all the packets until its WuRx is successfully woken upby one WuS Given the same nodes distribution as in the previous sectionswe have the packet delay shown in Figure 420 using the log-distance pathloss model

The end-to-end delay can be as much as 2 to 18 seconds This is signi-ficantly higher compared to the results with the Friis model which staysaround 01s (Figure 413) Failed wake-ups force more packets into the buf-fer Since wake-up signals are transmitted by 1 pkts each failed wake-upincreases the delay by 1 second Also the packets will be stored in the buf-fer without successfully received WuS Packets join the medium contentionwhen the previous packet in the buffered has been successfully transmittedSo a ldquobuffering backoffrdquo is introduced Finally the larger pathloss exponent(in log-distance model γ in Equation 45 is 276 while it is 2 for Friis model)also contributes slightly to the longer delay

The comparison of packet delivery ratio by using Friis model and log dis-tance model to calculate propagation loss is shown in Figure 421 It isworth mentioning that since multiple packets in the buffer can be transmit-ted after one single wake-up the packet delivery ratio can surpass 1

Deviation of packet delivery ratio using log-distance model is much moresevere than the one with Friis model We simulate 100 successful wake-upsto nodes at distance from 1 - 20 meters to the WuTx The mean value andstandard deviation of their wake-up delay is shown in Fig 422

There are two reasons why in Figure 421 when node density is highsystem using log-distance model has a higher packet delivery ratio Firstthe deviation of wake-up delay in Figure 422 spreads the starting point of

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

Figure 420 Packet delay with log-distance path loss model

0 50 100 150 200 250 300 3500

Packet Delivery Ratio of PWUR N = 10

logminusdistance model

Friis model

0 50 100 150 200 250 300 3500

Simulation Time (seconds)_

Packet Delivery Ratio of PWUR N =30

logminusdistance modelFriis model

0 50 100 150 200 250 300 3500

18Packet Delivery Ratio of PWUR N =50

0 50 100 150 200 250 300 3500

1Packet Delivery Ratio of PWUR N = 100

Figure 421 Influence of wireless channel performance on packet deliveryratio

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

2Wakeminusup delay different distances

Distance between WuTx and WuRx

eminusup

standard deviation

Figure 422 Mean of standard deviation of wake-up delay log-distance pathloss model

CSMACA randomly In the Friis model nodes with the same distanceto WuTx have the same fixed wake-up delay PWUR MAC does not havemuch impact on those nodes However as now the wake-up delay has alarge standard deviation even nodes with the same distance to WuTx canbenefit from the wake-up delay

Second as we have mentioned the buffering backoff it could also lead tothe fact that less packets content for the medium at the same time as someother packets have already completed their contention during a packetrsquosbuffering backoff

Overall a more dynamic wireless always channel provides worse latencyregardless of the node number and better packet delivery ratio when nodedensity is high Combining these two performances together the EPR com-parison is shown in Table 43

In conclusion when we apply a more realistic propagation loss model toour PWUR model we find out that the packet delay performance becomessignificantly worse while the delivery ratio performance maintains similaror sometimes even better In total the EPR drops In the real-world wire-less channel channel performance is even more dynamic and unpredictablePWURs overall network performance may be even worse

Node number

Log-distance

227524

105011

Pathloss Model

Table 43 EPR comparsion of log-distance and Friis pathloss model

Chapter 5

Conclusions and FutureWork

51 Conclusions

The primary goal of this thesis work is to investigate how PWUR can affectthe ZigBee networkrsquos overall performance We use OPNET and MATLABto build our simulations and focus on two specific problems to approach ourgoal

In terms of the comparison of duty cycling and PWUR in ZigBee net-work we find out that PWUR can indeed break the energy-latency tradeoffin duty cycling As for the latency performance the on-demand PWURcan indeed outperform the pre-scheduled duty cycling However as packetrate increases the energy consumption performance of the low duty cyclingscheme which also avoids the idle-listening activity is able to compete withPWUR We also proposed our CSMACA + TDMA combination scheme tohandle the multiple nodes scenario Large number of nodes are divided intoclusters by TDMA CSMACA is performed within a cluster for mediumaccess control When the number of nodes in a ZigBee network increasesthe latency performance gets worse while the node number does not affectenergy consumption performance that much

In terms of PWURrsquos influence on the MAC layer performance we find outthat the introduction of PWUR can improve the ZigBee networkrsquos effectivepacket rate up to 67 to 128 compared to CSMACA due to the jitterof wake-up delay for different end-nodes before the initial CSMACA back-off starts However the network size nodes location hardware design andpathloss model determine how much the performance improvement will beexactly

Figure 51 Network performance optimization loop

52 Future Work

A number of additions on this topic can be introduced in the future workFirst in Section 33 we proposed the combination scheme of TDMA

and CSMACA This scheme can be optimized by investigating the relationbetween cluster size and network size Clusters are divided by TDMA Nodeswithin the same cluster are divided by CSMACA Both these two processinclude overhead By selecting the right cluster size the sum of TDMA andCSMACArsquos overhead will be minimized

Second the wake-up delay of PWUR functions as an additional back-offbefore the CSMACA back-off starts Since the back-off length differs fromnode to node it improves the MAC layer performance by decreasing thetotal back-off delay and avoiding collisions in CSMACA We find out thatthere is an optimization loop in this process

From Equation 21 and 22 we understand that necessary amount of har-vested energy E and the receive power of wake-up signal Pr will determinehow long the wake-up process take E is dependent on the capacitance ofthe capacitor that stores the energy Pr is dependent on the location thedistance between WuRx and WuTx We have mentioned that wake-up delayis calculated by E Pr minus Ploss

Therefore we come up with the optimization loop shown in Figure 51With the right capacitance and location of WuRx the wake-up delay ofeach node will lead to a better system performance Vice versa both thePWUR circuit design and nodes distribution can be optimized based onsystem performance An optimization algorithm can be developed followingthis principle

Last but not least since PWUR analysis is a relatively new field by thetime this thesis work was completed we were not able to use PWUR-enabledZigBee device to validate our results In the future instead of purely based

on theoretical analysis a similar study can make use of measured data bydevices like WISP to make it more realistic

Bibliography

[1] httpwwwzigbeeorg[2] httpwwwbluetoothcomPagesBluetooth-Homeaspx[3] httpenhartcommorghcptecharticleswiHART_resources

witech_witechnotehtml[4] httpswwwisaorgisa100[5] httpwwwnxpcomtechzoneswireless-connectivityjn51xx

jn516xhtml[6] nrf24l01 single chip 24 GHz transceiver product specification httpwww

nordicsemicomindexcfmobj=productampact=displayamppro=89[7] OPNET httpwwwriverbedcomproducts

performance-management-controlopnethtml[8] EPCglobal UHF Class-1 Generation-2 Standard httpwwwgs1org

gsmpkcepcglobaluhfc1g2uhfc1g2_1_2_0-standard-20080511pdf2008

[9] IEEE 802154 macphy standard for low-rate wireless personal area net-works(lrwpans) httpwwwieee802org15pubTG4html 2013

[10] I F Akyildiz W Su Y Sankarasubramaniam and E Cayirci A surveyon sensor networks IEEE Communications Magazine 40(8)102ndash114 August2002

[11] Th Arampatzis J Lygeros and S Manesis A survey of applications ofwireless sensors and wireless sensor networks pages 719ndash724 2005 IEEE In-ternational Symposium Mediterrean Conference on Control and AutomationIntelligent Control 2005

[12] He Ba Ilker Demirkol and Wendi Heinzelman Passive wake-up radios Fromdevices to applications Ad Hoc Networks 112605 ndash 2621 August 2013

[13] M Buettner G Yee E Anderson and R Han X-mac a short preamblemac protocol for duty-cycled wireless sensor networks ACM Conference onEmbedded Sensor Systems 2006

[14] B Chandrasekaran Survey of Network Traffic Models httpwwwcs

wustledu~jaincse56706ftptraffic-models3indexhtml

[15] L Chen S Cool H Ba W Heinzelman I Demirkol U Muncuk K RChowdhury and S Basagni Range extension of passive wake-up radio systemsthrough energy harvesting pages 1549 ndash 1554 Proc IEEE ICC BudapestHungary June 2013

[16] H J Cho H C Kim X Yao M S Kim S W Kwon T J Park H S Kimand Y G Yang Highly sensitive cmos passive wake-up circuit pages 1 ndash 4Proc Asian Pacific Microw Conf (APMC 2008) 2008

[17] W Dargie and CPoellabauer Fundamentals of wireless sensor networkstheory and practice pages 168 ndash 183 191 ndash 192 John Wiley and Sons 2010

[18] J Demirkol C Ersoy and E Onur Wake-up receivers for wireless sensornetworks benefits and challenges IEEE Wireless Communications 16(4)88ndash 96 2009

[19] BVd Doorn W Kavelaars and K Langendoen A prototype low-costwakeup radio for the 868 mhz band International Journal on Sensor Net-works 522 ndash 32 2009

[20] L Gu and J Stankovic Radio-triggered wake-up capability for sensor net-works pages 27 ndash 36 Real-Time and Embedded Technology and ApplicationsSymposium 2004 IEEE 2004

[21] R Jurdak A G Ruzzelli and G M P OrsquoHare Multi-hop rfid wake-up radiodesign evaluation and energy tradeoffs pages 1 ndash 8 Proc of 17th ICCCNConference IEEE Aug 2008

[22] A Khatibi Y Durmus E Onur and I Niemegeers Event-driven mac pro-tocol for dual-radio cooperation pages 1 ndash 5 Vehicular Technology Conference(VTC Fall) IEEE September 2012

[23] P Le-Huy and S Roy Low-power 2 4 ghz wake-up radio for wireless sensornetworks pages 13ndash18 Int Conf on Wireless and Mobile Computing Net-working and Communication 2008

[24] Sang Hoon Lee Yong Soo Bae and L Choi On-demand radio wave sensorfor wireless sensor networks Towards a zero idle listening and zero sleepdelay mac protocol pages 560 ndash 566 Global Communications Conference(GLOBECOM) IEEE December 2012

[25] H Liu M Bolic A Nayak and I Stojmenovic Taxonomy and challenges ofthe integration of rfid and wireless sensor networks IEEE Network 22(6)26ndash 32 2008

[26] M Lont D Milosevic P G M Baltus A H M Van Roermund and G Dol-mans Analytical models for the wake-up receiver power budget for wirelesssensor networks pages 1 ndash6 Proc IEEE Global Telecommun Conf (GLOBE-COM09) 2009

[27] Charith Perera Arkady Zaslavsky Peter Christen and Dimitrios Georgako-poulos Context aware computing for the internet of things A survey IEEECommunications Surveys and Tutorials 16(1)414 ndash 454 2013

[28] N Pletcher S Gambini and J Rabaey A 65 uw 19 ghz rf to digital basebandwakeup receiver for wireless sensor nodes pages 539ndash542 Proc Custom IntegrCircuits Conf 2007

[29] P Radmand A Talevski S Petersen and S Carlsen Comparison of indus-trial wsn standards pages 632 ndash 637 4th IEEE International Conference onDigital Ecosystems and Technologies (DEST) 2010

[30] T S Rappaport Wireless communications principles and practices Prentice-Hall 2002

[31] AP Sample DJ Yeager PS Powledge AV Mamishev and JR SmithDesign of an rfid-based battery-free programmable sensing platform IEEETransactions on Instrumentation and Measurement 57(11)2608 ndash 2615 2008

[32] L Shan T Yunjian and Z Qin Passive wake-up scheme for wireless sensornetwork page 507 Proc 2nd Int Conf Innovative Comp Inform Control2007

[33] K Sohraby D Minoli and T Znati Wireless sensor networks technologyprotocols and applications pages 203 ndash 209 John Wiley and Sons 2007

[34] P Sthapit and Jae-Young Pyun Effects of radio triggered sensor mac protocolover wireless sensor network pages 546 ndash 5516 Computer and InformationTechnology (CIT) IEEE 2011

[35] L Zhang and Z Wang Integration of rfid into wireless sensor networksArchitectures opportunities and challenging problems pages 433 ndash 469 ProcGrid Coop Comput Workshops 2006

[36] Y Zhang L Huang G Dolmans and H de Groot An analytical model forenergy efficiencyanalysis of different wakeup radio schemes pages 1148 ndash1152PersonalIndoor and Mobile Radio Communications IEEE 2009

Preface

Introduction

Overview of WSN

WSN Supporting standards

Wake-up radio

Problem statement

Structure of the report

Background

Hardware design

MAC layer performance

Theoretical Model

Wake-up process

Analytical model for duty cycling and PWUR

Duty Cycling

PWUR

Medium access control

TDMA

CSMACA

Combination

Simulations

Simulation setup

ZigBee model in OPNET

Wake-up radio implementation

Simulation assumptions

Results and analysis

Duty cycling VS PWUR

PWUR on CSMACA

Conclusions and Future Work

Conclusions

Future Work

A Comprehensive Study of Passive Wake-up Radio

in Wireless Sensor Networks

Masterrsquos Thesis in Electrical Engineering

Embedded Software SectionFaculty of Electrical Engineering Mathematics and Computer Science

Delft University of TechnologyMekelweg 4 2628 CD Delft The Netherlands

Liang HuoLHuostudenttudelftnl

10072014

Abstract

Preface

Liang Huo

Contents

Preface v

1 Introduction 1

13 Wake-up radio 3

2 Background 9

321 Duty Cycling 18

322 PWUR 23

331 TDMA 26

332 CSMACA 27

333 Combination 29

4 Simulations 33

Chapter 1

Introduction

11 Overview of WSN

Energy

Latency

Reliability

ZigBee

13 Wake-up radio

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Abstract

Preface

Liang Huo

Contents

Preface v

1 Introduction 1

13 Wake-up radio 3

2 Background 9

321 Duty Cycling 18

322 PWUR 23

331 TDMA 26

332 CSMACA 27

333 Combination 29

4 Simulations 33

Chapter 1

Introduction

11 Overview of WSN

Energy

Latency

Reliability

ZigBee

13 Wake-up radio

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Abstract

Preface

Liang Huo

Contents

Preface v

1 Introduction 1

13 Wake-up radio 3

2 Background 9

321 Duty Cycling 18

322 PWUR 23

331 TDMA 26

332 CSMACA 27

333 Combination 29

4 Simulations 33

Chapter 1

Introduction

11 Overview of WSN

Energy

Latency

Reliability

ZigBee

13 Wake-up radio

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Preface

Liang Huo

Contents

Preface v

1 Introduction 1

13 Wake-up radio 3

2 Background 9

321 Duty Cycling 18

322 PWUR 23

331 TDMA 26

332 CSMACA 27

333 Combination 29

4 Simulations 33

Chapter 1

Introduction

11 Overview of WSN

Energy

Latency

Reliability

ZigBee

13 Wake-up radio

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Preface

Liang Huo

Contents

Preface v

1 Introduction 1

13 Wake-up radio 3

2 Background 9

321 Duty Cycling 18

322 PWUR 23

331 TDMA 26

332 CSMACA 27

333 Combination 29

4 Simulations 33

Chapter 1

Introduction

11 Overview of WSN

Energy

Latency

Reliability

ZigBee

13 Wake-up radio

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Contents

Preface v

1 Introduction 1

13 Wake-up radio 3

2 Background 9

321 Duty Cycling 18

322 PWUR 23

331 TDMA 26

332 CSMACA 27

333 Combination 29

4 Simulations 33

Chapter 1

Introduction

11 Overview of WSN

Energy

Latency

Reliability

ZigBee

13 Wake-up radio

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Contents

Preface v

1 Introduction 1

13 Wake-up radio 3

2 Background 9

321 Duty Cycling 18

322 PWUR 23

331 TDMA 26

332 CSMACA 27

333 Combination 29

4 Simulations 33

Chapter 1

Introduction

11 Overview of WSN

Energy

Latency

Reliability

ZigBee

13 Wake-up radio

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 1

Introduction

11 Overview of WSN

Energy

Latency

Reliability

ZigBee

13 Wake-up radio

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 1

Introduction

11 Overview of WSN

Energy

Latency

Reliability

ZigBee

13 Wake-up radio

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Latency

Reliability

ZigBee

13 Wake-up radio

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

13 Wake-up radio

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 2

Background

21 Hardware design

Range-based

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

V out =

radic2 times E

Cse(22)

ID-based

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 3

Theoretical Model

31 Wake-up process

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

loss included

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

321 Duty Cycling

Latency

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

T = Tactiveη (34)

T2 = Tlisten

T3 = Tsleep

T = T1 + T2 + T3

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Energy Consumption

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Algorithm

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

130 us

120582

322 PWUR

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Parameters Power

855 uw

339 mw

246 mw

354 mw

Epacket =EtotalM

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

331 TDMA

N = 1n (313)

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

332 CSMACA

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

10 4 006 s

100 75 065 s

333 Combination

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 4

Simulations

41 Simulation setup

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

elsebuffer the data

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

4211 Latency

T1 = 00028s T2 = 005s T3 = 04751s

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

0 01 02 03 04 050

Packet delay (s)

05 1 15 2 25 3

10 20 30 40 50 60

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

05 1 15 2 25 310

minus1

10 20 30 40 50 6010

minus1

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

10 888

15 1387

20 162

25 1976

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

10 20 30 40 50 600

N = 10N = 30N = 50N = 100N = 150N = 200N = 250

422 PWUR on CSMACA

CSMACA

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Node 1

Node 4

Node 3

Node 2

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

PWUR on CSMA

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

N = 10N = 30N = 50N = 100

10 0052 0040 0042

30 0100 0118 0082

50 0116 0190 0098

100 0124 0376 0150

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

0 50 100 150 200 250 300 3500

EPR N = 10

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 30

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 50

TDMACSMACAPWUR MAC

0 50 100 150 200 250 300 3500

EPR N = 100

TDMACSMACAPWUR MAC

EPR = ptimes 1

infin=

Depresents

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

N=10 N=30 N=50 N=1000

Node density

CSMACA

PWUR MAC

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

4πd) (44)

d0+Xg (45)

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

0 5 10 15 200

eminusup

Analytical result

Simulation result

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

0 50 100 150 200 250 300 3500

Simulation Time (s)

N = 10N = 30N = 50N = 100

0 50 100 150 200 250 300 3500

Friis model

0 50 100 150 200 250 300 3500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210

eminusup

standard deviation

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Node number

Log-distance

227524

105011

Pathloss Model

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Chapter 5

51 Conclusions

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

52 Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Bibliography

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Preface

Introduction

Overview of WSN


Wake-up radio

Problem statement


Background

Hardware design


Theoretical Model

Wake-up process


Duty Cycling

PWUR


TDMA

CSMACA

Combination

Simulations

Simulation setup






PWUR on CSMACA


Conclusions

Future Work

Documents

A Comprehensive Study of Passive Wake-up Radio in Wireless