Upload
others
View
27
Download
0
Embed Size (px)
Citation preview
Matlab/Simulink wireless HDMI model and simulation
Ruben Danilo Jesus Cabral
Dissertation submitted to obtain the Master (MsC) degree in
Electrical and Computer Engineering
Jury
President: Prof. Marcelino Santos
Supervisor: Prof. Helena Sarmento
Member: Prof. José Gerald
October 2011
ii
Acknowledgments
I want to express my deepest thanks to my supervisor, Professor Helena Sarmento for her guidance, advice,
suggestions, time and immeasurable patience. Her contributions improved this dissertation well beyond what I
could have ever done by myself.
I also want to thank André Glória and Ricardo Faria for their companionship throughout the development of
this work.
A special thanks to my mother and sister for their support and incentive.
Finally I would like to share my profound thanks to Cristina for her unconditional support.
This work has been performed under the project ”SideWorks”, no 3487, QREN - Projecto de I&DT em Co-
Promoção, the QREN and partially supported by FCT (INESC-ID multiannual funding) through the PIDDAC
Program funds.
iii
Abstract
The growing availability of High Definition video contents is driving the development of technologies capable
of multi-gigabit per second throughput, like High-Definition Multimedia Interface and Display Port. As wireless
communication systems have become common in everyday life, video transmission is also pushing wireless
technologies.
The 60 GHz band is extremely appealing due to the huge continuous bandwidth available, up to 7 GHz of
continuous bandwidth in some world regions, which provides enough throughput to transmit uncompressed
FullHD video. Several standards have been developed to take advantage of this frequency band like ECMA-387,
IEEE 802.15.3c and IEEE 802.11ad, and proprietary specifications like WirelessHD and Wireless Gigabit.
The objective of this work is to model in Matlab/Simulink a transceiver in order to assist the hardware
implementation, in FPGA, of an uncompressed HD video transceiver.
Key-words
Wireless, uncompressed video transmission, 802.15.3c, AV HRP, millimeter-wave, Simulink modeling
iv
Resumo
A crescente oferta de conteúdos vídeo de Alta Definição (HD) tem promovido o desenvolvimento de
tecnologias capazes de débitos na ordem de vários gigabits por segundo, como High-Definition Multimédia
Interface e Display Port. Á media que os sistemas de comunicações sem fios se têm tornado mais comuns no
dia-a-dia, estes são também influenciados pela necessidade de maiores débitos dos actuais conteúdos de
vídeo HD.
A banda dos 60 GHz oferece até 7 GHz de largura de banda contínua, disponibilizando largura de banda
suficiente para a transmissão de vídeo HD sem compressão. Várias normas têm vindo a ser desenvolvidas para
utilizar a banda dos 60 GHz, como o ECMA-387, IEEE 802.15.3c e IEEE 802.11ad, assim como formatos
proprietários como o WirelessHD e Wireless Gigabit.
Este trabalho destina-se a desenvolver um modelo em Matlab/Simulink de um transceiver de vídeo, capaz
de transmitir vídeo HD sem compressão, para auxiliar uma futura implementação em FPGA
Palavras-chave
Transmissão de vídeo sem fios, 802.15.3c, modelação Simulink
v
Contents
Introduction ...............................................................................................................................................................1
1.1 Motivation .......................................................................................................................................................1
1.2 Objectives.......................................................................................................................................................3
1.3 Dissertation outline.........................................................................................................................................3
Wireless HD video transmission...............................................................................................................................4
2.1 Digital video and audio signals.......................................................................................................................4
2.2 Wired HD video standards .............................................................................................................................5
2.3 Wireless video transmission technologies .....................................................................................................5
2.3.1 Technologies using frequencies up to 10 GHz.......................................................................................5
2.3.2 Technologies in the 60 GHz band ..........................................................................................................7
2.3.3 Conclusions on wireless HD video transmission ................................................................................. 12
IEEE 802.15.3c AV mode...................................................................................................................................... 15
3.1 HRP PHY frame structure ........................................................................................................................... 16
3.2 Splitting........................................................................................................................................................ 17
3.3 Forward error correction.............................................................................................................................. 18
3.3.1 Block encoding..................................................................................................................................... 18
3.3.2 Convolutional encoding........................................................................................................................ 18
3.4 Interleavers.................................................................................................................................................. 19
3.4.1 Outer interleaver .................................................................................................................................. 19
3.4.2 Bit interleaver ....................................................................................................................................... 20
3.4.3 Tone interleaver ................................................................................................................................... 21
3.5 Sub carrier modulation ................................................................................................................................ 21
IEEE 802.15.3c AV HRP mode transceiver Simulink model ................................................................................. 24
4.1 Transceiver blocks ...................................................................................................................................... 24
4.1.1 Transmitter and receiver controller ...................................................................................................... 24
4.1.2 Frame processing blocks..................................................................................................................... 24
4.1.3 Frame encoding and decoding blocks ................................................................................................. 28
4.1.4 Bit interleaver and deinterleaver blocks............................................................................................... 31
4.1.5 OFDM modulator and demodulator blocks .......................................................................................... 32
4.2 Complete Matlab/Simulink model................................................................................................................ 34
4.2.1 HRPDU frame generator...................................................................................................................... 34
4.2.2 Channel block ...................................................................................................................................... 34
4.2.3 Model settings...................................................................................................................................... 35
Tests and results ................................................................................................................................................... 37
Conclusion ............................................................................................................................................................. 40
References ............................................................................................................................................................ 42
Annex A. IEEE 802.15.3c specifications ......................................................................................................... 44
vi
Annex B. ECMA-387 specifications................................................................................................................. 47
Annex C. IEEE 802.1ad specifications ............................................................................................................ 49
Annex D. HRPDU detailed structure................................................................................................................ 50
D.1 HRP preamble........................................................................................................................................ 50
D.2 HRP header............................................................................................................................................ 50
D.2.1 PHY control ......................................................................................................................................... 50
D.2.2 Sub-frame header ............................................................................................................................... 50
D.3 Extended MAC header ........................................................................................................................... 51
Annex E. Matlab/Simulink detailed block diagrams......................................................................................... 59
vii
List of Tables
Table 1: Technologies up to 10 GHz ........................................................................................................................7
Table 2: 802.15.3c frequency channels [2] ..............................................................................................................8
Table 3: ECMA-387 frequency usage and channel bonding [5]............................................................................ 10
Table 4: 802.11ad OFDM modulation timing parameters [7] ................................................................................ 11
Table 5: Comparison of the 60 GHz technologies................................................................................................. 12
Table 6: Comparison criteria ................................................................................................................................. 13
Table 7: HRP data rates and coding [3] ................................................................................................................ 15
Table 8: HRP OFDM timing parameters [3]........................................................................................................... 16
Table 9: Puncturing pattern ................................................................................................................................... 19
Table 10: HRPDU header bit interleaver ............................................................................................................... 20
Table 11: HRPDU header bit interleaver (cont.).................................................................................................... 21
Table 12: HRPDU data bit interleaver ................................................................................................................... 21
Table 13: Tone interleaving table .......................................................................................................................... 21
Table 14: Real and imaginary values for QPSK modulation ................................................................................. 22
Table 15: Real and imaginary values for 16-QAM modulation.............................................................................. 22
Table 16: Real and imaginary values for 16-QAM modulation (cont.) .................................................................. 23
Table 17: Test HRPDU frame structure................................................................................................................. 37
Table 18: Test results ............................................................................................................................................ 38
Table 19: Determining minimum numerical precision for the IFFT block .............................................................. 38
Table 20: Determining minimum numerical precision for the FFT block ............................................................... 39
Table 21: Influence of the number of soft bits ....................................................................................................... 39
Table 22: SC PHY MCS parameters ..................................................................................................................... 44
Table 23: HSI MCS parameters [3] ....................................................................................................................... 44
Table 24: HSI MCS parameters [3] (cont.) ............................................................................................................ 45
Table 25: HSI OFDM timing parameters [3] .......................................................................................................... 45
Table 26: LRP frequency plan [3] .......................................................................................................................... 45
Table 27: LRP data rates and coding [3] ............................................................................................................... 46
Table 28: LRP OFDM timing parameters [3] ......................................................................................................... 46
Table 29: Work modes for Type A devices............................................................................................................ 47
Table 30: Working modes for Type B devices....................................................................................................... 48
Table 31: MCS modes for 802.11ad SC modulation [7]........................................................................................ 49
Table 32: MCS modes for 802.11ad OFDM modulation [7] .................................................................................. 49
Table 33: HRP header structure............................................................................................................................ 50
Table 34: PHY control structure ............................................................................................................................ 50
Table 35: Sub-frame structure............................................................................................................................... 51
Table 36: HRP data rates and coding ................................................................................................................... 51
Table 37: Extended MAC header .......................................................................................................................... 51
viii
Table 38: MAC header format ............................................................................................................................... 52
Table 39: Frame Control field ................................................................................................................................ 52
Table 40: Frame types........................................................................................................................................... 52
Table 41: ACK policy types ................................................................................................................................... 53
Table 42: Reserved device ID ............................................................................................................................... 54
Table 43: Fragmentation Control structure............................................................................................................ 54
Table 44: Stream Index reserved values............................................................................................................... 54
Table 45: Extended control header ....................................................................................................................... 55
Table 46: Regular class frame type values ........................................................................................................... 55
Table 47: MAC extension header format............................................................................................................... 55
Table 48: Allowed AC class sub frame types ........................................................................................................ 56
Table 49: ACK groups field.................................................................................................................................... 56
Table 50: Security header structure ...................................................................................................................... 56
Table 51: Security control field .............................................................................................................................. 57
Table 52: Video header field.................................................................................................................................. 57
Table 53: Video control field .................................................................................................................................. 57
Table 54: Non secure MAC body frame ................................................................................................................ 58
Table 55: Secure MAC body frame ....................................................................................................................... 58
ix
List of Figures
Figure 1: Wireless video network example...............................................................................................................2
Figure 2: HRPDU frame format ............................................................................................................................. 16
Figure 3: EEP splitting ........................................................................................................................................... 17
Figure 4: UEP splitting........................................................................................................................................... 18
Figure 5: Outer interleaver pattern for depth=2, N=56 .......................................................................................... 20
Figure 6: Outer interleaver pattern for depth=4, N=224 ........................................................................................ 20
Figure 7: HRP transceiver main blocks ................................................................................................................. 25
Figure 8: Frame conditioning block (transmitter)................................................................................................... 26
Figure 9: Frame conditioning block (receiver) ....................................................................................................... 26
Figure 10: HRP scrambler and descrambler block................................................................................................ 27
Figure 11: HRP transmitter Reed-Solomon encoder............................................................................................. 28
Figure 12: HRP receiver Reed-Solomon decoder block ....................................................................................... 29
Figure 13: Outer interleaver block internal structure (transmitter)......................................................................... 29
Figure 14: Individual convolutional encoder structure ........................................................................................... 30
Figure 15: Multiplexer/bit interleaver block internal structure................................................................................ 31
Figure 16: HRP demultiplexer and bit deinterleaver.............................................................................................. 32
Figure 17: OFDM modulator block ........................................................................................................................ 33
Figure 18: OFDM demodulator block .................................................................................................................... 33
Figure 19: Complete Simulation model ................................................................................................................. 34
Figure 20: HRPDU frame generator ...................................................................................................................... 35
Figure 21: Model top level view ............................................................................................................................. 36
Figure 22: Detailed HRP transmitter diagram ....................................................................................................... 59
Figure 23: Detailed HRP receiver diagram............................................................................................................ 60
Figure 24: Detailed HRP frame conditioning block (transmitter) ........................................................................... 61
Figure 25: Detailed HRP frame conditioning block (receiver) ............................................................................... 61
Figure 26: Detailed HRP scrambler/descrambler block ........................................................................................ 62
Figure 27: Detailed HRP Reed-Solomon encoder block (transmitter) .................................................................. 62
Figure 28: Detailed HRP Reed-Solomon decoder block (receiver)....................................................................... 63
Figure 29: Detailed HRP outerinterleaver block (transmitter)................................................................................ 63
Figure 30: Detailed HRP multiplexer/bit interleaver block (transmitter) ................................................................ 64
Figure 31: Detailed HRP demultiplexer block (receiver) ....................................................................................... 64
Figure 32: Detailed HRP OFDM modulator block (transmitter) ............................................................................. 65
Figure 33: Detailed HRP demodulator block (receiver)......................................................................................... 65
x
Glossary
ABR – Audio bit rate;
ADC - Analogue to Digital Converter;
ASIC – Application Specific Integrated Circuit;
AV - Audio/Visual;
AWGN – Additive White Gaussian Noise;
BER – Bit error rate;
BPSK - Binary Phase-Shift Keying;
CE - Consumer Electronics;
CPU - Central Processing Unit;
CRC - Cyclic Redundant Check;
DAC - Digital to Analogue Converter;
DC – Direct Current;
DVI - Digital Visual Interface;
EEP - Equal Error Protection;
FCS - Frame Check Sequence;
FEC - Forward Error Correction;
FFT - Fast Fourier Transform;
FPGA - Field Programmable Gate Array;
Gbps - Gigabits per second;
GPU - Graphics Processor Unit;
HCS - Header Check Sequence;
HD - High Definition;
HDCP - High Definition Content Protection;
HDMI - High-Definition Multimedia Interface;
HRP - High-Rate PHY;
HRPDU - High Rate Protocol Data Unit;
HSI PHY - High Speed Interface;
IEC - International Electrotechnical Commission;
IEEE - Institute of Electrical and Electronics Engineers;
IFFT - Inverse Fast Fourier Transform;
ISO - International Organization for Standardization;
LDPC - Low-Density Parity-Check;
LOS - Line of sight;
LRP - Low-Rate PHY;
LSB - Least Significant Bits;
MAC - Medium Access Control;
MB OFDM – Multiband Orthogonal Frequency Division Multiplexing;
xi
Mbps - Mega bits per second;
MCS - Modulation and Coding Scheme;
MCTA - Management Channel Time Allocation;
MSDU - MAC Service Data Unit;
MIMO – Multiple Inputs, Multiple Outputs;
MSB - Most Significant Bits;
MSK - Minimum-Shift Keying;
MUX – Multiplexer;
N.A. – Not applicable;
NLOS – Non line of sight;
OFDM - Orthogonal Frequency Division Multiplexing;
PAL - Protocol Adaptation Layer;
PCIe - Peripheral Component Interconnect Express;
PHY - Physical Layer;
PNC - PicoNet Controller;
PNCID - PicoNet Controller Identification;
PVR - Personal Video Recorder;
QAM – Quadrature Amplitude Modulation;
QPSK – Quadrature Phase-Shift Keying;
RF - Radio Frequency;
RS - Reed-Solomon;
SCBT - Single Carrier Block Transmission;
SC PHY - Single Carrier Physical Layer;
SD – Standard Definition;
SECID - Secure Session ID;
SFC - Secure Frame Counter;
SG3c – Study Group 3c;
SQPSK – Staggered Quadrature Phase-Shift Keying;
SNR - Signal to Noise Ratio;
TCM – Trellis Coded Modulation;
TDMA - Time Division Multiple Access;
TG3c - Task Group 3c;
TGac - Task Group ac;
TGad - Task Group ad;
UEP - Unequal Error Protection;
USB - Universal Serial Bus;
UWB - Ultra Wide Band;
VBR - Video bit rate;
xii
xiii
VESA - Video Electronics Standards Association;
VGA - Video Graphics Array;
VHT - Very High Throughput;
WHDI - Wireless Home Digital Interface;
WiDi - Wireless Display;
Wi-Fi – Wireless Fidelity;
WiGig - Wireless Gigabit;
WiHD – WirelessHD;
WiMAX - Worldwide Interoperability for Microwave Access;
WPAN - Wireless Personal Area Network;
WUSB - Wireless Universal Serial Bus;
WVAN - Wireless Video Area Network;
Chapter 1
Introduction
1.1 Motivation
The recent increase in High Definition (HD) video contents has brought the need to develop communication
standards capable of multi-gigabit per second throughput, like High-Definition Multimedia Interface (HDMI) [1]
and Display Port [2]. Consumer electronics (CE) users also want the flexibility provided by wireless connections
to set up and reconfigure multimedia systems, and to eliminate wired connections required by HD multimedia
systems, like home theatres. Driven by these needs, the CE industry is developing formats capable of delivering
uncompressed video, at the necessary data rates, via wireless connections. Simultaneously, 802.11 devices
have become ubiquitous and, in the latest specification (802.11n [21]), are capable of net data rates up to
600Mbps. This data rate is still insufficient for streaming uncompressed HD video or for transferring HD contents,
like a HD film, as fast as would be desirable. Uncompressed HD video transmission requires very high bit rates,
up to 3 Gbps for Full HD video.
Uncompressed HD video transmission avoids compression at the transmitter and decompression at the
receiver, therefore providing: lower latency which permits timing sensitive applications like multimedia
applications and gaming; higher interoperability between devices, because, unlike compressed video
transmission, the receiver device just displays the video content and does not need to be able to decoded the
video codec; and no degradation in picture quality due to compression losses in the transmission.
To address these needs several specifications have been created and several other are still being developed.
The development efforts have been focused in basically in two frequency bands: the 2.4 to 10 GHz and 60 GHz
bands. The former frequency band is used by specifications like Wireless Home Digital Interface (WHDI) [11] and
Intel's Wireless Display (WiDi) [12], WirelessUSB (WUSB) [9], and a draft version of IEEE 802.11ac [6], while the
later is used by IEEE 802.15.3c [3], WirelessHD (WiHD) [4], ECMA-387 [5], Wireless Gigabit (WiGig) [8] and a
draft version of IEEE 802.11ad [7]. Technologies using the lower frequency band have better range and are not
limited to in room transmissions. The propagation characteristic of the 60 GHz band limits the transmissions to a
maximum distance of 10 meters and to in room transmissions. However, the 60 GHz band has more available
bandwidth and therefore higher bit rates can be achieved using this frequency band.
The technologies specifically designed for wireless video transmission, like WHDI, IEEE 802.15.3c, ECMA-
387, WiHD, IEEE 802.11.ad (draft) and WiGig, organize the source devices (transmitters) and the sink devices
(receivers) into a wireless video network, as shown in Figure 1, that allows for example:
Point to point uncompressed video transmission.
Point to multi point uncompressed video transmission.
Office desktop, allowing to wirelessly transmitting a laptop/computer desktop to a HD display.
1
The wireless video network shown on Figure 1, has three type of devices: video sink (HD display), video
sources (set top box and mobile device) and devices that can perform both tasks (laptop).
Figure 1: Wireless video network example
As of the time of this writing there are some WHDI, WiGig, WiHD and IEEE 802.15.3c compliant products
commercially available ranging from external HDMI adapters to built-in solutions on HD displays and high-end
laptops. These consumer electronics devices are implemented with ASIC because it significantly reduces costs
per unit, significantly saves power, is smaller and works in higher speeds compared to FPGA. However, the
costs involved in creating an ASIC are very high and are only economically viable once the devices starts to be
massively produced. For prototyping, FPGA technology offers more flexibility to test, analyze and correct the
hardware implementation. A vast majority of designs intended for ASICs are originally prototyped in an FPGA.
Prototyping before definitive specifications creates conditions to reduce time-to-market.
Even though developing hardware using FPGA technology is faster than developing using ASIC, it is still a
time consuming process and a lot of development has to be done before a full system test can be executed.
System modelling and simulation can help reduce the development phase. By using already developed blocks
and high level programming languages the development time of the hardware prototype can be greatly reduced
by enabling an early detection of design problems, helping determining block specifications, producing and
validating system wide tests as well as individual block tests.
2
1.2 Objectives
Our goal is to model and simulate an OFDM base band transceiver unit capable of achieving throughputs
high enough to transmit uncompressed HD video. The model will aide the future development of a hardware
prototype using FPGA technology by generating test vectors for each transceiver block and validating the
results.
To achieve the main objective, specific objectives are defined:
- To analyse the available standards for wireless uncompressed HD video transmission.
- To choose an appropriate standard to use.
- To model the transmitter, receiver and channel.
- To define the numerical representation of the complex coefficients generated by the system mapper and
IFFT blocks.
- It will also be used to determine the specifications of the hardware blocks, specially the individual blocks
in which the implementation details are not addressed by the respective standard, e.g., the required
numeric precision for the IFFT and FFT blocks.
To achieve the main goal we defined several intermediate goals: determining the requirements for
transmitting uncompressed HD video and selecting a suited specification for the task.
1.3 Dissertation outline
The preliminary study conducted to determine the bandwidth requirements for wireless transmission of
uncompressed HD video and audio, the current capabilities and applications of wired HD video/audio interfaces,
and the current technical solutions for wireless transmission of video/audio is presented on Chapter 2.
The selected specification is described in detail in Chapter 3 and in the Matlab/Simulink transceiver model is
presented on Chapter 4. The results and conclusions are presented in Chapters 5 and 6 respectively.
3
Chapter 2
Wireless HD video transmission This chapter identifies the requirements for transmitting uncompressed video/audio, analyses the existing
wired and wireless specifications for transmitting HD video, and selects a specification to be used in this work.
2.1 Digital video and audio signals
Characterization of digital video and audio signals is important to assess the net data rate requirements. A
digital video signal data rate is defined by: resolution, i.e., the total number of pixels of each image, normally
referred as number of horizontal pixels by vertical pixels on the screen; colour depth, i.e., the number of bits
used to represent each of the three colours of a pixel; refresh rate, i.e., number of times per second the image is
completely reconstructed on the screen; progressive or interlaced formats, i.e., the way lines of an image are
displayed in the refreshing cycles, progressive formats display all the lines on all the refresh cycles, but
interleaved ones display even and the odd lines in alternated refresh cycles.
Currently available HD video formats are referred as “720i”, “720p”, “1080i”, and “1080p”1. These terms
indicate the number of lines and the display method2 used. Images used in HD video formats have a 16:93
aspect ratio, resulting in wider images than the conventional 4:3 aspect ratio used in Standard Definition (SD)
video. The number of pixels, np, in each HD image can be calculated by (1) where nl is the number of lines,
indicated by the video format designation:
np = (16/9) × nl2 (1)
From (1), it is possible to calculate that 720p and 720i images are formed by 1280×720 pixels; 1080i and
1080p images are formed by 1920×1080 pixels.
The bit rate required to transmit “Full HD” video, vbr, with progressive display can be calculated from (2),
where np is the number of pixels, ncchannels is the number of colour channels, cdepth is the number of bits
used to represent each colour and rfreq is the display refresh frequency. Video signals also contain audio
information and digital audio data is defined by: number of audio channels; sampling rate; number of bits used to
quantify each audio sample. The audio bit rate, abr, can be calculated from (3), where nac is the number of
audio channels used, srate is the sampling rate and sdepth is the number of bits used per sample.
vbr = np × ncchannels × cdepth × rfreq (2)
abr = nac × srate × sdepth (3)
1 1080p is also referred as “Full HD”. 2 “i” for interlaced and “p” for progressive display. 3 (Image length : image width)
4
Using (1), (2) and (3) is possible to calculate the required bit rate to transmit an uncompressed video and
audio signal. The net bit rate for Full HD video and audio is shown in (4) and (5).
vbrFullHD = 1920 ×1080 × 3 × 8 × 60 = 2.99 Gbps (4)
abrFullHD = 8 x 192 k x 24 = 36.8 Mbps (5)
2.2 Wired HD video standards
One of the most widespread HD interfaces is HDMI (Home Digital Multimedia Interface). It was introduced in
December 2002 and was designed to transmit uncompressed video, up to 8 channels of audio and control data.
HDMI is electrically compatible with the DVI (Digital Visual Interface) [24], introduced in 1999 as a digital
alternative to the analogue VGA (Video Graphics Array) interface. The latest version, HDMI 1.4 [1], supports
video resolutions up to 4096x2160 at 24 Hz or 3840x2160 at 30Hz with 12 bits for each colour channel, requiring
a maximum bit rate of approximately 7.64 Gbps and 8.96 Gbps respectively.
Display Port is a HD video and audio interface presented in 2006 by Video Electronics Standards Association
(VESA). In the latest specification it is capable of delivering digital video up to resolutions of 3840x2160 at 30 Hz
with 10 bit for each colour channel and 8 audio channels at 192 kHz with 24 bit samples [2]. Display Port version
1.2 supports resolutions up to 2560×1600 at 120Hz4 and 3840×2160 at 60 Hz with 10 bits for each colour
channel. These new video formats demand a bit rate of up to 14.9 Gbps.
2.3 Wireless video transmission technologies
In this section several wireless video transmission technologies will be presented. The technologies will be
divided into two categories: technologies using frequencies up to 10 GHz and technologies in the 60 GHz band.
2.3.1 Technologies using frequencies up to 10 GHz
Due to the characteristics of the frequency bands used, technologies using frequencies below 10 GHz
normally have greater range than 60 GHz band technologies and can be used across multiple rooms. However
the available bandwidth is narrower and therefore the maximum throughput is inferior. Over the next sub-
sections several technologies using frequencies up to 10 GHz will be presented
2.3.1.1 WirelessUSB (WUSB)
WUSB [9] is the wireless counterpart of the well known USB standard. The WUSB physical layer [20] based
on the WiMedia Alliance MB-OFDM specification [10], permits data rates ranging from 53.3 Mbps to 480 Mbps5.
There are some commercially available products that use WUSB to implement wireless High Definition
streaming transmission from a computer to a HD display. WUSB specifications are available but the existing
adaptation to transmit video and audio signals over WUSB are not public. Available information suggests
4 120 Hz refresh rates are used for 3D formats. 5 Up to 3 meters
5
compression is used and a fairly high-performing CPU6 is required.
2.3.1.2 Wireless Home Digital Interface (WHDI)
The WHDI specification, finalized in December 2009 is only available to WHDI Consortium members. This
technology is able of transmitting uncompressed HD video and achieves equivalent video data rates up to 3
Gbps.
The attenuation in the 5 GHz band is smaller than the attenuation in the 60 GHz band, therefore WHDI
ranges can go up to approximately 30 meters and transmission through walls and other obstacles is possible,
enabling the creation of networks spanning several rooms.
As of this date, there are some WHDI products commercially available like set-top boxes.
2.3.1.3 WirelessDisplay (WiDi)
WiDi is a proprietary format from Intel that enables Intel based computers7 to stream a computer's display
and audio to a set-top box connected to a high-definition display. The first version of the technology,
implemented on top of Wi-Fi (2.4 GHz), does not support 1080p or 1080i format8 [12] and Version 1.0 set-top
box's were produced by only one manufacturer. Version 2.0 presented at CES 2011 introduced additional
features like support for 1080p format.
2.3.1.4 IEEE 802.11ac
The Task Group ac (TGac) is the result of the work done by the 802.11 Very High Throughput (VHT) study
group between May and November 2008. The purpose of TGac is to produce modifications to the 802.11 PHY
and MAC layers to enable modes of operation that will allow throughputs of at least 1 Gbps measured at the
MAC layer. The 802.11ac amendment will enable operation below the 6 GHz, excluding the 2.4 GHz band, and
ensure backward compatibility and coexistence with legacy devices in the 5GHz band. The expected date for the
finalized amendments is December 2012.
According to the draft specification [6], IEEE 802.11ac devices can use 20, 40, 80 or 160 MHz channels, up
to seven spatial streams (MIMO), and will employ OFDM modulation. Each of the OFDM sub-carriers can be
modulated with four different modulations: BPSK, QPSK, 16-QAM, 64-QAM, and 256-QAM. The different
combination of channel bandwidth, sub-carrier modulations and number of spatial streams used originate
several throughputs ranging from 6.5 Mbps to approx. 6.9 Gbps9.
6 http://eu.veebeam.com/products/Veebeam-HD.html. 7 Equipped with 2010 generation CPUs and Intel Wi-Fi network adapters.
8 The supported resolutions, for version 1.0, are 1280 x 800, 1280 x 768, 1280 x 720, @ 30 frames per second. 9 Achieved with 160 MHz channel, 7 spatial streams and 256-QAM modulation.
6
2.3.1.5 Under 10 GHz technology summary
Table 1 presents the maximum data rates and frequency band of the technologies presented in the previous
sections. At the present moment WHDI has the highest date, 3 Gbps, and is, according to its promoters, able to
transmit uncompressed HD video. Future data rates in the frequency band below 6 GHz will reach approximately
7 Gbps with the completion of IEEE 802.11ac specification.
Table 1: Technologies up to 10 GHz
WUSB WHDI WiDi IEEE 802.11ac
Data rate (Mbps) 480 Approx. 300010 Not declared Approx. 7000
Frequency Band 3.1 to 10.6 5 2.4 < 6
2.3.2 Technologies in the 60 GHz band
The 60 GHz unlicensed band has 9 GHz of bandwidth available in most world regions, making it the choice
for implementing technologies able of multi-gigabit throughputs. ECMA-387, IEEE 802.15.3c and IEEE
802.11ad, still in development, use this frequency band to implement very high throughput WPANs. Industry
consortia like WirelessHD Consortium and Wireless Gigabit Alliance have also produced specifications to use
this band.
Although offering very high bandwidth, the use of the 60 GHz band poses several challenges. First of which is
the high propagation attenuation that occurs in this frequency band, limiting the technologies to short range,
typically up to 10 meters, and to in-room transmissions. The other big challenge is the use of highly directional
antennas, needed to overcome the propagation attenuation, which make the technologies very sensitive to line
of sight (LOS) obstacles and thus the need to introduce transmission methods able to use non line of sight
(NLOS) channels.
2.3.2.1 IEEE 802.15.3c
Working Group 802.15 is specialized in wireless personal area networks (WPANs) standards, networks of
interconnected devices with a typical diameter inferior to ten meters. According to [13], in 2003 WG 802.15
published the Standard 802.15.3-2003 defining the Physical Layer (PHY) and Medium Access Control (MAC) for
high throughput WPAN in the 2.5 GHz frequency band. In 2004, Study Group 3c (SG3c) started analysing the
possibility of using the 60 GHz band for very high throughput WPANs. SG3c was elevated to Task Group 3c
(TG3c) in 2005 and was commissioned to introduce modifications to the 802.15.3-2003 specification in order to
use the unlicensed 60 GHz band. IEEE 802.15.3c standard [2] was published in 2009.
2.3.2.1.1 Specification overview
10 Equivalent video data rate.
7
PHY layer defines four channels in the frequency band from 57.0 GHz to 66.0 GHz. The available bandwidth
was divided into 2.16 GHz channels as shown in Table 2. To accommodate the regulatory differences in some
world regions, the requirements allow the usage of fewer channels.
Table 2: 802.15.3c frequency channels [2]
Channel ID Start frequency Centre frequency Stop frequency
1 57.240 GHz 58.320 GHz 59.400 GHz
2 59.400 GHz 60.480 GHz 61.560 GHz
3 61.560 GHz 62.64 GHz 0 63.720 GHz
4 63.720 GHz 64.800 GHz 65.880 GHz
Three working modes are defined, allowing different applications and different device types: the Single
Carrier mode (SC PHY), designed for short range communication; the High Speed Interface (HSI PHY) primarily
designed for low-latency, bidirectional data transfers; and the Audio/Visual mode (AV PHY) specifically designed
for transmitting HD video and audio.
2.3.2.1.1.1 SC mode
Three modulation and coding schemes (MCS) classes are defined addressing different applications: Class 1
devices are low-power/low-cost mobile devices capable of achieving data rates up to 1.5 Gbps; Class 2 devices
are intermediate performance devices whose data rates vary from 1.5 to 3 Gbps; and Class 3 devices designed
for high performance with data rates topping at approximately 5 Gbps.
The data rates for the three classes in SC PHY are presented in Table 22. There are 14 different MCS
resulting from the combination of pilot word lengths, carrier modulation and forward error correction (FEC) used.
The carrier modulations defined for SC PHY are Binary Phase Shift-Keying (BPSK), Gaussian Minimum Shift-
Keying, Quadrature Phase Shift-Keying, 8-Phase Shift-Keying and 16-Quadrature Amplitude Modulation11. The
forward error correction codes used are the Reed-Solomon error correction code (RS) and the Low-Density
Parity-Check code (LDPC).
2.3.2.1.1.2 HSI mode
The HSI PHY uses OFDM and was designed for devices that require low-latency, full-duplex high speed
connections. Eleven MCS are defined, presented in Table 23 and Table 24 with data rates ranging from 32.1
Mbps to 5.775 Gbps. The FEC code used in all MCS is LDPC with bit rates from 1/2 to 7/8. MCS index 0 to 7
apply equal error protection (EEP) to all the bits, while MCS indexes 8 to 11 apply unequal error protection
(UEP) to the most significant bits (MSB) and less significant bits (LSB).
The OFDM timing parameters for the HSI PHY are presented in Table 25. It is defined that HSI PHY occupies
11 All these modulations are shifted by Π/2.
8
a 1.815 GHz bandwidth and the reference sampling frequency is 2.640 GHz. The OFDM
modulation/demodulation is performed by a 512 point IFFT/FFT12.
2.3.2.1.1.3 Audio/Visual mode
This mode is specifically designed for transmitting uncompressed audio and video content. There are two
modes defined, the High-Rate PHY (HRP) and the Low-Rate PHY (LRP), both using OFDM and coordinated in a
TDMA manner by the MAC layer because these modes operate in overlapping frequency bands (Table 26 and
Table 2). The former is highly directional and used for streaming video contents, while the later is omni
directional and is used to set up high rate channels. The LRP and HRP maximum data rates are 10.2 Mbps
(Table 27) and 3.8 Gbps respectively (Table 7).
2.3.2.1.2 WirelessHD (WiHD)
WirelessHD [4] is very closely related to the AV mode of the 802.15.3c-2009 Standard. In fact they both use
OFDM modulation on the high-rate PHY (HRP) and low-rate PHY (LRP) and share the same frequency band
plan, base band and general parameters.
Devices using WiHD form a Wireless Video Area Network (WVAN), consisting of one Coordinator and zero or
more Stations. In most cases the Coordinator will be a device that is a sink for audio or video data, like a display
or a Personal Video Recorder (PVR). A Station is a device that can source and/or sink audio and video data,
simultaneously if needed. The Coordinator can also act as a Station.
Currently there are some CE products available on the market like high end HD TVs and WiHD set-top boxes
(transmitter/receiver). A second generation WirelessHD specification [4] will include data rates up to 28Gbs;
support for 3D and 4k resolutions; backward compatibility with WirelessHD 1.0 devices; and High Definition
Content Protection 2.0 (HDCP) 2.0 over WirelessHD.
2.3.2.2 ECMA-387
ECMA-387 [5] is a standard addressing WPANs in the 60 GHz band. It was finished in December 2008 and
was adopted as an ISO/IEC 13156 [14] standard in late 2009.
2.3.2.2.1 Specification overview
ECMA-387 defines working modes that can use Single Carrier Block Transmission (SBCT) or OFDM
modulation. For transmission modes using SCBT, the available bandwidth is divided into four 2.16 GHz
channels. Channels can be aggregated to increment the maximum data rate by a factor of two, three or four13.
Table 3: ECMA-387 frequency usage and channel bonding [5] shows the frequency usage of single and
aggregated channels, where channel bonding represents the aggregation.
12 Timing constraints are the same for both FFT and IFFT blocks. 13 Unlike the other 60 GHz technologies presented, IEEE 802.15.3c and the draft version of IEEE 802.11ad.
9
Table 3: ECMA-387 frequency usage and channel bonding [5]
Band ID Channel Bonding Lower Frequency
(GHz)
Upper
Frequency
(GHz)
1 no 57.240 59.400
2 no 59.400 61.560
3 no 61.560 63.720
4 no 63.720 65.880
5 Channels 1 & 2 57.240 61.560
6 Channels 2 & 3 59.400 63.730
7 Channels 3 & 4 61.560 65.880
8 Channels 1, 2 & 3 57.240 63.720
9 Channels 2, 3 & 4 61.560 65.880
10 Channels 1, 2, 3 & 4 57.240 65.880
ECMA-387 specifies two types of devices: Type A devices, which use LOS and NLOS channels and have
trainable antennas and perform more complex base band digital signal processing like equalization; and Type B
devices, which do not have trainable antennas, only use the LOS channel and perform minimal base band digital
signal processing. ECMA-387 networks can be formed by any combination of Type A and Type B devices.
Twenty two working modes are defined for type A devices (Table 29), of which fourteen modes use SCBT, and
eight use OFDM. For type B devices, there are four modes all using SCBT (Table 30).
Table 29 shows the data rates, carrier modulation and FEC codes for all modes defined for Type A devices. SCBT modes can use channel bonding to increase the maximum throughput. The FEC encoding is performed by Reed-Solomon algorithm and convolutional codes; some modes use Trellis Coded Modulation instead of convolutional coding. Several modes perform unequal error protection for MSBs and LSBs.
Table 30 shows data rates, carrier modulation and FEC coding used for all Type B device modes. All modes
use Differential Binary Phase-Shifting Keying, except mode B3 which uses QPSK and unequal error protection
To ensure compatibility between both device types, all Type A devices most support, at least, modes A0 and
B0, and Type B devices most support, at least mode B0.
2.3.2.3 IEEE 802.11.ad
The Task Group TGad is the result of the work done by the 802.11 Very High Throughput (VHT) study group
between May and November 2008 for the 60 GHz band. The 802.11ad amendment enables operation in the
10
60GHz band and coexistence with other systems in this frequency band, e.g. 802.15.3c systems. The expected
date for the finalized amendments is December 2012.
2.3.2.3.1 Specification Overview
According to [7], available bandwidth is divided in four channels of 2.160 GHz as shown in Table 1, and IEEE
802.11ad will support OFDM and SC modulations. Table 31 and Table 32 show the MCS modes for the SC and
OFDM modulations respectively. Both modes use LDPC as the FEC code. SC modes allow data rates up to
approximately 4.6 Gbps, while OFDM modes allow a maximum rate of approximately 6.8 Gbps.
The timing parameters for the OFDM modulation are listed in Table 4.
Table 4: 802.11ad OFDM modulation timing parameters [7]
Parameter Value
Number of data sub-carriers (NSP) 336
Number of pilot sub-carriers (NSP) 16
Number of DC sub-carriers (NDC) 3
Total Number of sub-carriers (NST) 355
Number of sub-carriers occupying half of the overall BW (NSR) 177
Sub-carrier frequency spacing 5.15625 MHz
OFDM sample rate (Fs) 2640 MHz
OFDM Sample Time (Ts) 0.38 ns=1/Fs
IDFT/DFT period (TDFT) 0.194 µsec
Guard Interval duration (TGI) 48.4 ns= TDFT/4
Symbol Interval (TSYM) 0.242µs= TDFT+TGI
2.3.2.3.2 Wireless Gigabit (WiGig)
WiGig [8] is a specification developed by the Wireless Gigabit Alliance. It was based on the IEEE802.11
standard and operates on the 60GHz band. WiGig devices equipped with multi-band radios will be able to switch
seamlessly to other versions of 802.11 standard like, for example, 802.11ac, 802.11n or 802.11g.
Currently, the Wireless Gigabit Alliance has established a cooperation agreement with the Wi-Fi Alliance to
share specifications in order to develop Wi-Fi certified products in the 60 GHz band. The Wireless Gigabit
Alliance has also announced reaching an agreement with Video Electronics Standards Association (VESA) to
define a standard for wireless display technology.
According to [29], this specification was officially selected as the basis for the 802.11ad draft in May 2010.
The target applications are: fast wireless sync; streaming HD over HDMI or DisplayPort; wireless computing;
Internet access, using native Wi-Fi support.
The specification addresses the 60 GHz band, allowing: data rates up to 7 Gbps; native 802.11 support and
11
transparent switching between current 802.11 networks (2.4 GHz, 5 GHz) and 60 GHz networks; wireless
implementations of HDMI, DisplayPort, USB and PCIe; transmission of compressed and uncompressed video.
As 802.11.ad, the WiGig specification defines four 2.16 GHz channels corresponding to (band ID 1 to 4 in
Table 2) can utilize OFDM or SC modulation and should deliver approx. 7 Gbps of maximum throughput.
2.3.2.4 60 GHz technology summary
Table 5 presents a summary of the 60 GHz technologies presented on the section. WiGig is not included in
Table 5 because of the small amount of information available on this technology. It is interesting to note that all
technologies, except WiHD, have single carrier modes. In spite of this, only ECMA-387 requires all devices to
support single carrier modes14.
ECMA-397 is the only technology with the capability of aggregating frequency channels enabling it to reach
data rates in excess of 25 Gbps.
Table 5: Comparison of the 60 GHz technologies
802.15.3c WiHD ECMA-387 802.11.ad
(draft [7])
AV mode
SC mode HSI
mode LRP HRP
LRP HRP SCBT OFDM SC mode OFDM
Modulation Single
Carrier OFDM OFMD OFDM OFDM OFDM
Single
Carrier OFDM
Single
Carrier OFDM
Maximum Data
Rate 5.280 Gbps
5.775
Gbps
10.2
Mbps
3.807
Gbps
10.2
Mbps
3.807
Gbps
6.350
Gbps15
4.032
Gbps 4.62 Gbps 6.756 Gbps
Sampling Rate
(samples per
second)
- 2.640 G 2.538 M 2.538 G 2.538 M 2.538 G 1.728 G 2.592 G - 2.640 G
FFT time (ns) N.A. 193.94 403.47 201.73 403.47 201.73 N.A. 197.53 ns N.A. 19416
FFT
sub-carriers
total (data)
N.A. 512
(336)
128
(30)
512
(336)
128
(30)
512
(336) N.A.
512
(360) N.A.
512
(336)
2.3.3 Conclusions on wireless HD video transmission
High Definition video and audio require high bit rates. As shown in Section 2.1, the minimum data rate for
transmitting 1080p uncompressed video and audio is approximately 3 Gbps. Newer video formats, like HD 3D
and 4k resolutions, require even higher dates, close to 15 Gbps.
14 Mode A0 for Type A and mode B0 for Type B devices.
15 Over 25 Gbps with channel bonding
16 The draft specification mentions DFT instead of FFT.
12
Current wireless technologies in the 2.4-10 GHz band are unable of accommodating such high throughput.
The only exception is WHDI, a proprietary specification which is able of 3 Gbps equivalent video data rates.
Future version of the IEEE 802.11 standard, IEEE 802.11ac, will reach data rates of approximately 7 Gbps, using
MIMO technology with 7 spatial streams and higher order carrier modulation constellations, but will also be
incapable of supporting the newer video formats.
The 60 GHz band offers a large bandwidth, almost 7 GHz, in several world regions. This allows the
development of technologies with multi-gigabit data rates but poses several design challenges due to the high
propagation attenuation in these frequencies: lower range, up to 10 meters, and highly directional channels, due
the necessity to employ high gain directive antennas.
Table 6 compares all the studied technologies in the following criteria:
Maximum data rate;
Standard body;
Availability of the specification;
Existence of transmission modes designed for video.
Table 6: Comparison criteria
WUSB WHDI WiDi 802.11ac
(draft) 802.15.3c WiHD ECMA-387
802.11ad
(draft)
Data Rate
(Gbps) 0.48 3
Not
defined 6.9
5.28 (SC)
5.39 (OFDM) 3.817
6.35/ 25 (SCBT)
4.032 (OFDM)
4.63 (SC)
6.8 (OFDM)
International
Standard No No No
Yes
(after Dec.
2012)
Yes No Yes Yes
(after Dec. 2012)
Specifications
available No18 No No Yes (draft) Yes No19 Yes Yes (draft)
Specific video
modes Yes Yes Yes No Yes Yes Yes Yes
As mentioned before, with the exception of WHDI and the upcoming IEEE 802.11ac standard, 2.4-10 GHz
band technologies do not support uncompressed HD video transmission. Current WHDI version and upcoming
IEEE 802.11ac specification are not able to support newer HD video formats. Even tough WiHD 1.0 is
compatible with IEEE 802.15.3c-2009 and WiGig was used as the starting point for IEEE.802.11ad amendment,
only ECMA-387, IEEE 802.15.3c are international standards and have available specifications; IEEE 802.11ad is
expected to become an international standard in December 2012. All 60 GHz technologies are able to achieve
data rates higher than 3 Gbps, and have specially designed video modes and therefore are adequate for
17 28 Gbps in next version 18 WUSB specification is available. 19 Only an overview of the first version is available. The full document is only accessible to members.
13
transmitting uncompressed HD video. Presently only ECMA-387, by the use of bonded channels, can support
data rates up to approximately 25 Gbps. WiHD next version is expected to achieve throughputs up to 28 Gbps
but little more information is available about it.
Taking into consideration the mentioned criteria, ECMA-387 is the more appropriate standard on which to
base the implementation of a wireless uncompressed HD video transmission system. However, if the
implemented system is to be based on OFDM modulation, then ECMA-387 would no longer be the most suitable
standard because it requires that all devices support a SCBT mode resulting in additional required hardware and
the maximum throughput of the OFDM modes is similar to IEEC 802.15.3 HSI and AV HRP modes and inferior to
IEEE 802.11ad OFDM modes. In this context IEEE 802.15.3c AV mode is the best alternative because the
specification is already published and is specifically designed for transmission of HD video; the IEE802.11ad
specification was discarded because it is still being developed.
14
Chapter 3
IEEE 802.15.3c AV mode As mentioned before the HRP mode was designed for uncompressed HD video transmission. Table 7 shows
the data rates and coding for HRP mode. Modes 0 to 2 use EEP on both MSB and LSB; modes 3 and 4 use
UEP to provide different protection levels to the MSB and LSB, thus providing a more appropriate protection for
streaming video contents20.
Table 7: HRP data rates and coding [3]
Code rate HRP mode
index
Coding
mode Modulation
MSB 4b LSB 4b
Raw data
rate (Gbps/s)
0 QPSK 1/3 0.952
1 QPSK 2/3 1.904
2
EEP
16-QAM 2/3 3.807
3 QPSK 4/7 4/5 1.904
4 UEP
16-QAM 4/7 4/5 3.807
5 QPSK 1/3 N.A. 0.952
6
MSB-only
Retransmission QPSK 2/3 N.A. 1.904
As in the HSI PHY, the OFDM modulation is performed by a 512 point FFT. Due to the lower bandwidth of the
signal, 1.76 GHz, the AV HRP mode has lower timing constraints than HSI PHY. The HRP timing parameters are
summarized in Table 8.
Each HRP channel can include three 3 LRP channels, but only one can be used at a time. Table 26 presents
the LRP frequency plan referenced to the HRP channels (Table 2).
All the LRP modes use BPSK modulation and the data rates vary from 2.5 Mbps to 10.2 Mbps as showed in
Table 27. The omni directional LRP link is created by repeating the OFDM symbol and its associated cyclic prefix
in eight (LRP modes 0 to 2) or four (LRP mode 3) spatial directions. The LRP OFDM timing requirements are
considerably lower than the timing requirements of the HRP mode and are presented in Table 28.
20 As mentioned on section 2, each pixel is constituted by 3 colour channels and in the case of Full HD each colour channel
sample is coded with 8 bits, so an error on the 4 MSB will cause a bigger impact on the image than an error on the 4 LSB.
15
Table 8: HRP OFDM timing parameters [3]
Parameter Value
Occupied bandwidth 1.76GHz
Reference Sampling (fs(HR)) 2.538 Gsamples/s
Number of sub-carriers (Nsc(HR)) 512
FFT period (TFFT(HR)) Nsc(HR)/ fs(HR) ~ 201.73 ns
Sub-carrier spacing (Δfsc(HR)) 1/ TFFT(HR) ~ 4.957 MHz
Guard interval (TGI(HR)) 64/ fsc(HR) ~ 25.22 ns
Symbol duration (TS(HR)) TFFT(HR) + TGI(HR) ~ 226.95 ns
Number of data sub-carriers
(Ndsc(HR)) 336
The focus will be on the HRP mode because it’s the actual mode used to transmit video and has more
demanding and challenging requirements than the LRP. The next sections are dedicated to presenting the
structure of the HRP frame and the most important characteristics of the HRP mode.
3.1 HRP PHY frame structure
Each MAC frame received at the HRP PHY layer is converted into a HRP data unit (HRPDU). The HRP PHY
layer adds the HRP preamble, the HRP header and the header check sequence (HCS) to the Extended MAC
header. The MAC frame body is sent as HRPDU payload. The HRPDU format is shown in Figure 2. All the
HRPDU fields are described in detail in Annex D and [3].
MAC Frame Body HCS Extended MAC
Header
HRP Header HRP Preamble
Last in Time First in time
Figure 2: HRPDU frame format
The HRP preamble, HRP header, MAC header and HCS fields are transmitted using HRP mode 0. The HRP
payload is divided into up to seven blocks, called HRP sub-frames. Each sub-frame can use a different HRP
mode and has a maximum length of 220 (1.048.576) octets, allowing a maximum HRPDU frame size of
approximately 7MB.
16
3.2 Splitting
Current HD video formats, like 720p and 1080p, use 8 bit representation for each colour channel. The MSBs
in each colour channel are more influential to the displayed pixel colour than the LSBs. All the HRPDU frame bits
are grouped into groups of 8 bits, referred to as octets in [2], and the 4 MSBs are separated from the 4 LSBs and
independently encoded. This design allows introducing higher forward error protection to the MSBs.
The HRP mode splits the octets differently depending whether UEP modes are used or not. Figure 3 and
Figure 4 show the splitting pattern for EEP and UEP modes. The pattern period is eight octets and the MSBs are
always mapped to the upper bit stream and the LSBs always mapped to the lower bit stream. The analytical
expressions for the splitting patterns are presented in (6) through (9), where in(i,b) is the input array of octets; i is
the index of the input octet and is defined as i=0,1,...,N-1; N is the number of octets in the input stream; b=0 is
the LSB and b=7 is the MSB in each octet; and n is the index of the output octet.
The output arrays generated by the HRP modes 0, 1 and 2 are defined by (6) and (7):
lower(n, b) = in({floor[b/4] + floor[n/2] × 4}, {mod[b, 4] + mod[n, 2] × 4}) (6)
upper(n, b) = in({floor[b/4] + floor[n/2] × 4 + 2}, {mod[b, 4] + mod[n, 2] × 4}) (7)
The output arrays generated by the HRP modes 3 and 4 are defined by (8) and (9):
lower(n, b) = in({floor[b/4] + n × 2}, mod[b, 4]) (8)
upper(n, b) = in({floor[b/4] + n × 2}, {mod[b, 4] + 4) (9)
The modulo function, mod (x,y), in (6), (7), (8) and (9) is defined in (10), where n is the closest integer smaller
than or equal to n/y:
mod(x,y) = x – n × y (10)
Figure 3: EEP splitting
17
Figure 4: UEP splitting
3.3 Forward error correction
The HRP mode includes forward error protection methods enabling the receiver to correct errors introduced
in the communication by the noise in the transmission channel. Two different methods are used in the HRP
mode: block encoding and convolutional encoding.
3.3.1 Block encoding
Reed-Solomon, referred as the outer code in [3], is a block encoding algorithm, which works by dividing the
message in several blocks and adding parity protection information to each block. A Reed-Solomon code is
characterized by the length of the code word, n, original length of the message block, k, the number of bits in
each symbol, m, and the Galois field generator polynomial, gfield. The Reed-Solomon decoder algorithm can
recover (n-k)/2 incorrect symbols in each codeword, even if all the bits in the symbols are corrupted. It is
common to use the abbreviated notation RS (n,k) to refer to a Reed-Solomon code.
For the AV HRP mode, the Galois field generator polynomial is given by (11) and each symbol has eight bits.
gfield(x)=x8+x4+x3+x2+1 (11)
Each codeword is generated by appending to the message the remainder, r(x), of the multiplication of the
message, by xn-k and divided by gfield(x), as expressed in (12), represented as a polynomial by using the nth
symbol as the coefficient for the xn factor. Where r(x) is the remainder of (12):
M (x )× xn− k
gfield (x ) (12)
3.3.2 Convolutional encoding
The convolutional code is applied to individual bits. The code used is non-recursive and has a code rate of
1/3. It is defined by the polynomial generators; g0=133o (1011011b); g1=171o (1111001b) and g2=165o
(1110101b).
18
3.3.2.1 Variable code rates
Puncturing is employed to modify the code rates allowing the system to adjust to the noise characteristics of
the transmission channel. The defined code rates are presented in Table 9.
Table 9: Puncturing pattern
Code
Rate
Puncturing
Pattern
Transmitted
sequence
1/3
X:1
Y:1
Z:1
X1Y1Z1
1/2
X:1
Y:1
Z: 0
X1Y1
4/7
X:1111
Y:1011
Z:0000
X1Y1X2X3Y3X4Y4
2/3
X:11
Y:10
Z:00
X1Y1X2
4/5
X:1111
Y:1000
Z:0000
X1Y1X2X3X4
3.4 Interleavers
To increase the performance of the forward error correcting methods, by reducing the occurrence of errors
bursts, several interleaving schemes are used throughout the transmission chain.
3.4.1 Outer interleaver
The symbols in the code words generated by the Reed-Solomon encoder are interleaved before being
distributed to the convolutional encoders. The distribution order is defined in (13), where b represents the
encoded octets stream; i takes values from 0 to depth-121; k varies from 0 to N-1; N is the length of the Reed-
Solomon code word, M is the number of parallel convolutional encoders used by the HRP mode; and m is the
index of the convolutional encoder that receives the octet: 21 depth is the number of Reed-Solomon codewords that are interleaved by the outer interleaver.
19
b(i,k+M+m) is delivered to the mth convolutional encoder (13)
In practice (13) results in two different patterns because the interleaving depth used for the HRPDU header is
2 and the overall RS code word length is 56 symbols and the interleaving depth for the HRPDU payload is 4 and
the RS code word length is 224. The two patterns are represented in Figure 5 and Figure 6.
Figure 5: Outer interleaver pattern for depth=2, N=56
Figure 6: Outer interleaver pattern for depth=4, N=224
3.4.2 Bit interleaver
The bit streams produced by the convolutional encoders are interleaved according to Table 10, for HRP
header bits, and Table 12 for HRP payload bits.
Table 10: HRPDU header bit interleaver
Before After Before After Before After Before After Before After Before After
0 0 8 16 16 32 24 1 32 17 40 33
1 2 9 18 17 34 25 3 33 19 41 35
2 4 10 20 18 36 26 5 34 21 42 37
3 6 11 22 19 38 27 7 35 23 43 39
20
Table 11: HRPDU header bit interleaver (cont.)
Before After Before After Before After Before After Before After Before After
4 8 12 24 20 40 28 9 36 25 44 41
5 10 13 26 21 42 29 11 37 27 45 43
6 12 14 28 22 44 30 13 38 29 46 45
7 14 15 30 23 46 31 15 39 31 47 47
3.4.3 Tone interleaver
Before the modulated sub carriers are processed by the IFFT block all the sub carriers are shuffled. The tone
interleaving operation reverses the binary representation of each complex coefficient index. Table 13 presents
some examples of the interleaving done to the sub-carriers.
Table 12: HRPDU data bit interleaver
Before After Before After Before After Before After Before After Before After
0 0 8 44 16 47 24 14 32 10 40 25
1 28 9 36 17 39 25 6 33 2 41 5
2 20 10 16 18 31 26 34 34 42 42 45
3 12 11 8 19 11 27 26 35 22 43 37
4 40 12 7 20 3 28 18 36 21 44 17
5 32 13 35 21 43 29 46 37 13 45 9
6 24 14 27 22 23 30 38 38 41 46 1
7 4 15 19 23 15 31 30 39 33 47 29
Table 13: Tone interleaving table
Sub-carrier
Original index
Sub-carrier
Interleaved index
Binary Decimal Binary Decimal
000000000 0 000000000 0
000000001 1 100000000 256
000000010 2 010000000 128
… … … …
111111111 511 111111111 511
3.5 Sub carrier modulation
QPSK and 16QAM modulations are used on the HRP mode. The complex coefficients are represented by
(14).
21
(d1+jd2) × KMOD (14)
The relation between d1 and d2 depends on the error protection mode applied, EEP or UEP, and the value of
KMOD depends on the modulation used.
For QPSK modulation the serial bits are aggregated into groups of two bits, and KMOD is defined by (15).
KMOD =1/( 22d+d 2
1) (15)
For EEP d1=d2 and for UEP d1=1.25 d2. Table 14 shows the correspondence between the two bit
combinations and the value of the real and complex parts of the associated complex number.
For 16-QAM modulation the serial bits are aggregated into groups of four bits and KMOD is defined by (16).
KMOD= 1/ 22d+d 2
15 (16)
For EEP d1=d2 and for UEP d1=1.25 d2. Table 15 shows the relationship between the four bit combinations
and the value of the real and complex parts of the associated complex number.
Table 14: Real and imaginary values for QPSK modulation
QPSK EEP UEP
b0 b1 Real Img. Real Img.
00 -0.707 -0.707 -0.781 -0.625
01 0.707 -0.707 0.781 -0.625
10 -0.707 0.707 -0.781 0.625
11 0.707 0.707 0.781 0.625
Table 15: Real and imaginary values for 16-QAM modulation
16-QAM EEP UEP
b0 b1 b2 b3 Real Img. Real Img.
0000 -0.948 -0.948 -1.048 -0.838
0001 -0.948 -0.316 -1.048 -0.279
0010 -0.948 0.948 -1.048 0.838
0011 -0.948 0.316 -1.048 0.279
22
Table 16: Real and imaginary values for 16-QAM modulation (cont.)
16-QAM EEP UEP
b0 b1 b2 b3 Real Img. Real Img.
0100 -0.316 -0.948 -0.349 -0.838
0101 -0.316 -0.316 -0.349 -0.279
0110 -0.316 0.948 -0.349 0.838
0111 -0.316 0.316 -0.349 0.279
1000 0.948 -0.948 1.048 -0.838
1001 0.948 -0.316 1.048 -0.279
1010 0.948 0.948 1.048 0.838
1011 0.948 0.316 1.048 0.279
1100 0.316 -0.948 0.349 -0.838
1101 0.316 -0.316 0.349 0.279
1110 0.316 0.948 0.349 0.838
1111 0.316 0.316 0.349 0.279
23
Chapter 4
IEEE 802.15.3c AV HRP mode transceiver Simulink model The HRP transmitter and receiver main blocks are presented in Figure 7. The incoming octet stream
undergoes some initial processing in the “Frame Conditioning” block, like splitting and scrambling the MSBs and
the LSBs into upper and lower branches octet streams. In the “Frame Encoding” block both streams are
encoded with a Reed-Solomon code, interleaved and divided into four parallel streams that are convolutionally
encoded and punctured. The multiplexer block interleaves and serializes all eight parallel bit streams. In he
OFDM block the bit stream is converted to a complex coefficients stream by the system mapper and pilot, DC
and null coefficients are inserted to form an OFDM symbol. The tone interleaver shuffles the order of the all sub
carriers and the IFFT block converts the OFDM symbol to the time domain. The data received by the radio block
is delivered to the receiver input, labelled “receiver in”, and processed by the “OFDM demodulation” block, where
the time signal is converted to the frequency domain by the FFT block. The pilots, DC and null carriers are
removed and the system demapper creates a bit stream from the receiver data coefficients. The demultiplexer
block deinterleaves and separates the bit stream into eight parallel streams that are decoded by eight Viterbi
decoders, deinterleaved and decoded again by 2 Reed-Solomon decoders. In the final block the two bit octet
streams are de-scrambled and merged into a unique octet stream and provided to the upper layer.
Detailed information on all the main blocks will be presented in the subsequent sections. The structure of the
data packets transmitted is described in Figure 2.
4.1 Transceiver blocks
4.1.1 Transmitter and receiver controller
The configuration of all the transmitter blocks and sub blocks is performed by the transmitter controller block.
It decodes the information on the HRPDU header and configures the system accordingly.
The receiver controller block is responsible for configuring all the blocks on the receiver model. It decodes the
transmitted HRPDU header and generates all the signals to configure appropriately all the remaining blocks.
4.1.2 Frame processing blocks
The frame conditioning block, represented in Figure 8, is formed by three sub blocks: the stuff bits block, the
splitter block and the scrambler block. The bit stuffer is responsible for ensuring that the transmitter produces an
integer number of OFDM symbols; the splitter only processes payload data; separates and reorders each
incoming octet in the 4 MSBs and 4 LSBs; the scrambler scrambles all data bits using a polynomial scrambler.
24
Figure 7: HRP transceiver main blocks
25
Figure 8: Frame conditioning block (transmitter)
The reciprocal block in the receiver is the also called frame conditioning block, illustrated in Figure 9, and its
main blocks are the destuff bits block; the desplitter block and the descrambler block.
Figure 9: Frame conditioning block (receiver)
4.1.2.1 Stuff and “destuff” bits
Stuff bits are inserted in the payload to guarantee that an integer number of OFDM symbols will be created
for each HRPDU and to guarantee that the outer interleaver frame is complete. The additional bits are added to
the bit stream before any operation is performed on the data. All stuff bits are set to zero.
The stuff bits block guarantees that the payload sub-frames includes a multiple of 1728 octets, number
required for the outer interleaver blocks. No stuff bits are added to the HRPDU header octets – the first 92 octets
in each HRPDU frame.
On the receiver side, the decoded HRPDU header contains the length of all the transmitted sub frames and
26
the destuff bits block uses this information to discard all stuff bits.
4.1.2.2 Splitter and desplitter
The splitter block is used to reorder and divide HRPDU sub frames into two different bit streams, designated
as lower and upper bit streams. The HRP header and sub-frames transmitted with modes 5 and 6 are not
processed by the splitter and desplitter blocks.
Sub-frames using HRP modes 0 to 2 are divided into lower and upper bit streams according to (6) and (7);
sub-frames using HRP modes 3 to 4 are split according to (8) and (9).
4.1.2.3 Scrambling and descrambling
Input stream bits, including the HRP header, are scrambled by a polynomial scrambler defined by (17).
P(x) = x15+x14+1 (17)
The initial values in the delay line are given by eleven fixed seeds and four variable seeds, S0, S1, S2 and S3:
[x-1,x-2,..,x-15]=[1101 0000 101S3 S2S1S0]. The HRP header is scrambled with S0S1S2S3= [0101]. The payload bits,
including the stuff bits, are scrambled using the seeds specified in the PHY Control field22. The first eight bits of
the scrambler are applied to the lower branch and the next eight bits to the upper branch, alternating every eight
bits between the two branches. Figure 10 illustrates the inner structure of the scrambler and descrambler blocks.
Figure 10: HRP scrambler and descrambler block
The scrambling process for the HRP header is different from the scrambling done to the data because it is not
22 Detailed description of the PHY Control field in Annex D.
27
split into upper and lower branches. For this reason it was decided to process the header and the data payload
using two different blocks, the header scrambler and the scrambler. Each of these blocks implements a delay
line with constraint length K=16 sixteen and 15 delay memory elements. The output of the scrambler is the result
of the 1-bit sums without carry of each data bit and the values in the sixteenth, fifthteenth and first positions in
the delay line.
4.1.3 Frame encoding and decoding blocks
The frame encoding block has three main sub blocks: incoming octets are encoded by the Reed-Solomon
encoding block, and then the outer interleaver block groups, reorders and distributes the data by the eight
convolutional encoders that make up the convolutional encoder block.
On the receiver side, the data undergoes the reverse processing: first the eight Viterbi decoders recover the
data, which is then deinterleaved and regrouped by the outer deinterleaver and then decoded by the Reed-
Solomon decoding block.
4.1.3.1 Reed-Solomon encoder and decoder
Three different codeword lengths are used for encoding the HRPDU frame: the first 48 octets of the header
are encoded with a RS (56,48) code, the remaining 44 HRP header octets are encoded with a RS (52,44) code
and the payload octets are encoded using a RS(224,216) code.
Figure 11 and Figure 12 show the internal structure of the Reed-Solomon encoding and decoding blocks. The
main blocks of the Reed-Solomon encoding block are two Reed-Solomon encoders working in parallel encoding
simultaneously the upper branch data and the lower branch data. The implemented RS encoder blocks can
perform all three encoding schemes. The additional blocks are used for splitting, routing the header data to both
RS encoders and appending four null symbols to the portion of the header that is encoded using the RS(52,44)
code.
Figure 11: HRP transmitter Reed-Solomon encoder
28
Figure 12: HRP receiver Reed-Solomon decoder block
The RS decoding block performs the data decoding and error correction using two parallel RS decoders.
These blocks use the native matlab Reed-Solomon decoder function: rsdec.m. Additional data routing blocks
were necessary to use both RS decoders to decode the HRP header.
4.1.3.2 Outer interleaver and outer deinterleaver
The outerinterleaver block uses two interleavers, one for the upper branch data and another for the lower
branch data. Each interleaver distributes data to four convolutional encoders. The internal structure of the
outerinterleaver block in represented in Figure 13.
Figure 13: Outer interleaver block internal structure (transmitter)
29
The main blocks are the upper and lower branch interleaver blocks. The additional blocks are used to allow
the use of just the upper interleaver block as required for processing the HRP header and data for HRP modes 5
and 6. For all HRP modes the depth is defined to be four; the depth used in the HRP header is two. The matlab
function sfunct_outerinterleaver.m groups depth × N sets of octets and implements (13) for depths 2 and 4.
Figure 5 and Figure 6 show the octet interleaving scheme used in the header and data respectively.
In the receiver the reverse operation is implemented by the outer deinterleaver blocks.
4.1.3.3 Convolutional encoder and decoder
After being interleaved the data is encoded by a convolutional code. The encoder block is composed by eight
convolutional encoders in parallel, four for the data of each interleaver block. The structure of each individual
encoder is presented in Figure 14.
The decoding operation on the receiver is done by eight Viterbi decoders in parallel. The Matlab function that
implements the convolutional decoder, sfunct_viterbi_decoder.m, uses the Matlab built in function, vitdec.m, to
perform the decodification and data puncturing. This block is prepared to handle soft decoding in which the logic
values “0” and “1” are represented by a value between 0 and 2nbits-1, where nbits is the number of bits in each
soft bit.
Figure 14: Individual convolutional encoder structure
30
4.1.3.3.1 Puncturing
The convolutional encoder native code rate is 1/3; the other coded rates defined in Table 7 are generated by
the puncturing block which selects which bits will be transmitted from each of the three streams. The data bits
are selected and serialized according to the puncturing pattern and transmission sequence in Table 9.
On the receiver side the puncturing is done by the Viterbi decoders.
4.1.4 Bit interleaver and deinterleaver blocks
The bit streams generated simultaneously by the convolutional encoders are serialized and interleaved in the
multiplexer/bit interleaver block. The structure of the multiplexer and bit interleaver block is presented in Figure
15. The main blocks are the header multiplexer/bit interleaver and data multiplexer/bit interleaver, the additional
blocks just perform data routing.
Figure 15: Multiplexer/bit interleaver block internal structure
The HRP header is processed by a separate block because it utilizes a different grouping and interleaving
scheme: 12 sequential bits from each stream are grouped and interleaved according to Table 10 to form a single
stream with 48 bits.
The data multiplexer/bit interleaver has two different working modes: the EEP and UEP modes. In the EEP
mode all eight puncturing blocks use the same code rate and the incoming bits are multiplexed and interleaved
every 48 bits. The EEP multiplexer groups six bits, in increasing order of time, from each encoder, from A to H,
31
and serializing them by ordering bits A1 to A6, B1 to B6, C1 to C6, D1 to D6, E1 to E6, F1 to F6, G1 to G6, H1 to
H6 and interleaving them according to Table 12.
In the UEP mode, the four upper puncturing blocks use a code rate of 4/7 while the four lower puncturing
blocks use a code rate of 4/5. The encoded bits are multiplexed and bit-interleaved every 96 bits in a two phase
cycle. In the first half of the cycle, the UEP multiplexer groups sequentially seven bits from each one of the four
upper puncturing blocks, A to D, and five bits from each of the lower puncturing blocks, E to H, and creates a
serial stream: A1 to A7, B1 to B7, C1 to C7, D1 to D7, E1 to E5, F1 to F5, G1 to G5, and finally H1 to H5. The
bits are then interleaved according to Table 12. In the second phase, the UEP multiplexer, groups again seven
bits from each of the upper puncturing blocks, and five bits from each of the lower puncturing blocks and
serializes the bits from B8 to B14, C8 to C14, D8 to D14, A8 to A14, F6 to F10, G8 to G10, H8 to H10, and finally
E7 to E10'. These 48 bits are also interleaved according to Table 12.
The HRP demultiplexer/bit deinterleaver block is represented on Figure 16. Similarly to the multiplexer/bit
interleaver block the main blocks are the header demultiplexer and data demultiplexer. These blocks
deinterleave the data stream received from the system demapper and separate it into parallel bit streams. The
deinterleaving tables can be derived from Table 10 and Table 12 by interchanging the values in the “Before” and
“After” columns.
Figure 16: HRP demultiplexer and bit deinterleaver
4.1.5 OFDM modulator and demodulator blocks
The OFDM modulator, Figure 17, includes the following blocks: system mapper, preamble, pilots insert
blocks, tone interleaver, the IFFT block and some additional routing blocks.
32
The serial data stream is mapped to complex coefficients by the system mapper and the insert pilots block
adds the null, dc and pilots. The carriers are interleaved by the tone interleaver block and the IFFT block
converts the OFDM symbol to the time domain and the signal is sent to the radio block. Each OFDM symbol has
512 sub carriers: 336 modulated by data, 157 null carriers, 16 pilot carriers and 3 DC carriers.
Figure 17: OFDM modulator block
The OFDM demodulator block, Figure 18, performs the reverse processing: the received signal is converted
to the frequency domain by the FFT block, the complex coefficients are deinterleaved and the pilots, null and DC
coefficients are separated from the data coefficients and the system demapper block converts the complex
coefficients into a bit stream.
Figure 18: OFDM demodulator block
4.1.5.1 Mapper and Demapper
The system mapper maps groups of bits from the HRP multiplexer/bit interleaver to complex coefficients that
will be used to modulated the OFDM data sub carriers. The QPSK and 16-QAM modulation tables are presented
in Table 14 and Table 15.
The system demapper performs soft decoding. Soft decoding represents logical levels with several bits, nbits,
to indicate the uncertainty associated with the demapping of each bit. The demapper output for each bit varies
from 0, certain “0” bit, to 2nbits-1, certain “1”.
33
4.1.5.2 Tone interleaver and deinterleaver
The tone interleaver reorders the 512 complex coefficients before the IFFT is performed according to the
algorithm described on Chapter 3 section 4.3. The deinterleaving operation is the same and therefore can be
implemented by the same function.
4.1.5.3 IFFT and FFT blocks
The IFFT block transforms the 512 complex coefficients, corresponding to a representation in the frequency
domain, into 512 complex coefficients in the time domain. The FFT block recovers the coefficients in the
frequency domain. The IFFT and FFT blocks are built around the native Matlab functions ifft.m and fft.m.
4.2 Complete Matlab/Simulink model
The top level overview of the Matlab/Simulink model is presented in Figure 19. To simulate the transceiver
operation two additional blocks were introduced: the HRDPU frame generator block and a channel block.
Figure 19: Complete Simulation model
The HRPDU frame generator block creates the data that will be transmitted and the channel block simulates
several transmission channel models.
4.2.1 HRPDU frame generator
The frame generator block is not included in transmitter block and is responsible for generating the data to be
transmitted. This block structure is shown in Figure 20.
This block reproduces the frame structure described in Section 6.5, and the information on the HRP PHY
control and sub frames headers can be configured through the HRPDU frame settings block.
4.2.2 Channel block
The channel block implements three different channels:
34
Ideal channel;
Random noise channel, which adds random noise to the amplitude and phase of the sub carriers;
Additive White Gaussian Noise channel;
Configurable attenuation and phase channel, which allows defining the attenuation and phase values
for each individual sub carrier.
Figure 20: HRPDU frame generator
This block reproduces the frame structure described in Section 6.5, and the information on the HRP PHY
control and sub frames headers can be configured through the HRPDU frame settings block.
4.2.3 Model settings
The model top level view, shown in Figure 21, has four different settings blocks that allow changing
configuration parameters in the Matlab/Simulink model (Figure 19).
Each block allows changing predefined model parameters. The HRPDU frame settings block allows the
configuration of the following parameters:
The number of HRPDU frames to generate;
The scrambler seed to use in the data scrambler;
The mode and length of each payload sub-frame.
The transmitter settings block allows modifying the numeric representation of the complex coefficients used in
the IFFT block the between the default Matlab double precision and customizable fixed-point representation.
The receiver settings block enables configuring:
The number of soft bits to use in the system demapper, demultiplexer and Viterbi decoders blocks;
The numerical representation used in the FFT block: either double precision or customizable fixed
35
point representation.
The channel settings block allows choosing and configuring the type of channel to use in model.
Ideal channel;
Random noise channel;
Additive White Noise Channel;
Configurable attenuation and phase channel, which allows defining the attenuation and phase values
for each individual sub carrier.
Figure 21: Model top level view
36
Chapter 5
Tests and results In order to verify the correct function of the model, a series of tests were conducted: a HRPDU frame
configured as shown in Table 17 was transmitted over an ideal channel, the Random noise channel and the
AWGN channel. The maximum available numeric precision was used in the IFFT and FFT blocks and the
system demapper used sixteen levels (four soft bits). The frame structure was chosen to provide an overview of
the performance of the main HRP modes when transmitted over noisy channels. The random noise channel was
configured to have 3 dB of maximum noise amplitude and 0.2 radians of maximum phase variation, and the
AWGN channel was configured to have a SNR of 43 dB.
The test was executed three times and the average bit error rate (BER) and the average number of bit errors
at the output of the system demapper (1) and at the output of the receiver (2) are presented in Table 18. The
BER was measured at the output of the system demapper and at the overall system output to measure the effect
of the decoding block on the performance of the system.
Table 17: Test HRPDU frame structure
Mode
Length
(octets)
Sub-frame 1 0 1000
Sub-frame 2 1 1000
Sub-frame 3 2 1000
Sub-frame 4 3 1000
Sub-frame 5 4 1000
Sub-frame 6 0 1000
Sub-frame 7 1 1000
The results show that in, these conditions, modes two and four are the most sensitive to channel noise. This
result was expected because 16-QAM modulation is less permissive to channel noise than QPSK and, in the
case of sub-frame 5, the use of lower rate convolutional code rate. These results also show that the decoding
blocks were able to correct some of the errors introduced by the channel.
To determine the minimum numerical precision to use in the IFFT block a sub-frame using mode 4 was
transmitted over the ideal channel while varying the numerical representation used in the IFFT block. The
numerical precision used on the FFT block and the number of soft bits remained constant and set to maximum
and four respectively. These tests conditions were chosen because they represent the worst case scenario
verified in the previous test.
37
Table 18: Test results
Sub
frame 1
BER
(nr. errors)
Sub
frame 2
BER
(nr. errors)
Sub
frame 3
BER
(nr. errors)
Sub
frame 4
BER
(nr. errors)
Sub
frame 5
BER
(nr. errors)
Sub
frame 6
BER
(nr. errors)
Sub
Frame 7
BER
(nr. errors)
1 2 1 2 1 2 1 2 1 2 1 2 1 2
Ideal Channel 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Random noise
Channel 0 0 0 0
0.26%
(230)
130
(0.44%) 0 0
2.86%
246
1.96%
157 0 0 0 0
AWGN Channel 0 0 0 0 0.16%
(100)
0,11
(9) 0 0
0,16%
134
1.41%
113 0 0 0 0
The results are presented in Table 19 and the chosen threshold criterion was that no errors were detected at
the output of the system demapper, i.e., BER=0. This criterion guarantees that no errors are introduced in the
system due to the numerical representation used on the IFFT block.
Table 19: Determining minimum numerical precision for the IFFT block
Integer bits
(1 signal bit)
Fractional
bits
BER at
system demapper
output
2 10 1.522%
3 5 5.018/
3 6 1.414%
3 7 0.023%
3 8 0
3 9 0
3 10 0
3 11 0
The same procedure and criterion was used to determine the minimum numerical precision required in the
FFT block: the IFFT block precision was fixed at the maximum value and varying numerical precision of the FFT
block. Again the sub-frame used mode 4 and the same Random noise channel as the previous tests.
The results are presented in Table 20. Both results indicate the fixed point representation 3.8 as the minimum
in order not to introduce errors to the Viterbi decoders.
To determine the influence of the number of soft bits on the overall system BER, the numeric precision on
both IFFT and FFT blocks was set to one signal bit, two integer bits and eight fractional bits; the same HRPDU
sub-frame was transmitted over the random channel, which on the previous tests proved to introduce more
errors than the AWGN channel, while varying the number of soft bits used.
The results are presented in Table 21 and show that the influence of using fixed point representation on the
IFFT and FFT blocks can be compensated by the increase in the number of soft bits used by the system
demapper and Viterbi decoders. These results show that the use of five soft bits and the minimum fixed point
representation determined previously yields overall BER similar to the BER achieved with hard decoding on the
38
system demapper and floating point representation on the IFFT and FFT blocks.
Table 20: Determining minimum numerical precision for the FFT block
Integer
Bits
(1 signal bit)
Fractional
bits
BER at
system demapper
output
2 10 1.536%
3 5 4.070%
3 6 0.868
3 7 0.007%
3 8 0
3 9 0
3 10 0
3 11 0
Table 21: Influence of the number of soft bits
Number of soft bits/
numeric precision
BER
at receiver output
1/Floating point 0.85%
1/ 3.8 8.90%
2/ 3.8 4.56%
3/ 3.8 2.44%
4/ 3.8 1.41%
5/ 3.8 0.67%
39
Chapter 6
Conclusion Our goal was to implement a Matlab/Simulink model for an uncompressed HD video transceiver. To
accomplish this, we started by analysing the characteristics of video and audio signals and establishing the
requirements for transmitting uncompressed HD video and audio. The next step was to analyse the state of the
art in wireless video transmission to choose the standard to use. The analysed technologies are separated into
two different frequency bands: the 2.4 GHz to 10 GHz band which is used by WHDI, WUSB, WiDi and IEEE
802.11ac; and the 60 GHz band used by IEEE 802.15.3c, ECMA-387 and IEEE 802.11ad. On the 2.4-10 GHz
band only WHDI and IEEE 802.11ac fulfil the requirements for Full HD video transmission but WHDI is a
proprietary specifications and IEEE 802.11ac is still in draft. Due to the very high amount of continuous
bandwidth all the standards on the 60 GHz band provide enough throughput to transmit Full HD video.
We choose a standard from the 60 GHz band because the amount of bandwidth available enables all the
standards to achieve higher bit rates than the 2.4 GHz - 10 GHz band standards/specifications. Among the three
possible options, we choose IEEE 802.15.3c because it provided enough throughput to transmit uncompressed
HD video and has a dedicated OFDM mode with a forward error protection scheme more appropriate for video
transmission.
A Matlab/Simulink model was developed to simulate the end-to-end transmission system. The model
includes the transmitter base band processor, the receiver base band processor, a frame generator block and
several channel models. The frame generator emulates the MAC layer output and generates data frames with
the structure described in Annex D which can be used to perform end-to-end simulations. This is important
because it enables to evaluate the behaviour of the system in the presence of channel noise, and in future
developments, can also be used to evaluate the performance of the hardware blocks in the same channel
conditions.
Simulations using this model permitted to determine the minimum numerical precision of the complex
coefficients in the OFDM modulation/demodulation blocks. A fixed point representation with 11 bits, one for the
signal, two for the integer part and eight to represent the fractional part, proved to be the minimum numerical
representation in which no errors were detected at the output of the system demapper, meaning that the
quantization error introduced by this numerical representation was small enough as to not interfere with the
performance of the system demapper. The influence of the number of bits in each soft bit generated by the
system demapper and used by the Viterbi decoder was also analysed. The simulations show that using 5 soft
bits and the 3.8 fixed point representation on the transmitter and receiver OFDM blocks, yields a lower BER in
the overall communication than the BER obtained when using floating point representation and hard decision in
the OFDM blocks.
We expect the IFFT/FFT, Viterbi and Reed-Solomon decoders to prove the most difficult to implement in
hardware because of the complex algorithms they implement and the very high amount of data that most be
40
41
processed in these blocks. We also expect the system mapper/demapper and HRP multiplexer/demultiplexer to
be challenging to implement in hardware because these blocks must process the bit stream resulting from the
serialization of the eight parallel convolutionally coded bit streams.
The model is a useful tool to guide a future hardware implementation of a transceiver’s base band processor,
because it enables the behavioural validation of the hardware blocks, and also enables evaluating the influence
of different hardware implementations on the overall system performance for any hardware block. Finally the
model provides a platform for testing the overall performance influence of different solutions from the specified in
the IEEE 802.15.3c standard, e.g.: different block codes; different polynomial generators for the convolutional
encoders/decoders; different sub-carrier modulations schemes.
The model was simulated using channels that add random noise to the transmitted signal, however, TG3c
recommends, a more realistic channel model which includes the typical use case scenarios for residential and
office environments. Therefore, future work includes the introduction of this model in the developed transceiver
model. Also, as future work LRP mode be used. The inclusion of the LRP mode would enable the exchange of
control signals between the transmitter and receiver enabling for example, the transmitter to choose the HRP
mode according to the channel’s noise characteristics.
This model was developed as an initial step for the hardware implementation of the transceiver. Therefore,
the work will continue: defining hardware architectures that can fulfil timing requirements for each block, with
special attention to the OFDM modulator/demodulator, Viterbi and Reed-Solomon decoders, mapper/demapper
and HRP multiplexer/demultiplexer. Future work also includes the validation of the implemented hardware blocks
and the evaluation of the overall system performance, i.e. transmitter’s baseband processor, channel and
receiver’s baseband processor.
References
[1] Steve Venuti, "Introducing HDMI 1.4 Specification Feature", Press Release, 2009
(http://www.hdmi.org/download/press_kit/PressBriefing_HDMI1_4_Final_083109.pdf).
[2] "Display Port Technical overview", Press Release, May 2010
(http://www.displayport.org/cms/sites/default/files/downloads/DisplayPort_Technical_Overview.pdf).
[3] "Part 15.3: Wireless Medium Access Control (MAC and Physical Layer PHY) Specifications for High Rate
Wireless Personal Area Networks WPANs", IEEE 802.15.3c. PHY specification, October 2009.
[4] "WirelessHD Specification Overview", WiHD Specification Overview, August 2009
(http://www.wirelesshd.org/wp-content/uploads/2009/12/WirelessHD-Specification-Overview-v1-0-4-
Aug09.pdf)
[5] "High Rate 60 GHz PHY, MAC and PALs", ECMA-387 Specification, December 2010.
[6] “Proposed TGac Draft Amendment", IEEE 802.11ac specification draft, November 2010.
[7] "PHY/MAC complete Proposal Specification", 802.11ad specification draft, May 2010.
[8] "Defining the Future of Multi-Gigabit Wireless Communications", White Paper, July 2010.
[9] "Wireless Universal Serial Bus Specification 1.1", WUSB Specification, September 2010.
[10] "MultiBand OFDM Physical Layer Specification", MB OFDM PHY specification, August 2009
[11] "WHDI technology Overview", April 2011 (http://www.whdi.org/Technology/).
[12] "Intel Wireless Display", April 2011
(http://download.intel.com/network/connectivity/products/prodbrf/323116.pdf) .
[13] Reed Fisher, "60 GHz WPAN Standardization within IEEE 802.15.3c", in Proc ISSSE ’07- International
Symposium on Signals, Systems and Electronics, August 2007.
[14] "Information technology - Telecommunications and information exchange between systems - High rate 60
GHz PHY, MAC and HDMI PAL", November 2009.
[15] Woosung Lee et al., "Beamforming Lens Antenna on a High Resistivity Silicon Wafer for 60 GHz WPAN",
IEEE Transactions on antennas and propagation, vol.. 58, no. 3, March 2010, pp. 706-713.
[16] Junyi Wang et al., "A Pro-Active Beamforming Protocol for Multi-Gbps Millimetre-Wave WPAN Systems", in
Proc. of WCNC 2010 – 2010 IEEE Wireless Communications and Networking Conference, Sydney,
Australia, April 2010.
[17] Jie-Wei Lai, Alberto Valdes-Garcia, "A 1V 17.9dBm 60GHz Power Amplifier in Standard 65nm CMOS", in
Proc. of ISSCC- IEEE international Solid State Circuits Conference, San Francisco, USA, February 2010.
[18] Oliver Hoffmann et al., "Coded Performance of OFDM and SC PHY of IEEE 802.15.3c for Different FEC
Types", in Proc of IEEE-Globecom workshops 2009, Honolulu, USA, December 2009.
[19] Shen-Jui Huang et al., "A Green FFT Processor with 2.5-GS/s for IEEE 802.15.3c (WPANs)", in Proc
ICGCS-International Conference on Green Circuits and Systems 2010, Shanghai, P.R. China, June 2010.
[20] Nicholas Hinitt, Taskin Kocak, "GPU-Based FFT Computation for Multi-Gigabit WirelessHD Baseband
Processing", EURASIP Journal on Wireless Communications and Networking, June 2010.
42
43
[21] "Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications", November
2007.
[22] Chang-Soon Choi et al., "60-GHz OFDM systems for multi-gigabit wireless LAN applications", in Proc. 7th
Annual IEEE Consumer Communications and Networking Conference, Las Vegas, USA, January 2010.
[23] Kang, Fujiang Lin, "A 20-GHz Integer-N Frequency Synthesizer for 60-GHz Transceivers in 90nm CMOS",
in Proc 2010 IEEE International Conference on Ultra-Wideband, Nanjing, P.R. China, September 2010.
[24] "Digital Visual Interface DVI Revision 1.0", April 1999 (http://www.ddwg.org/lib/dvi_10.pdf)
[25] Editors Su-Khiong Yong, Pengfei Xia and Alberto Valdes Garcia “60 Ghz technology for Gbps WLAN and
WPAN, From Theory to Practice”, 2011, Wiley.
[26] Qimei Cui et al., “Gbps Wireless Communication System Design and Transmitter Implementation”, in Proc.
3rd IEEE International Conference on Broadband Network & Multimedia TechnologyJanuary, Beijing,
P.R.China October 2011.
[27] Khaled Sobaihi et al., “FPGA Implementation of OFDM Transceiver for a 60 Ghz Wireless Mobile Radio
System”, in Proc ReConFig 2010 - International Conference on Reconfigurable Computing and FPGAs,
Cancum. Mexico, December 2010.
[28] "Part 15.3: Wireless Medium Access Control (MAC) and Physical Layer (PHY) Specifications for High Rate
Wireless Personal Area Networks (WPANs)", IEEE 802.15.3 PHY Specification, September 2003.
[29] "White Paper Defining the Future of Multi-Gigabit Wireless Communications",July 2010,
(http://wirelessgigabitalliance.org/?getfile=1510)
Annex A. IEEE 802.15.3c specifications Table 22: SC PHY MCS parameters
MCS
class
MCS
Identifier
Data rate
(Mbps) with
pilot word
length=0
Data rate
(Mbps) with
pilot word
length=64
Modulation FEC type
0 25.8 -
1 412 361
2 825 722
3 1650 1440
RS(255,239)
4 1320 1160 LDPC(672,504)
5 440 385 LDPC(672,336)
Class1
6 880 770
Π/2 BPSK/(G)MSK
7 1760 1540
LDPC(672,336)
8 2640 2310 LDPC(672,504)
9 3080 2700 LDPC(672,588)
10 3290 2870 LDPC(1440,1344)
Class2
11 3300 2890
Π/2 QPSK
RS(255,239)
12 3960 3470 Π/2 8-PSK Class3
13 5280 4620 Π/2 16-QAM
LDPC(672,504)
Table 23: HSI MCS parameters [3]
FEC type c
Data rate
(Mbps) Modulation
MSB 8b LSB 8b
0 32.1 QPSK LDPC(672,336); bit rate 1/2
1 1540 QPSK LDPC(672,336); bit rate 1/2
2 2310 QPSK LDPC(672,504); bit rate 3/4
3 2695 QPSK LDPC(672,588); bit rate 7/8
4 3080 16-QAM LDPC(672,336); bit rate 1/2
5 4620 16-QAM LDPC(672,504); bit rate 3/4
6 5390 16-QAM LDPC(672,588); bit rate 7/8
7 5775 64-QAM LDPC(672,420); bit rate 5/8
44
Table 24: HSI MCS parameters [3] (cont.)
FEC type c
Data rate
(Mbps) Modulation
MSB 8b LSB 8b
8 1925 QPSK LDPC(672,336);
bit rate 1/2
LDPC(672,504);
bit rate 3/4
9 2503 QPSK LDPC(672,504) LDPC(672,588);
bit rate 7/8
10 3850 16-QAM LDPC(672,336);
bit rate 3/4
LDPC(672,504);
bit rate 3/4
11 5005 16-QAM LDPC(672,504);
bit rate 3/4
LDPC(672,588);
bit rate 7/8
Table 25: HSI OFDM timing parameters [3]
Parameter Value
Occupied bandwidth 1815 GHz
Reference Sampling (fs) 2.640 Gsamples/s
Number of sub-carriers
(Nsc) 512
FFT period (TFFT) Nsc(HR)/ fs(HR) ~ 193.94 ns
Sub-carrier spacing
(Δfsc) 1/ TFFT(HR) ~ 5.15625 MHz
Guard interval (TG) 64/ fsc(HR) ~ 24.24 ns
Symbol duration (TS) TFFT(HR) + TGI(HR) ~ 226.95 ns
Number of data sub-
carriers (Nds)) 336
Table 26: LRP frequency plan [3]
Channel index Start frequency Centre frequency Stop frequency
1 fc(HRP) – 207.625 MHz fc(HRP) – 158.625 MHz fc(HRP) – 109.625 MHz
2 fc(HRP) – 49 MHz fc(HRP) fc(HRP) + 49 MHz
3 fc(HRP) + 109.625 MHz fc(HRP) + 158.625 MHz fc(HRP) + 207.625 MHz
45
46
Table 27: LRP data rates and coding [3]
LRP mode
index Modulation FEC
Data rate
(Mbps) Repetition
0 1/3 2.5 8x
1 1/2 3.8 8x
2 2/3 5.1 8x
3
BPSK
2/3 10.2 4x
Table 28: LRP OFDM timing parameters [3]
Parameter Value
Occupied bandwidth 92 MHz
Reference Sampling (fs(LR)) 2.538 Msamples/s
Number of sub-carriers (Nsc(LR)) 128
FFT period (TFFT(LR)) Nsc(HR)/ fs(HR) ~ 403.47 ns
Sub-carrier spacing (Δfsc(LR)) 1/ TFFT(LR) ~ 4.957 MHz
Guard interval (TGI(LR)) 64/ fsc(LR) ~ 88.26 ns
Symbol duration (TS(LR)) TFFT(LR) + TGI(LR) ~ 491.73 ns
Number of data sub-carriers
(Ndsc(LR)) 30
Annex B. ECMA-387 specifications Table 29: Work modes for Type A devices
Base Data Rate (Gbps)23
Mode No
channel
bonding
2 bonded
channels
3 bonded
channels
4 bonded
channels
Modulatio
n
Carrier
modulation FEC
FEC
Code
Bit Rate
A0 0.397 0.794 1.191 1.588 SCBT BPSK RS & CC 1/2
A1 0.794 1.588 2.381 3.175 SCBT BPSK RS & CC 1/2
A2 1.588 3.175 4.763 6.350 SCBT BPSK RS 1
A3 1.588 3.175 4.763 6.350 SCBT QPSK RS & CC 1/2
A4 2.722 5.443 8.166 10.886 SCBT QPSK RS & CC 6/7
A5 3.175 6.350 9.526 12.701 SCBT QPSK RS 1
A6 4.234 8.467 13.701 16.934 SCBT NS 8QAM RS & TCM 5/6
A7 4.763 9.526 14.288 19.051 SCBT 16QAM RS 1
A8 4.763 9.526 14.288 19.051 SCBT TCM-16QAM RS & TCM 2/3
A9 6.350 12.701 19.051 25.042 SCBT 16QAM RS 1
A10 1.588 3.175 4.763 6.350 SCBT QPSK RS & UEP-CC 1/2 (MSB)
A11 4.234 8.467 12.701 16.934 SCBT 16QAM RS & UEP-CC 4/7 (MSB) 4/5 (LSB)
A12 2.117 4.234 6.350 8.467 SCBT UEP-QPSK RS & CC 2/3
A13 4.234 8.467 12.701 16.934 SCBT UEP-16QAM RS & CC 2/3
A14 1.008 N/A24 N/A N/A OFDM QPSK RS & CC 1/3
A15 2.016 N/A N/A N/A OFDM QPSK RS & CC 2/3
A16 4.032 N/A N/A N/A OFDM 16QAM RS & CC 2/3
A17 2.016 N/A N/A N/A OFDM QPSK RS & UEP-CC 4/7 (MSB) 4/5 (LSB)
A18 4.032 N/A N/A N/A OFDM 16QAM RS & UEP-CC 4/7 (MSB) 4/5 (LSB)
A19 2.016 N/A N/A N/A OFDM UEP-QPSK RS & CC 2/3
A20 4.032 N/A N/A N/A OFDM UEP-16QAM RS & CC 2/3
A21 2.016 N/A N/A N/A OFDM QPSK RS & CC 2/3 (MSB)
23 Assuming a cyclic prefix length of zero. 24 OFDM modes do not support channel bonding.
47
48
Table 30: Working modes for Type B devices
Base Data Rate (Gbps)
Mode No
channel
bonding
2 bonded
channels
3 bonded
channels
4 bonded
channels
Carrier
modulation FEC
B0 0.794 1.588 2.381 3.175 DBPSK RS & Diff
B1 1.588 3.175 4.763 6.350 DBPSK RS & Diff
B2 3.175 6.350 9.526 12.701 DBPSK RS & Diff
B3 3.175 6.350 9.526 12.701 UEP-QPSK RS
Annex C. IEEE 802.1ad specifications Table 31: MCS modes for 802.11ad SC modulation [7]
MCS Index Carrier Modulation FEC type Data Rate (Mbps)
1 π/2-BPSK LDPC(672,336) 385
2 π/2-BPSK LDPC(672,336) 770
3 π/2-BPSK LDPC(672,420) 962.5
4 π/2-BPSK LDPC(672,504) 1155
5 π/2-BPSK LDPC(672,546) 1251.25
6 π/2-QPSK LDPC(672,336) 1540
7 π/2-QPSK LDPC(672,420) 1925
8 π/2-QPSK LDPC(672,504) 2310
9 π/2-QPSK LDPC(672,546) 2502.5
10 π/2-16QAM LDPC(672,336) 3080
11 π/2-16QAM LDPC(672,420) 3850
12 π/2-16QAM LDPC(672,504) 4620
Table 32: MCS modes for 802.11ad OFDM modulation [7]
MCS index Carrier Modulation FEC type Data Rate (Mbps)
13 SQPSK LDPC(672,336) 693.00
14 SQPSK LDPC(672,420) 866.25
15 QPSK LDPC(672,336) 1386.00
16 QPSK LDPC(672,420) 1732.50
17 QPSK LDPC(672,504) 2079.00
18 16-QAM LDPC(672,336) 2772.00
19 16-QAM LDPC(672,420) 3465.00
20 16-QAM LDPC(672,504) 4158.00
21 16-QAM LDPC(672,546) 4504.50
22 64-QAM LDPC(672,420) 5197.50
23 64-QAM LDPC(672,504) 6237.00
24 64-QAM LDPC(672,546) 6756.75
49
Annex D. HRPDU detailed structure
D.1 HRP preamble
The HRP preamble is introduced to signal the beginning of a new HRPDU and improve the synchronization
of the receiver device. It is formed by 8 OFDM symbols, the first four OFDM symbols are generated by an 8th
order polynomial, defined by p(x) = x8+x7 +x2 +x+1, re-sampled at 3/2 rate. The last four OFDM symbols are
predefined, in the frequency domain, in Table 156 from [3].
D.2 HRP header
The HRP header includes information about how the transmission is performed. It includes information about
the error protection used, HRP header scrambling seed and the HRP mode used to transmit all the sub-frames.
The HRP header format is presented in Table 33.
Table 33: HRP header structure
3 octets ... 3 octets 3 octets 1 octet
Sub-frame header 7 ... Sub-frame header 2 Sub-frame header 1 PHY control
The HRP header sub fields are described in the following sections.
D.2.1 PHY control
The PHY control field is formed by the UEP mapping field and the HRP header scrambling seeds, S3, S2, S1
and S0 as showed in Table 34.
Table 34: PHY control structure
Bits: b7-b6 b5 b4 b3 b2 b1 b0
Reserved S3 S2 S1 S0 UEP mapping Reserved
The UEP mapping indicates which UEP mode is used, if set to '1', the UEP coding described in sections
12.4.2.10.3 and 12.4.2.12 from [3], and if set to '0', the UEP mapping described in section 12.4.2.10.2 from [3].
The value of S3, S2, S1 and S0 is used as the seed value for scrambling the HRP header.
D.2.2 Sub-frame header
The sub-frame header has two fields as indicated in Table 35, the HRP mode index field and the length field.
The latter contains the number of octets in the respective sub-frame, and the former indicates the mode that will
50
be used to transmit the respective sub-frame. The valid HRP modes are presented in Table 36.
Table 35: Sub-frame structure
Bits: b23-b4 b3-b0
Length HRP mode index
Table 36: HRP data rates and coding
Code rate HRP mode
index
Coding
mode Modulation
MSB 4b LSB 4b
Raw data
rate
(Gbps/s)
0
EEP QPSK 1/3 0.952
1 QPSK 2/3 1.904
2 16-QAM 2/3 3.807
3
UEP QPSK 4/7 4/5 1.904
4 16-QAM 4/7 4/5 3.807
5 MSB-only
retransmission QPSK 1/3 N/A 0.952
6 QPSK 2/3 N/A 1.904
D.3 Extended MAC header
The MAC header for the AV HRPDU frame is the Extended MAC header format defined in section 7.2.9.1 [3],
that extends the MAC header defined in section 7.2 [28] with specific fields for Audio and Video transmission.
The Extended MAC header format is represented in Table 37.
Table 37: Extended MAC header
16 octets 24 octets 5 octets 5 octets 2 octets 10 octets
Reserved Video Header Security Header MAC extension
header
Extended
control header MAC header
51
Each of the Extended MAC header sections are described in the following sections.
D.3.1 MAC header field
The MAC header field is defined in section 7.2 from the original specification [28] and the format is
represented in Table 38.
Table 38: MAC header format
1 octet 3 octets 1 octet 1 octet 2 octets 2 octets
Stream
index
Fragmentation
control SrcID DestID PNID
Frame
Control
D.3.1.1 Frame Control
The frame control field provides basic information about the payload, like the type of data, type of security.
The frame control field structure is presented in Table 39.
Table 39: Frame Control field
Bits:b15-b11 b10 b9 b8-b7 b6 b5-b3 b2-b0
Reserved More Data Retry ACK policy SEC Frame type Protocol
Version
The Protocol Version field is used to indicate the protocol version used. The Frame type field indicates the
type of frame that will be sent. The possible frame types are presented in Table 40.
Table 40: Frame types
Value
(b5b4b3) Frame type
000 Beacon frame
001 Immediate ACK(Imm-ACK) frame
010 Delayed ACK (Dly-ACK)frame
011 Command frame
100 Data frame
101-111 Reserved
52
The SEC field value indicates if the frame is protected using the key specified in the secure session ID
(SECID) field25.
The ACK policy field indicates the acknowledgment procedure that should be used by the destination device.
The different ACK policies are listed in Table 32.
Table 41: ACK policy types
Value
(b8b7)
ACK policy
type Description
00 No ACK
The recipient does not acknowledge the frame and the sender
considers that the frame was successfully sent, irregardless of
the actual result.
01 Immediate ACK
(Imm-ACK)
The recipient sends an Imm-ACK upon successfully receiving
the frame.
10 Delayed ACK (Dly-
ACK)
The recipient keeps track of the received frames with this policy
until the senders sends a Dly-ACK request.
11 Delayed ACK
request The receiver send either an Imm-ACK or a Dly-ACK.
The Retry field is set to one if the frame is a re-transmission of an earlier data or command frame. It will be
set to zero in all other situations.
The More Data field is set to '1' to signal that the sender will continue to use the channel. The '0' value
indicates that the sender will free the channel after transmitting the current HRPDU.
The remaining bits, bit 11 through bit 15, are reserved for future uses and will be set to '0'.
D.3.1.2 PicoNet ID (PNID)
The PNID contains the piconet identifier, determined by the piconet controller (PNC).
D.3.1.3 DestID and SrcID
The identification of the source device (SrcID) and the destination device (DestID). The piconet controller
(PNC) will use the 0x00 ID and the remaining devices will use identifiers from 0x01 to 0xEC. All other ID values
are reserved and they use is shown in Table 42.
25 The SECID is defined in the security control field from the security header (section 6.4.3.2.1.1).
53
Table 42: Reserved device ID
Device ID Description
0x00 Reserved for the PNC.
0xED to 0F6 Reserved for future use.
0xF7 to 0xFC Reserved for neighbor piconets.
0xFD Reserved for multicast frames.
0xFE Reserved for use by all unassociated devices, until an identifier is
allocated by the PNC.
0xFF Reserved for broadcast frames.
D.3.1.4 Fragmentation control
The fragmentation control field is used in the fragmentation and reassembly of the MAC service data unit
(MSDU). The fragmentation control structure is presented in Table 43.
Table 43: Fragmentation Control structure
Bits:b23 b2-b16 b15 b8-b0
Reserved Last fragment
number Fragment number MSDU number
The MSDU number contains the frame sequence number in the current MSDU and the Fragment number
indicates the order of the frame in the MSDU. The Last fragment number is the total number of fragments minus
one. The last field is set to zero if the frame is not fragmented.
D.3.1.5 Stream index
The PNC allocates a unique stream index for each isochronous stream. The allowed values belong to the
interval [0x01, 0xFC]. The remaining values are reserved for other purposes as presented in Table 44.
Table 44: Stream Index reserved values
Stream index Purpose
0x00 Asynchronous data
0xFD Management channel time allocation (MCTA) traffic
0xFE Unassigned stream
54
D.3.1.6 Extended control header
The extended control header contains information about other fields in the Extended MAC header and its
structure is presented in Table 45.
Table 45: Extended control header
Bits:b15-b9 b8-b6 b5 b4 b3 b2-b0
Reserved Frame type Video header
present
Security
header
present
MAC
extension
header
present
Frame class
The frame class field indicates if the frame is a regular frame type, 0b000, or an AV frame, 0b001. The fields
MAC extension header present, Security header present and Video header present, indicate that the respective
field is present in the Extended MAC header, 0b1, or not, 0b0. The frame type field is only relevant for regular
class frames, and will be set to 0b000 for all other frame classes. The allowed values, for regular class frames,
are presented in Table 46. AV frames will use the Type field in the MAC extension header.
Table 46: Regular class frame type values
Frame type value
(b8b7b6) Description
000 MAC commands
001 Data
010 Audio
011-111 Reserved
D.3.2 MAC extension header
The MAC extension header contains information about the type of data in each sub-frame for AV class
frames. The complete MAC extension header is presented in Table 47.
Table 47: MAC extension header format
Bits:b39-b32 b31-b28 b27-b24 b23-b20 b19-b16 b15-b12 b11-b8 b7-b4 b3-b0
ACK groups Reserved Type 7 Type 6 Type 5 Type 4 Type 3 Type 2 Type 1
55
Each AV class sub-frame Type can have one of the following values, presented in Table 48.
Table 48: Allowed AC class sub frame types
Type value Description
0000 MAC commands
0001 Data
0010 Audio
0011 Video
0100-1111 Reserved
The ACK groups field is constituted by 8 bits: one for each sub-frame and an additional bit for indicating if the
lsb FCS is used in the verification if the sub-frame were received correctly. The ACK groups structure is
represented in Table 49.
Table 49: ACK groups field
Bits:b7 b6 b5 b4 b3 b2 b1 b0
lsb FCS Sub-frame 7 Sub-frame 6 Sub-frame 5 Sub-frame 4 Sub-frame 3 Sub-frame 2 Sub-frame 1
If a sub-frame belongs to the same ACK group as the previous sub-frame, then the corresponding sub-frame
bit is set to '1', otherwise, it will be set '0'. The bit, b0, must always be set to '0', as it is the beginning of the first
ACK group. There can only be up to five ACK groups defined, meaning that only up to five bits cat be set to '0'. If
the lsb FCS bit is set to '1', them the lsb FCS is used to verify if the sub-frames were correctly received.
D.3.2.1 Security header
The security header field is formed by two sub fields, Security control and Security Frame Check (SFC) as
shown in Table 50. Both sub fields purposes are explained in the following sections.
Table 50: Security header structure
2 octets 3 octets
SFC Security control
D.3.2.1.1 Security control
The security control field has the following structure shown in Table 51.
56
Table 51: Security control field
Bits: b23-b22 ... b11-b10 b9-b8 b7-b0
Sub-frame 1
security ...
Sub-frame 7
security Reserved SECID
The secure session ID (SECID) field is used to indicate the key set used to encrypt and/or authenticate the
sub-frame. Each sub-frame security field indicates the type of security used for the respective sub-frame. The
allowed values for these fields are:
0b00: no security applied to the sub-frame;
0b01: encrypted sub-frame;
0b10-0b011: are reserved.
D.3.2.2 Secure frame counter (SFC)
The secure frame counter is used to ensure that every nonce is unique in a secure frame, meaning that
devices must not reuse the SFC in the same time token. The SFC will be implemented for all sub-frames, even
unsecured sub-frames.
D.3.2.3 Video header
The video header structure is defined as shown in Table 52.
Table 52: Video header field
4 octets 5 octets 5 octets 5 octets 5 octets
Reserved Video control 4 Video control 3 Video control 2 Video control 1
D.3.2.3.1 Video control field
The video control field structure is as presented in Table 53.
Table 53: Video control field
Bits: b39-b36 b35-b33 b32 b31-b16 b15-b0
57
58
Reserved Video frame number Interlaced field
indication H position V position
The V and H position fields contain, respectively, the vertical and horizontal position of the first pixel in the
sub frame, where (0, 0) is the topmost left position on the screen. The Interlaced field indicates, by being set to
'1', that the sub frame carries pixels from the bottom field. If the frame carries pixels from the top field or pixels
for non-interlaced video formats the interlaced field will be set to '0'. The video frame number is a counter that
indicates the video frame to which the pixels in the sub frame belong.
D.3.3 MAC frame body
The MAC frame body has two defined structures, for the unsecured and secure frames, respectively. These
are presented in Table 54 and Table 55.
Table 54: Non secure MAC body frame
4 octets Ln
FCS Frame payload
Table 55: Secure MAC body frame
4 octets 8 octets Ln 2 octets 2 octets
FCS Integrity code Secure payload
Secure Frame
counter SECID
Frame payload
D.3.4 Header Check Sequence (HCS)
The HCS is a four octet CRC and is defined as the one's complement of the modulo 2 sum of two terms:
Term a: the remainder resulting from ((xk*(x
31+x30+...)) divided (modulo 2) by polynomial G(x),
presented in (18). The value k is the number of bits in the calculation field, which is the number of bits in
the HRP header and Extended MAC header, 704 bits.
Term b: the remainder resulting from the calculation field contents, treated as a polynomial, is
multiplied by x32 and then divided by G(x).
The polynomial G(x) is defined as:
G(x)=x32+x26+x23+x22+x16+x12+x11+x10+x8+x7+x5+x4+x2+x+1 (18)
Annex E. Matlab/Simulink detailed block diagrams
Figure 22: Detailed HRP transmitter diagram
59
Figure 23: Detailed HRP receiver diagram
60
Figure 24: Detailed HRP frame conditioning block (transmitter)
Figure 25: Detailed HRP frame conditioning block (receiver)
61
62
Figure 27: Detailed HRP Reed-Solomon encoder block (transmitter)
Figure 26: Detailed HRP scrambler/descrambler block
Figure 28: Detailed HRP Reed-Solomon decoder block (receiver)
Figure 29: Detailed HRP outerinterleaver block (transmitter)
63
Figure 30: Detailed HRP multiplexer/bit interleaver block (transmitter)
Figure 31: Detailed HRP demultiplexer block (receiver)
64
Figure 32: Detailed HRP OFDM modulator block (transmitter)
Figure 33: Detailed HRP demodulator block (receiver)
65