An Integrated Overlay Architecture BasedMulti-GNSS Front-end

Alexander Rügamer∗, Santiago Urquijo∗, Markus Eppel∗, Heinrich Milosiu∗, Johannes Görner†, Günter Rohmer∗∗Fraunhofer IIS Nuremberg, Germany

†Technische Universität Dresden, Germany

Abstract—This paper presents the overall architecture, firsttest structure implementations, and measurement results of anintegrated GNSS front-end based on intentional path overlay. Thefront-end ASIC supports simultaneous multiband, multi-systemGNSS reception of GPS L5 / Galileo E5 / GLONASS G3 andGPS L1 / Galileo E1 / GLONASS G1 signals with up to 52 MHzbandwidth while using only one common baseband path thanksto an intentional analog signal overlay. Test structures of theRF and baseband parts were realized in a 1.8 V, 150 nm RF-CMOS technology packaged in a QFN48 housings with full ESDprotection. Both chips are described in detail regarding theirdesign and their actual measurement results.

Index Terms—Satellite navigation systems, Global PositioningSystem, Receivers, CMOS integrated circuits, GNSS front-end

I. INTRODUCTION

Global navigation satellite system (GNSS) receivers greatlybenefit from the modernization of existing GNSS constel-lations such as GPS and GLONASS as well as from thelaunch of new ones such as Galileo and COMPASS. First, thecombining of these constellations can significantly improvethe navigation solution availability in urban canyons andheavily shadowed areas. Second, increased satellite availabilitytranslates into higher measurement redundancy and improvedreliability. Additionally, the excellent inherent noise and multi-path mitigation capacity of the new and modernized widebandsignals notably improves accuracy in both measurement andposition domains.

Single-frequency users can receive GNSS signals and ser-vices but there are several advantages to multi-frequencyprocessing: The frequency diversity offers superior protectionagainst jamming and interference since, if one frequency bandis corrupted, the receiver is still able to provide a navigationsolution relying on another frequency band. Moreover a fasterreception of the navigation messages is often possible sincethe same information is transmitted on several bands (e.g.the Galileo I/NAV message broadcast on both E1B and E5B)using page swapping [1]. Finally, multi-frequency can be usedto form ionosphere-free pseudorange measurements that canremove the first-order ionospheric bias and therefore providea higher positioning accuracy.

The challenges of multiband reception are a much higherrequired bandwidth, higher sampling rates, often several re-ception chains, a higher digital bandwidth (the raw samplerate from the front-end output to the baseband signal process-ing) and more self-generated interferences (e.g. when several

frequency synthesizers for different local oscillator frequen-cies are needed). This all leads to a noticeable increase inreceiver complexity, size, and power consumption, especiallyfor integrated radio frequency (RF) front-ends.

Traditional GNSS front-ends but also current mass-marketGNSS receivers typically feature a low intermediate frequency(low-IF) architecture with an RF-bandwidth of approx. 2 to4 MHz and a low-resolution analog-to-digital converter (ADC)of 1 to 3 bit [2], [3]. This is sufficient for the legacy GPSL1 C/A or the narrowband Galileo E1 BOC(1,1) signals butnot for most of the new GNSS signals, especially if theirfull potential in terms of accuracy and multipath resistance isto be reached. For these, considerably larger bandwidths arenecessary (e.g. at least 16 MHz for the GPS/Galileo L1/E1MBOC(6,1,1/11), 20 MHz for L5/E5A BPSK(10) signals,and up to 52 MHz for processing the complete Galileo E5AltBOC(15,10) signal) which leads to higher sampling raterequirements.

The straightforward approach is to widen the bandwidthand to use higher sampling rates for each desired GNSSsignal while keeping the original low-IF architecture. Thesesolutions can already be found as integrated circuits and caneasily be tuned to the required GNSS signal band [4], [5], [6].However real multiband reception is only possible by adding acomplete extra receiver for each additional frequency band tobe received, either as a separate chip as proposed in [4] or byintegrating more or less several standalone receivers on one dieas proposed in [6] and [7]. This makes their implementationstraight forward but is inefficient especially for an integratedcircuit implementation.

For a wideband BOC signal such as the Galileo E5AltBOC(15,10) a zero-IF architecture can be very advanta-geous since the inherent zero-IF problems, namely DC-offsetand flicker noise, are not so relevant to the DC-free BOCsignals. Moreover, quasi zero-IF architectures can be usedto enable simultaneous reception of the L1/E1 GPS/Galileosignals and the GLONASS G1 frequency division multipleaccess (FDMA) signals by placing the local oscillator betweenboth signal bands [8].

All the architectures mentioned so far are generic in away that none exploit any properties of the GNSS signalscode division multiple access (CDMA) structure: the usefulsignals are below the thermal noise floor, spread with longpseudo-random noise (PRN) sequences, and have differentbandwidths. The complexity of a multiband RF front-end

978-1-4673-0387-3/12/$31.00©2012 IEEE 50

L1 SPSBPSK(1)

Inph

ase

Quad

ratur-

phas

e

Inph

ase

Quad

ratur-

phas

e

NAVSTAR-GPS

GLONASS

Inph

ase

Quad

ratur-

phas

e

Galileo

PPSBPSK(10)

BPSK(5.11)

E5a -BPSK(10)

1176.45 MHz

E5bBPSK(10)

1207.14 MHz

E5AltBOC(15,10)1191,795 MHz

E6bcBPSK(5)

E6aBOCcos(10,5)

BOC(2,2)

1176.45 MHz

BPSK(0.511)1598.0625 ….…1605.3750 MHz

E1aBOCcos(15,2.5)

E1bcCBOC(6,1,1/11)

L1C-IBOC(1,1)

M-CodeBOC(10,5) L1 M-Code

BOC(10,5)

L2CBPSK(0.511)

L1C-QTMBOC(6,1,4/33)

G1G3G5

BPSK(5.11)L1OC

L1SC L1OF

L1SF

L5

BPSK(10)

PPSBPSK(10)

1227.60 MHz1575.42 MHz1575.42 MHz

L2 L1

L5OC

BPSK(10)1202.025

L3OC

L3SC

G2

L2OF

L2SF

1575.42 MHz

BOC(2,2)

1278.75 MHz 1575.42 MHz

E5 E6 E1

L5/E5/G3 Band: ARNS and RNSS L1/E1/G1 Band: ARNS and RNSSL2 / E6 Band: RNSS only

BOC(4,4)1176.45 MHz

BPSK(0.511)1242.9375 ……1248.6250 MHz

Fig. 1. GNSS signals of the future

can be considerably reduced by sharing front-end componentsor stages using intentional signal overlay while exploitingthe aforementioned properties of the GNSS signals [9]. Theproposed overlay front-end architecture exploits the propertiesof the direct sequence spread spectrum GNSS signals byintentionally overlaying two signal reception paths. Thanksto the overlay, only one common baseband path is needed.This allows significant savings in terms of cost, size, powerconsumption, and digital bandwidth. The possibility to receivewideband signals with up over 52 MHz bandwidth enables tofully benefit from the inherent noise and multipath resistanceof the new Galileo AltBOC and GPS/Galileo MBOC signals.To the authors knowledge there are no publications of inte-grated GNSS front-end receiver chip measurements capableof processing the complete AltBOC signal.

The paper is organized as follows: First the selection ofappropriate frequency bands for this receiver type is explained.Then the overlay architecture with its frequency plan andsome implementation details is described. In the fourth sectionthe used semiconductor technology of the ASIC is brieflyintroduced before the architecture and measurement results ofboth the RF and the baseband part are discussed in detail.Finally, conclusions are drawn and the performance of thisoverlay front-end is compared with that of current state-of-the-art integrated multi-GNSS front-end implementations.

II. GNSS SIGNALS AND STATUS

Currently, two fully operational GNSS are available: theAmerican NAVSTAR-GPS and the Russian GLONASS sys-tem. Both systems are continuously evolving to improve thesignals and services they will provide in the coming years.Aside, two other GNSS are being developed - the European

Galileo and the Chinese COMPASS systems. Therefore, in afew years, at least four independent but interoperable GNSSwill be available to support an increasing range of applications.

Figure 1 shows the L-band spectrum of the current andplanned GNSS signals with the notation of their modulationnames and carrier frequencies. The red and green signals areclassified, the blue ones are open signals. All current andupcoming GNSS signals are within the protected Radio Navi-gation Satellite Services (RNSS) band but only the L1/E1/G1and L5/E5/G5/G3 bands are within the even better protectedspectrum allocated to Aeronautical Radio Navigation Services(ARNS). The other three GNSS bands (i.e. E6, G2, and L2)therefore suffer from radar, military transmissions and otherpotentially strong interferers.

The combination of both E5 (including the GPS L5,GLONASS L5OC and L3OC, Galileo E5a and E5b) and E1(featuring the GPS L1 C/A and L1C, GLONASS L1OC, andGalileo E1-B/C) signals is deemed the most appropriate for theadvanced open-service, multi-constellation, multi-frequencyGNSS receiver that is intended to be realized with thisintegrated GNSS overlay front-end. For fast acquisition the rel-atively narrowband L1/E1 signals (GPS C/A and Galileo E1-B/C with BPSK(1) and BOC(1,1) modulation, respectively)are typically used. The estimated Doppler and code delay canthen be used for high performance tracking with the widebandL1/E1 MBOC(6,1,1/11), L5/E5a BPSK(10), L5OC BOC(4,4)or E5b/L3OC BPSK(10) signals [10].

III. OVERLAY FRONT-END ARCHITECTURE

The front-end architecture consists of a zero-IF down-conversion path for the L5/E5/G3 bands and a double hetero-dyne low-IF path for L1/E1/G1 bands as shown in Figure 2.

51

LNA

LNA

IF_AMP_E5

IF_AMP_E5ADC

ADC

F_LO1_E1 =1192 MHz

F_LO_E5 =1192 MHz

E1_IF =-13.913 MHz

IF1_E1 =383.42 MHz

IF_AMP_E1

F_sampling = F_OSC = 74.5 MHz

Q

I

E5 PathHomodyne

E1 PathHeterodyne

F_LO2_E1 =397.333 MHz

Mixer 1

Mixer 2

/3

74.5 MHz

F_VCO =2384 MHz

FrequencySynthesizer

Combiner

I/QChange

E1VGA

E5VGA

E5_IF =0 MHz

0°

90°/2

CPCharge Pump

LFLoop Filter

VCOVoltage Crtl. Osc

/2/16

PDPhase

DetectorBuffer

F_LO_E5 =1192 MHz

F_LO2_E1 =397.333 MHz

Serial Data Interface

Digital Core 0°

90°/2

RF Part Baseband Part

AGC

VGA

Anti-Aliasing10-40 MHz

VGA

VGA I

VGA Q

Filter-BW

TCXO

Power Chip on / offPower E1 / E5 on / offI/Q ChangeCombiner VGA ControlAGC ControlADC Configuration

Fig. 2. Proposed overlay GNSS front-end architecture

The incoming L5/E5/G3 band is amplified and directly shiftedto baseband by a quadrature down-conversion mixer with alocal oscillator (LO) frequency of 1192 MHz and amplifiedagain by an IF gain block.

A double heterodyne architecture is selected for theL1/E1/G1 reception path to circumvent the image problem ofa pure low-IF architecture and to reuse the LO frequenciesas it will be explained in the following paragraph. In the firstmixing stage the RF signal is down-converted to an IF ofapprox. 400 MHz using a real mixer. The image frequencyis more than 750 MHz away from the RF and can thereforebe easily attenuated by an RF bandpass filter preceding thefront-end input. An IF gain block both amplifies the IF signalsand provides sufficient isolation to the next mixing stage. Thesecond mixing stage consists of a complex mixer which shiftsthe IF signals to a low-IF frequency, as shown in Figure 3.

Thanks to a judicious frequency plan all LO frequenciescan be directly derived with integer-N dividers from a singlevoltage controlled oscillator (VCO), as shown in the frequencysynthesizer block diagram in Figure 2. This architecture reusesthe E5-LO frequency for the first E1 mixing stage. The secondE1 mixing stage uses exactly one third of the first E5-LO

frequency. By using an off-chip reference oscillator of e.g.74.5 MHz the frequency synthesizer can be realized with astraight forward integer-N PLL type implementation.

The combiner overlays both (complex) IF-signals from theE5/L5/G3 path and the E1/L1/G1 path in the analog domain.This is possible since all GNSS signals are direct sequencespread spectrum based signals with a high spreading gain anda negative signal-to-noise ratio (SNR) before the correlation.Preceding the overlay, the signals can be relatively amplifiedor attenuated using a variable gain amplifier (VGA) combinercontrollable by the digital signal processing - in most cases theGNSS baseband receiver. Thanks to this combiner only onecommon baseband stage, consisting of an anti-aliasing lowpassfilter, an automatic gain control loop (AGC), and analog-to-digital-converters (ADC) is needed.

The downside of the intentional signal overlay is that thenoise from the overlayed signal bands is folded into thebaseband range and directly affects the signals C/N0. Using anappropriate power control before the overlay, the overlay noisecan be controlled and at least partly mitigated as explained in[11]. Moreover, due to the overlay, jammers present in onefrequency band also impact the other, previously undisturbed,

52

Galileo E5a/GPS L5 -BPSK(10)

1176.45 MHzE5b -

BPSK(10)1207.14 MHz

E5 Filter, BW=72 MHz

E5a/L5-15.14 MHz

E5b15.55 MHz

LO=1192 MHz 1575 MHz

0 MHz

E1 / L1-13.91 MHz

809 MHz

E1 ImageFrequency

GLONASSG1 FDMA

1602 MHz

complex E5 mixer andfirst real E1 mixer with

LO=1192 MHz

second complex E1 mixer withLO=1192/3 MHz= 397.33 MHzOverlay, and Anti-Aliasing LP

filter with 35 MHz

Galileo E1 / GPS L1 C/A

E5a/L5 -BPSK(10)

-15.14 MHz

E5b -BPSK(10)15.55 MHz

GLONASS G1+12.67 MHz

782 MHz

G1 ImageFrequency

E5a/L5

GLONASS410 MHz

E1 / L1383 MHz

LO=397.33 MHz

GLONASSG3 CDMABPSK(10)

1202.025 MHz

E5bG3

G3 CDMA10.025 MHz

E5a/L5 E5bG3

GLONASS G3+10.025 MHz

E5 - AltBOC(15,10)1191,795 MHz

Fig. 3. Signal down-conversion scheme

frequency band. But the advantage of being able to receive allthe open-service GNSS signals together with just one ADCis obvious and makes the overlay architecture interesting foradvanced mass-market multiband, multi-system receivers.

IV. SYSTEM AND IC DESIGN

For the integrated circuit implementation the new LF150process from the German foundry LFoundry was used. LF150features a 0.15 µm RF-CMOS process with up to 6 metal layersincluding a thick metal e.g. for high Q inductors and optimizedRF library devices. The complete chip was implemented in thisprocess using an 1.8 V operation supply.

To test and verify the different parts of this front-endarchitecture, its first implementation was split into three teststructures described in the following sections: A low noiseamplifier (LNA) and mixer ”pipe-cleaner”, an analog RF re-ceiver part, and a mixed signal baseband part. These blocks areshown in Figure 2. The pipe-cleaner die was directly bondedon a printed circuit board (PCB) without any packaging whilethe other two test structure chips were packaged in a QFN48

housing including full ESD protection and measured on adedicated PCB separately.

V. RF PART

It is assumed that an active antenna precedes the front-endchip. According to Friis’ Formula, the active antenna LNAconsiderably lowers the noise figure (NF) requirements for theon-chip LNAs. The analog RF receiver test structure uses twofully differential LNAs for the upper and lower band amplifica-tion, followed by complex Gilbert mixing stages, intermediateamplifiers, and test-buffers to make off-chip measurementspossible.

A. LNA, Mixer, and IF-Amplifier Architectures

For both the E1 and E5 band LNA a modified version ofthe fully differential topology described in [12] and [13] isused to significantly suppress on-chip interferences cominge.g. from the mixed signal baseband part on coupling over thesubstrate. The LNAs were not designed to achieve the lowestpower consumption or NF but mainly for functionality. TheLNAs consist of two stacked stages as shown in Figure 5. The

53

E5 LNA

E1 LNA

Mixer I

Mixer Q

AMP

AMP

BufBuf

1150 µm

1900

µm

Fig. 4. Chip photo of the RF receiver test structure

input stage is a cascode topology with transistors M1 to M4. Acascode topology is advantageous since it inherently providesa high reverse isolation [14] which is particularly important inthe E5 zero-IF architecture, where the LO signal of the mixermatches the LNA’s input frequency. For better input impedancematching an inductive source degeneration is used. Lookinginto the gates of the transistors M1 and M2 the inductors L1and L2 contributes to the real part of the impedance. Insteadof using two separate inductors a common tapped one is used,see the chip photo in Figure 4. With the series connection ofthe capacitor C1 and C2, and the real parts of the impedanceseen through the gates of M1 and M2, respectively a highpassRC-filter is formed. The inductors L3 and L4 are used to tunethe output of the first stage. The capacitor C5 together withL3 and L4 sets the resonance frequency to the signals carrierfrequencies. The differential amplifier between Vb and R1 andR2 adaptively controls the bias points and effectively mitigatesprocess, temperature, and supply variations. A common-sourceoutput stage of the LNA is formed by the transistors M5 andM6.

Both the first and the second mixers are realized in astandard Gilbert-Cell topology. Gilbert-Cells are a fully dif-ferential, double balanced topology providing a very high LOto IF isolation [14]. Being fully differential fits perfectly to thedifferential LNA output. The high isolation is very beneficialfor the intended E5 zero-IF architecture.

The IF-amplifiers are standard source coupled differentialpairs with a resistive load designed for a gain-bandwidthproduct (GWB) of approx. 10 GHz. A coarse lowpass filtering

M1 M2

M3 M4

M5 M6

L1 L2

L3 L4

C1 C2

C3 C4C5

R1 R2

R6 R7 R8 R9

Inp Inn

Op On

Ibias

+–

Vb

Fig. 5. LNA circuit diagram, modified version of [12], [13], some biasingdetails omitted

at the output is realized with first order RC-filters.

B. RF Part Measurement Results

These analog RF components are most critical and at thesame time crucial for the whole receiver functionality. Tocheck the behavior of the design kit’s RF-models a first”pipe-cleaner” run was done with test-structures of the LNAsand mixers already including full ESD protection. For themeasurements the originated chip-die was directly bonded ona PCB without any packaging.

In order to measure the differential LNA input with astandard single-ended network analyzer, an RF-balun was usedto convert between the single-ended and differential signaldomains. Moreover the LNAs were matched using a lumpedcomponent LC-network before the balun. The measurementresults of the input reflection coefficient S11 are shown inFigure 6 for both the lower band E5 and upper band E1 LNAwith plots of the theoretical, normalized power spectral density(PSD) of the GNSS signals to be received. It can be seenthat the 3 dB matched frequency bandwidth is greater than200 MHz for both LNAs, enabling the reception of all thelower and upper GNSS L-band signals.

The impedances on the chip components are generally notdesigned for 50Ω interfaces. The on-chip structures are sosmall in relation to the lower GHz frequencies in the L-band that any transmission line effects can be neglected.Therefore the on-chip LNAs outputs are not intended todrive a 50Ω load, e.g. given by the standard measurementequipment. These LNAs only need to drive the much higher

54

1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8

x 109

−25

−20

−15

−10

−5

0

Frequency [Hz]

S11

[dB

]

Measured LNA S11

Lower Band LNAUpper Band LNAE1 MBOC PSDGLONASS G1 PSDE5 AltBOC PSD

Fig. 6. Measured S11 of the lower band E5 and upper band E1 LNA withplots of the theoretical, normalized PSD of the GNSS signals to be received

impedance of the following active mixer stage. To enable off-chip measurements, a NMOS buffer with a transconductanceof 20 mS was designed at the output of the mixer leadingto 0 dB gain when a load impedance of 50Ω is applied.Designing such a buffer for the mixer output is facilitated bythe fact that the mixer output frequencies are within the IFrange and no longer are at RF.

Therefore, the gain of the LNA can only be measuredindirectly. First the gain of the active mixer was determinedto be 9.5 dB for both the E1 and E5 mixers using the NMOS-buffer as previously explained. The active mixer LO inputwas an external rectangle signal with a differential peak-to-peak amplitude of 700 mV. Then the complete chain includ-ing the LNA and mixer was measured. The implementationlosses of the balun and cables were deembedded. With thepreviously determined mixer gain, the LNA gain alone couldbe calculated. The results of the pipe-cleaner measurementsof the LNA and mixer are shown in Table I and fit to thesimulated values. The current consumption is around 3 mAfor the E5 LNA, and around 2 mA for the E1 LNA, which isalso consistent with the simulations.

The 1 dB input compression point of the combined LNAwith mixer measurement was determined to be −30 dBm forboth E1 and E5. In comparison to the linearity of off-the-shelf LNAs this may seem very low, but keeping in mind thatin an undisturbed environment only thermal noise within thebandwidth of interest is received (approx. −100 dBm) and thatthe front-end incorporates a 3 bit ADC which provides approx.18 dB of dynamic, even a −30 dBm input compression pointis sufficient.

Having validated the pipe-cleaner designs, the complete RFreceiver including the intermediate frequency amplifier wasmanufactured and packaged in the intended QFN48 housing. Aphoto of the chip is shown in Figure 4. Still the aforementionedbuffers are included in parallel to each component’s output forthis test structure to allow separate component measurements.

TABLE IMEASUREMENT AND SIMULATION RESULTS OF THE RF RECEIVER

COMPONENTSE5 LNA1 E1 LNA1 Mixer1 IF AMP2

Gain sim. [dB] 20.0 15.0 10.0 15.0Gain meas. [dB] 19.0 12.0 9.5 16.0

NF sim. [dB]

Dig.Core

ADC

AGC+VGA

Lowpass Filter

Combiner

700 µm

1300

µm

Fig. 7. Chip photo of the baseband part

and turn off various blocks and components of the chip. Itsfunctionality is summarized in Table III. The digital core notonly allows an effective debugging and testing of the chip, italso gives to the GNSS receiver the possibility to modify orcontrol the front-end hardware according to its environmente.g. switching off one path if a severe interferer is detected orcontrolling the combiner gain values to minimize the overlayloss as described in [11].

2) Combiner: As shown in the combiner block diagram inFigure 2, before the E1 and E5 signal paths are overlayed theycan be amplified or attenuated by the combiner VGA cell. Thisblock can be controlled in 64-steps (6 bit) via the SPI-bus ofthe digital core. Each signal path has a tunable 25 dB dynamicrange from −10 to 15 dB, see Figure 8. In order to improvethe area implementation of the cell, the steps show a non-linear characteristic depicted as differential gain in Figure 8.Since the combiner VGAs should be regulated by an adaptivecontrol algorithm running on the baseband GNSS receiver, thisnon-linearity is irrelevant.

According to the frequency down-conversion scheme de-picted in Figure 3 the E1 signal overlays the E5a/L5 mainlobe. If the RF-filter exclusively selects the E1 frequency band,only the left main-lobe of E5 (in essence the Galileo E5a andGPS L5 signals) is affected by the overlay while the right

TABLE IIIDIGITAL CORE FUNCTIONALITY

Logical Block Component MethodLNA E1 Power Down

E5 Power DownMixer E5 I and/or Q Power Down

First E1 I and/or Q Power DownSecond E1 I and/or Q Power Down

IF-amplifier E5 I and/or Q Power DownE1 I and/or Q Power Down

Combiner I/Q Change Mirror E1 SpectrumVGA E1 6 bit −10 to 15 dBVGA E5 6 bit −10 to 15 dBAdder I and/or Q Power Down

AGC Low-pass Filter Power DownADC VGA I and/or Q Power DownInternal AGC Loop Power DownControl Method CW or GNSS-noisePulse Blanking on / offOverload Control on / off

ADC I and/or Q Path Power DownADC Buffer Power DownResolution 3 or 2 bitOutput FormatOutput Driver Strength

10 20 30 40 50 60−10

−5

0

5

10

15

20

Com

bine

r G

ain

[dB

]

Combiner Values (6 bit)10 20 30 40 50 60

0

1

2

Diff

eren

tial G

ain

[dB

]

Fig. 8. Simulation of the combiner block digital programmable gain for bothE1 and E5 paths

main-lobe comprising Galileo E5b remains untouched. Thiscan also be inverted by changing either the polarity of oneE1 signal I/Q component (swapping + and - signals in thein-phase and quadrature branch) or its phase (swapping thein-phase and quadrature branch) before the combiner stage.Doing so leads to a mirroring of the E1 signal around thezero frequency axis. Depending on the intended application,this method can be beneficial since the overlay loss can becompletely switched off on either E5a or E5b assuming anadequate E1 signal RF bandpass filter.

3) Anti-Aliasing Lowpass Filter: A 4th order active lowpassfilter was realized in a Chebyshev topology using two op-amps and RC-networks. The default 3 dB cutoff frequencyis around 33 MHz. Since the bandwidth setting integratedRC values vary by ±20% due to CMOS process variationsa self-calibration was realized with an external reference

56

0 5 10 15 20 25 30 35 40 45 50−25

−20

−15

−10

−5

0

5

10

15

Frequency [MHz]

Am

plitu

de R

espo

nse

[dB

]

120 kΩ115 kΩ110 kΩ105 kΩ100 kΩ95 kΩ90 kΩ85 kΩ80 kΩ75 kΩ70 kΩ65 kΩ60 kΩ55 kΩ50 kΩ45 kΩ40 kΩ

Fig. 9. Simulation results of the low-pass filter bandwidth together with thecombiner in dependency of an external reference resistor

resistor. This external resistor can also be used to adjust thefilter bandwidth between 3 to 35 MHz. Figure 9 shows thesimulation result of the filter bandwidth in dependency of theexternal reference resistor value swept from 40 to 120 kΩ.

4) Automatic Gain Control: The automatic gain controlensures that the ADC’s input signals are within the optimalrange. The AGC senses the distribution of the ADC bits andsets the ADC VGA gain appropriately. The control algorithmscan be configured to achieve better performance in presence ofCW signals or for a minimum C/N0 loss when GNSS-noiselike signals are applied. The dynamic range of the VGAs is−12 to 40 dB according to the simulations.

5) Pulse Blanking: The baseband chip is equipped with anenergy detector setting a blanking flag output signal if ADCclipping is occurring. Moreover, the ADCs can be configuredto blank the signal if the blanking flag is active. This featureenables pulse blanking mitigation at a very early stage makingit very effective. Finally, the AGC can be protected from beingstimulated from pulses using the overload control switch.

6) ADC: The analog-to-digital-converters (ADC) digitizethe output differential signals of the variable gain amplifiers.Typical GNSS receivers use low bit quantization. The im-plemented converters employ 3 bit which is a good trade-off between complexity and low implementation loss. Dueto the low number of bits requirement, the classical flashtopology was selected. The cell includes a buffer, a resistorladder, comparators and a digital core cell for decoding. Thecomponents are optimized in terms of current consumption andspeed. The comparators include an amplification stage and adecision stage. The amplification stage reduces the couplingto the input signal of the noise generated at the decision stage.The decision stage consumes current only at the rising edgeof the clock signal. The current consumption of the differentstages is minimized in order to preserve the ADCs linearity.The resistor ladders of the ADCs are placed together in thelayout in order to reduce the mismatch.

Fig. 10. Photo of the baseband part test PCB

B. Baseband Measurement Results

The baseband functionality was measured and verified withthe test board shown in Figure 10. The PCB provides two com-plex input signals (E5 I/Q, E1 I/Q) which are converted fromsingle-ended to differential with an external balun including abias point. Moreover some output buffers for the clock input,the digital ADC outputs, and SPI interface are employed asan interface for the measurement equipment.

It has to be noted that due to the baseband part teststructure implementation, it is not possible to test single blocksseparately. Therefore, all reported test results were measured atthe dual-ADC outputs using input signals at the combiner. Themeasurement results include all effects from all the basebandpart components (from combiner to ADCs).

1) Combiner: To test the combiner with its E1 and E5VGAs, two complex GNSS baseband signals - one 16 MHzbandlimited E1 MBOC and one E5 AltBOC signal - weregenerated and fed to the baseband part inputs with a highSNR to be able to identify their main-lobes in the plots. TheE1 signal was shifted to an IF of −13.91 MHz to be compliantwith the frequency down-conversion scheme proposed in Fig-ure 3. Using the combiner VGAs, the E1 signal amplificationwas varied from 13.5 to −5.3 dB while the E5 combiner VGAwas hold constantly on 2 dB. Figure 11 shows the overlayedsignals with different E1 combiner VGA settings. Since theAGC automatically amplifies the overlay output signal to fitto the optimum ADC range, the PSDs shown in Figure 11were scaled in a way that the E5b main lobe is approximatelyalways at the same level. This enables a better comparison ofthe combiner VGA effect. It can be concluded, that the overlayworks appropriately and that sufficient overlay control of thepath is provided by the implemented combiner VGAs.

2) E1 I/Q Change: The blue colored plot in Figure 12shows the recorded PSD when both paths are simultaneouslyactivated with intentional overlay of the E1 MBOC and E5aBPSK signals according to the frequency down-conversionscheme proposed in Figure 3. As the blue colored plot shows,the left main-lobe of E5a is completely affected by the overlaywhile the right side remains untouched. Activating the E1 I/Qchange feature, only the E1 path is mirrored around the zerofrequency axis. The PSD with activated E1 I/Q change is

57

−25 −20 −15 −10 −5 0 5 10 15 20 25−95

−90

−85

−80

−75

−70

−65

−60

Frequency [MHz]

Sca

led

Pow

er S

pect

ral D

ensi

ty [d

B/H

z]

Fig. 11. Measured overlay PSD of E1 and E5 signals in dependency of theE1 combiner VGA settings

−25 −20 −15 −10 −5 0 5 10 15 20 25−95

−90

−85

−80

−75

−70

−65

−60

Frequency [MHz]

Pow

er S

pect

ral D

ensi

ty [d

B/H

z]

E1 and E5a overlayedE1 and E5a overlayed with I/Q change

Fig. 12. Measurement of E1 I/Q change activation

depicted in the same Figure in green color and shows thatthe E5a main-lobe is now no longer affected by the overlay.

3) AGC and ADC VGAs: The AGC loop continuouslymodifies the ADC VGAs control voltage to guarantee anoptimal ADC input power. The implemented circuitry enablesto control the VGAs (I and Q) independently or together, asshown in Figure 2. The voltages at the I and Q VGA controlpins were measured while the input power was swept. TheAGC loop controls the VGAs in a way that the ADC inputpower is kept constant. Therefore, the VGAs dynamic rangecan be measured observing the range of the control voltagesto be approx. 50 dB between the signal input powers of 5 to−45 dBm as shown in Figure 13. However, it should be notedthat for normal operation the I and Q VGA control pins areconnected together. Nevertheless Figure 13 also demonstratesthe excellent matching of the I and Q baseband paths.

Moreover, the combiner VGAs were set to their minimumvalue (approx. −10 dB) and afterwards to their maximumvalue (approx. 15 dB). The combiner VGAs dynamic range

−50 −40 −30 −20 −10 0 10450

500

550

600

650

700

750

800

850

900

950

1000

Pin [dBm]

AD

C V

GA

Vol

tage

[mV

]

VGA I; Combiner VGA=min; AGC=noiseVGA I; Combiner VGA=min; AGC=CWVGA I; Combiner VGA=max; AGC=noiseVGA I; Combiner VGA=max; AGC=CWVGA Q; Combiner VGA=min; AGC=noiseVGA Q; Combiner VGA=min; AGC=CWVGA Q; Combiner VGA=max; AGC=noiseVGA Q; Combiner VGA=max; AGC=CW

Fig. 13. Measured ADC VGAs values of the I and Q paths in dependencyof the input power

0 0.5 1 1.5 2 2.5 3 3.5 4

x 107

−70

−60

−50

−40

−30

−20

−10

0

Frequency [Hz]

Am

plitu

de [d

B]

BW=39.801 MHz, SNDR=18.2247 dB; ADC 3−bit ENOB=2.7346 bit

Fig. 14. Measured dual-ADC output spectrum with 10 MHz input tone anda sampling frequency of 80 MHz

can also be confirmed by measuring the offset between theresults with minimum and maximum combiner VGAs settings.

And finally, the AGC control method was changed fromCW-mode to the GNSS-noise mode. In CW-mode more clip-ping of the ADCs occurs. Therefore the VGAs gain in CWAGC mode is always approx. 5 dB higher than in the GNSS-noise mode.

4) ADC: Figure 14 shows the measurement results ofthe signal-to-noise-and-distortion-ratio (SNDR) of the wholebaseband part at nominal conditions. A 10 MHz input signalwas applied to the test-chip and sampled with 80 MHz. UsingEquation 1 for a sinusoidal input signal, the effective numberof bits (ENOB) was calculated to be 2.74 bit which is inaccordance with the simulation results.

ENOB = (SNDR − 1.76 dB)/6.02 dB (1)Although the receiver architecture shown in Figure 2 has anominal dual-ADC sampling rate of 74.5 MHz, different tests

58

TABLE IVCOMPARISON WITH PREVIOUSLY PUBLISHED INTEGRATED MULTIBAND GNSS FRONT-ENDS[4] [15] [5] [7] [6] [8] this work

Process 130 nm CMOS 180 nm CMOS 65 nm CMOS 130 nm CMOS 180 nm CMOS 65 nm CMOS 150 nm CMOS1st GNSS band L1/E1 L1/E1 L1/E1 L1/E1 L1/E1/B1, L1/E1/B1, L1/E1/G1

2nd GNSS band L5/E5aa L5/E5a b G1 L5/E5ac L5/E5a, B2/E5b B2/E5b L5/E5a, E5b/G33rd GNSS band - - - - L2 c,d L2 c,d E5

Bandwidth 14/20 MHz 9 MHz 2/4/8 MHz 4.53/24 MHz 2/4/20 MHz 2 to 8 MHz 52 MHzADC resolution 2 bit I/Q , n.a. 2x 3 bit 2 bit 4 bit I/Q 2 to 4 bit I/Q 3 bit

ADC sampl. rate 66.188 MHz 24 MHz 32.736 MHz 49.104 MHz 62 MHz n.a. 74.5 MHzSize 2.89 mm2 a 16.0 mm2 4.65 mm2 11.4 mm2 7.2 mm2 10.5 mm2 6.76 mm2 e

a One chip needed for each GNSS band to be receivedb Switching receiver architecture; only one band can be received at the same timec Integration of two almost independent reception chainsd Only two GNSS bands can be received simultaneouslye Die size of complete receiver including frequency synthesizer

showed a reliable maximum ADCs sampling rate of over150 MHz.

VII. CONCLUSION

The presented RF front-end parts enable the simultaneousreception of GPS L5 / Galileo E5 / GLONASS G3 and GPS L1/ Galileo E1 / GLONASS G1 signals with broad bandwidth. Byusing the proposed frequency plan and sharing the basebandparts with an overlay concept, the underlying architectureenables an efficient implementation in terms of digital datarate, chip size, and power consumption.

In comparison to other recently published integrated multi-GNSS front-end implementations (see Table IV) the receptionbandwidth has been significantly increased. The presentedimplementation supports the first time the reception of thecomplete 52 MHz bandwidth Galileo E5 AltBOC(15,10) signalin an integrated front-end enabling to fully benefit from theinherent noise and multipath resistance of this sophisticatedmodulation scheme.

For future work, the presented RF and baseband parts willbe revised and combined with the frequency synthesizer toform a fully integrated multi-GNSS front-end ready to coher-ently and simultaneously receive most open GNSS signals.The currently foreseen layout of the complete receiver chipincluding the frequency synthesizer has a die size of 6.76 mm2.The already small implementation size shows the potentialof this overlay architecture as compared to other recentlypublished integrated multi-GNSS front-end architectures.

ACKNOWLEDGMENT

This work was supported by a Fraunhofer IIS internalresearch program (IIFOP). The authors express their appre-ciations to all members of this project.

REFERENCES

[1] “European GNSS (Galileo) Open Service Signal In Space InterfaceControl Document (OS SIS ICD) Issue 1,” European Union/EuropeanGNSS Supervisory Authority (GSA), Tech. Rep., February 2010.

[2] “Fully integrated RF front-end receiver for GPS applications, STA5620Data Sheet,” STMicroelectronics, 2008.

[3] “Single-Chip Global Positioning System Receiver Front-End, MAX2742Data Sheet,” Maxim, 2008.

[4] G. Wistuba, A. Vasylyev, S. Haas, J. Driesen, L. Dalluge, D. Peitscher,B. Kramer, C. Dohmen, and F. Henkel, “A highly integrated configurableGNSS receiver frontend design for high bandwidth operation on E1/L1and E5a/L5,” in Localization and GNSS (ICL-GNSS), 2011 InternationalConference on, June 2011, pp. 164 –168.

[5] G. Rivela, P. Scavini, D. Grasso, M. Castro, A. Calcagno, G. Avellone,A. Di Mauro, G. Cali, and S. Scaccianoce, “A low power RF front-end for L1/E1 GPS/Galileo and GLONASS signals in CMOS 65nmtechnology,” in Localization and GNSS (ICL-GNSS), 2011 InternationalConference on, June 2011, pp. 7 –12.

[6] D. Chen, W. Pan, P. Jiang, J. Jin, J. Wu, J. Tan, C. Lu, and J. Zhou,“Reconfigurable dual-channel tri-mode all-band RF receiver for nextgeneration GNSS,” in Proceedings of the Solid State Circuits Conference(A-SSCC) 2010 IEEE Asian, 2010.

[7] Y. Moon, S. Cha, G. Kim, K. Park, S. Ko, H. Park, J. Park, and J. Lee, “A26mW dual-mode RF receiver for GPS/Galileo with L1/L1F and L5/E5abands,” in SoC Design Conference, 2008. ISOCC ’08. International,vol. 01, Nov. 2008, pp. I–421 –I–424.

[8] N. Qi, Y. Xu, B. Chi, Y. Xu, X. Yu, X. Zhang, and Z. Wang, “Adual-channel GPS/Compass/Galileo/GLONASS reconfigurable GNSSreceiver in 65nm CMOS,” in Custom Integrated Circuits Conference(CICC), 2011 IEEE, Sept. 2011, pp. 1 –4.

[9] A. Ruegamer, S. Urquijo, and G. Rohmer, “Multi-band GNSS Front-end Architecture Suitable for Integrated Circuits,” in Proceedings of the2010 International Technical Meeting of The Institute of Navigation -ION ITM 2010, January 25 - 27, 2010, San Diego, CA, 2010.

[10] C. Mongredien, M. Overbeck, and G. Rohmer, “Development andIntegration of a Robust Signal Tracking Module for the Triple-FrequencyDual-Constellation GAMMA-A Receiver,” in Proceedings of the 23thInternational Technical Meeting of the Satellite Division of the Instituteof Navigation, ION GNSS 2010, Portland, Oregon, September 20-24,2010.

[11] A. Ruegamer, C. Mongredien, S. Urquijo, and G. Rohmer, “OptimalPath-Control for Dual-Frequency Overlay GNSS Receivers,” in Local-ization and GNSS (ICL-GNSS), 2011 International Conference on, June2011, pp. 158 –163.

[12] A. Shahani, D. Shaeffer, and T. Lee, “A 12-mW wide dynamic rangeCMOS front-end for a portable GPS receiver,” Solid-State Circuits, IEEEJournal of, vol. 32, no. 12, pp. 2061–2070, 1997.

[13] D. Shaeffer, A. Shahani, S. Mohan, H. Samavati, H. Rategh, M. delMar Hershenson, M. Xu, C. Yue, D. Eddleman, and T. Lee, “A 115-mW,0.5-µm CMOS GPS receiver with wide dynamic-range active filters,”Solid-State Circuits, IEEE Journal of, vol. 33, no. 12, pp. 2219–2231,1998.

[14] F. Ellinger, Radio Frequency Integrated Circuits and Technologies.Springer, 2007.

[15] M. Detratti, E. Lopez, E. Perez, and R. Palacio, “Dual-Band RF Front-End Solution for Hybrid Galileo/GPS Mass Market Receivers,” inConsumer Communications and Networking Conference, 2008. CCNC2008. 5th IEEE, Jan. 2008, pp. 603–607.

59

Main MenuAttendeesTable of Contents2012 IEEE/ION Position, Location and Navigation SymposiumSession A1: High-Performance Inertial Sensor TechnologiesIntelligent Polynomial Curve Fitting for Time-Domain Triggered Inertial DeviceAngular Motion and Attitude Estimation Using Fixed and Rotating Accelerometers ConfigurationBreakthrough in High Performance Inertial Navigation Grade Sigma-Delta MEMS AccelerometerHigh Performance Gyro with Fast Startup Time, High Range, Wide Bandwidth, Low Noise and Excellent Vibration Immunitymilli-HRG Inertial Navigation SystemAtom Interferometric Gravity Gradiometer System

Session B1: Receiver & Antenna Technology IA High Sensitivity VDFLL Utilizing Precise Satellite Orbit/Clock and Ionospheric ProductsAdvanced Signal Processing Scheme for GNSS Receivers Under Ionospheric ScintillationAn Integrated Overlay Architecture Based Multi-GNSS Front-endCombined Doppler and Time-Free Positioning Technique for Low Dynamics ReceiversE-textile Microstrip Patch Antennas for GPSGeneralization of the Code Tracking Performance Theory to Receivers using Non-Standard Local ReplicasImplementation of a Multi-constellation and Multi-mode Navigation Terminal Equipment

Session C1: Indoor Personal & First-Responder NavigationRapid Mesh Network Setup for Indoor RF TrackingGLANSER: Geospatial Location, Accountability, and Navigation System for Emergency Responders - System Concept and Performance AssessmentScenario-based Evaluations of High-accuracy Personal Positioning SystemsMaps-based Angular PDFs used as Prior Maps for FootSLAMOpportunistic Radio SLAM for Indoor Navigation using Smartphone SensorsUnscented Kalman Filter and Magnetic Angular Rate Update (MARU) for an Improved Pedestrian Dead-ReckoningFoot-Mounted INS for Everybody-An Open-Source Embedded Implementation

Session D1: Aviation Positioning & Navigation Applications IInvestigation of APNT Optimized DME/DME Network Using Current State-of-Art DMEs: Ground Station Network, Accuracy, and CapacityTropospheric Error Modeling for High Integrity Airborne GNSS NavigationTesting of a Statistical Approach to Local Ionospheric Disturbances DetectionDifferential RAIM and Relative RAIM for Orbit Ephemeris Fault DetectionGNSS-based Curved Landing Approaches with a Virtual ReceiverPerformance Study of a New Crosslink-aided User Range Accuracy (URA) Integrity Monitor Algorithm for LPV-200A New Cascade Method for Detecting GPS Multiple Outliers Based on Total Residuals of Observation EquationsEnhanced Low Visibility Operations - Increasing Flight Operations Services in the National Airspace Systems in Low Visibility Conditions

Session A2: Low-Cost Inertial Sensor TechnologiesHigh-Performance, Low Cost Inertial MEMS: A Market in Motion!Parameter Study of Loosely Coupled INS/GNSS Integrity PerformanceCharacterization and Control of a High-Q MEMS Inertial Sensor using Low-cost HardwareA Novel Optimal Redundant Inertial Sensor Configuration in Strapdown Inertial Navigation SystemHigh-Q and Wide Dynamic Range Inertial MEMS for North-Finding and Tracking ApplicationsNovel Mismatch Compensation Methods for Rate-integrating GyroscopesMiniature IMU/INS with Optimally Fused Low Drift MEMS Gyro and Accelerometers for Applications in GPS-denied Environments

Session B2a: Atmospheric Effects & ModelingEvolution to Modernized GNSS Ionospheric Scintillation and TEC MonitoringTime-Frequency Analysis of Ionosphere Scintillations Observed by A GNSS Receiver Array

Session B2b: Receiver & Antenna Technology IIAdaptive Code Tracking Loop Design for GNSS ReceiversTowards a Robust Multi-Antenna Mass Market GNSS Receiver

Session C2: Urban Personal & Vehicular NavigationParticle Filter Based Positioning with 3GPP-LTE in Indoor EnvironmentsSystem Parameters Testing Methodology for ITS Applications of GNSS SystemsTracking in Dynamic Anchorless Wireless Networks Based on Manifold FlatteningA Robust Pedestrian Dead-reckoning Positioning Based on Pedestrian Behavior and Sensor ValidityCooperative Navigation in Transitional EnvironmentsA Near-Far Effect Canceller for GPS High Sensitivity Receiver

Session D2: Aviation Positioning & Navigation Applications IIInterference Characteristics for the Civil Airport Environment using Time-Frequency AnalysisMultipath Study On The Airport SurfaceAlgorithm for the Detection and Isolation of Systemic Measurement Errors in Vision Based/Aided Navigation SystemsReal-time Estimation of Projectile Roll Angle using Magnetometers: In-flight Experimental ValidationRadar Altimeter as a Navigation Aid Using Hierarchical Elevation Map ClusteringThe GNSS Signal Phase Measurement to Determine the Trajectory Disturbances for a Small Air Vehicle

Session A3: Multisensor Integrated Systems & Sensor Fusion Technologies IAn Adaptive Unscented Kalman Filter for Tightly Coupled INS/GPS IntegrationApproaches to Evaluate and Improve Short-term Relative Accuracy of GPS/INS SystemsIndoor Positioning Through Integration of Optical Angles of Arrival with an Inertial Measurement UnitRobust Height Tracking by Proper Accounting of Nonlinearities in an Integrated UWB/MEMS-based-IMU/Baro SystemRange Sensor Aided Inertial Navigation Using Cross Correlation on the Evidence GridThe IMRE Kalman FILTER - A New Kalman Filter Extension for Nonlinear ApplicationsSINS/GPS/CNS Information Fusion System Based on Improved Huber Filter with Classified Adaptive Factors for High-speed UAVs

Session B3: Interference, Spectrum Issues & Robust NavigationCompatibility Analysis Between LightSquared Signals and L1/E1 GNSS ReceptionDevelopment and Demonstration of a TDOA-Based GNSS Interference Signal Localization SystemA USRP-Based GNSS and Interference Signal Generator and Playback SystemGNSS Spoofing Detection in Handheld Receivers Based on Signal Spatial CorrelationMulti-Satellite Time-Delay Estimation for Reliable High-Resolution GNSS ReceiversA Multiple-frequency GPS Software Receiver Design Based on a Vector Tracking Loop

Session C3: Vision/Integrated Navigation Systems for Indoor ApplicationsEfficient Outlier Removal in Vision Based NavigationImproved Fusion of Visual Measurements Through Explicit Modeling of OutliersInertial and Imaging Sensor Fusion for Image Aided Navigation with Affine Distortion PredictionExploiting Ground Plane Constraints for Visual-aided Inertial NavigationFusing Visual Tags and Inertial Information for Indoor NavigationCHAMELEON: Visual-inertial Indoor NavigationOptical Flow Measurement of Human Walking

Session D3: Consumer Positioning & Smartphone Navigation TechnologyUsing Inertial Sensors of iPhone 4 for Car NavigationWiFi AP Position Estimation using Contribution from Heterogeneous Mobile DevicesAn Empirical Solar Radiation Pressure Model for Autonomous GNSS Orbit PredictionUsing Unlocated Fingerprints in Generation of WLAN Maps for Indoor Positioning3D Personal Navigation in Smart Phone using Geocoded ImagesAn Adaptive Map-Matching Based on Dynamic Time Warping for Pedestrian Positioning using Network Map

Session A4: Multisensor Integrated Systems & Sensor Fusion Technologies IIPerformance Investigation of Barometer Aided GPS/MEMS-IMU IntegrationUse of Magnetic Quasi Static Field (QSF) Updates for Pedestrian NavigationA Quaternion-based Initial Orientation Estimation System Suitable for One Special Dynamic State Using Low Cost Magnetometers and AccelerometersLoosely-Coupled GPS-INS State Estimation in Precision ProjectilesModeling with Noises for Inertial Sensors:Circular Data Processing Tools Applied to a Phase Open Loop Architecture for Multi-Channels Signals Tracking:

Session B4: Precise Positioning, Multlipath Mitigation & Advanced Processing AlgorithmsFixed Ambiguity Precise Point Positioning (PPP) with FDE RAIMHigh-precision GNSS Orbit, Clock and EOP Estimation at the United States Naval ObservatoryIntegrity Risk of Cycle Resolution in the Presence of Bounded FaultsSynthetic Aperture Navigation Algorithms Applied to a Driving User in Multipath EnvironmentsBounding Integrity Risk Subject to Structured Time Correlation Modeling UncertaintyUrban Multipath Detection and Mitigation with Dynamic 3D Maps for Reliable Land Vehicle LocalizationA Novel Bayesian Ambiguity Resolution Technique for GNSS High Precision PositioningTrajectory Estimation Improvement Based on Time-series Constraint of GPS Doppler and INS in Urban AreasM-Epoch Ambiguity Resolution Technique for Single Frequency Receivers with INS Aid

Session C4: Vision/Integrated Navigation for Vehicular & Robotic ApplicationsReal-time Implementation of Visual-aided Inertial Navigation Using Epipolar ConstraintsCollaborative Image Navigation Simulation and Analysis for UAVs in GPS Challenged ConditionsFilter-Based Calibration for an IMU and Multi-Camera SystemLane Detection Based on a Visual-Aided Multiple Sensors PlatformBundle Adjustment Without Iterative Structure Estimation and its Application to NavigationAn Efficient Refinement for Relative Pose Estimation with Unknown Focal Length from Two ViewsGraph-Based Cooperative Navigation Using Three-View Constraints: Method ValidationArchitecture for Asymmetric Collaborative Navigation

Session D4: Maritime Positioning & Navigation ApplicationsShip Attitude Accuracy Trade Study for Aircraft Approach and Landing OperationsPrecise Determination of Sediment Dynamics Using Low-cost GPS FloatersHeading-determination using the Sensor-fusion Based Maritime PNT UnitA New Method of Initial Alignment and Self-calibration Based on Dual-axis Rotating Strapdown Inertial Navigation SystemAn Automatic Compensation Method for States Switch of Strapdown Inertial Navigation SystemFour-position Drift Measurement of SINS Based on Single-axis Rotation

Session A5: Sensor Manufacturing, Error Modeling & TestingStatistical Modeling of Rate Gyros and AccelerometersHoneywell Micro Electro Mechanical Systems (MEMS) Inertial Measurement Unit (IMU)Inertial Sensors: Further Developments in Low Cost Calibration and TestingA Framework for Inertial Sensor Calibration Using Complex Stochastic Error ModelsEstimation of Deterministic and Stochastic IMU Error ParametersAnalysis of Coning Motion Caused by Turntable's Vibration in Rotation Inertial Navigation System

Session B5: Modernized GNSSCharacterizing the GNSS Correlation Function using a High Gain Antenna and Long Coherent Integration - Application to Signal Quality MonitoringSpectral Compatibility of BOC(5,2) Modulation with Existing GNSS SignalsRange Domain Auto-regression Method on Generalized Multivariate Time-varying Series AnalysisFMT Signal Options and Associated Receiver Architectures for GNSSStaggered InterplexOn Generalized Multidimensional Geolocation Modulation Waveforms

Session C5: Weak Signal ProcessingA Novel Quasi Open Loop Frequency Estimator for GNSS Signal TrackingA Weighted Combing Method for GNSS Antenna DiversityCharacterization of the Impact of Indoor Doppler Errors on Pedestrian Dead ReckoningTechnique for MAT Analysis and Performance Assessment of P2P Acquisition EnginesNon-Coherent, Differentially Coherent and Quasi-Coherent Integration on GNSS Pilot Signal Acquisition or Assisted AcquisitionThe Use of High Sensitivity GPS for Initialization of a Foot Mounted Inertial Navigation System

Session D5: Terrestrial Positioning & Navigation ApplicationsA Location-Based Services System (LBSS) Designed for an ArboretumIncreased Ranging Capacity using Ultra-Wideband Direct-path Pulse Signal Strength with Dynamic RecalibrationUnknown Source Localization using RSS in Open Areas in the Presence of Ground ReflectionsPerformance Analysis of a Civilian GPS Position Authentication SystemCollaboration-Enhanced Receiver Integrity Monitoring with Satellite EstimationUsing the Microsoft Kinect for 3D Map Building and TeleoperationIntegrated INS/GPS for Camera Tracking in Large Scale Outdoor Cinematic Applications

Session A6: Emerging & Alternative Sensor Technologies & Precision Timing SystemsA Consistent Zero-Configuration GPS-Like Indoor Positioning System Based on Signal Strength in IEEE 802.11 NetworksA Segmentation-based Radio Tomographic Imaging Approach for Interference Reduction in Hostile Industrial EnvironmentsImprovement of TERCOM Aided Inertial Navigation System by Velocity CorrectionCompact Atomic Magnetometer for Global Navigation (NAV-CAM)Cold Atom Micro Primary Standard (CAMPS)Factors Influencing the Noise Floor and Stability of a Time Domain Switched Inertial DeviceDevelopment of a Test Bed for UWB Radio Indoor Localization of First Responders

Session B6: GNSS Augmentation SystemsAutomated Verification of Potential GPS Signal-In-Space Anomalies Using Ground Observation DataAssessment of Integrity Monitoring Test-bed Module for the Airport EnvironmentDetecting Ionospheric Gradients for GBAS Using A Null Space MonitorDetermining and Measuring the True Impact of C/A Code Cross-correlation on Tracking - Application to SBAS GeorangingAutomated Determination of Fault Detection Thresholds for Integrity Monitoring Algorithms of GNSS Augmentation SystemsThe Application of the Time Series Theory to Processing Data from the SBAS Receiver with Safety of Life ServiceGPS-III L1C Signal Reception Demonstrated on QZSS

Session C6: Terrestrial Radionavigation & RF-PositioningASF Quality Assurance for eLoranA Terrestrial Positioning and Timing System (TPTS)WPI Precision Personnel Locator: Inverse Synthetic Array Reconciliation TomographyJoint Location and Parameter Tracking of Mobile Nodes in Wireless NetworksAccuracy Indicator for Fingerprinting Localization SystemsConstructing a Continuous Phase Time History from TDMA Signals for Opportunistic NavigationNext Generation Low Frequency Solutions for Alternative Positioning, Navigation, Timing, and Data (PNT&D) Services and Associated Receiver TechnologyRSSI Ranging Model and 3D Indoor Positioning with ZigBee Network

Session D6: Robotic Positioning, Navigation, Control & Sensor TechnologyAn Analysis of Observability-constrained Kalman Filtering to Vision-aided NavigationCharacterization of Non-Line-of-Sight (NLOS) Bias via Analysis of Clutter TopologyIn-Cabin Localization Solution for Optimizing Manufacturing and Maintenance Processes for Civil AircraftMobile ad-hoc Communication in Machine Swarms for Relative Positioning Based on GNSS-raw Data ExchangeLow Cost IMU Based Indoor Mobile Robot Navigation with the Assist of Odometry and Wi-Fi Using Dynamic ConstraintsStudy on the Use of Microsoft Kinect for Robotics ApplicationsRelative Navigation for Formation Flying Spacecrafts Using X-ray PulsarsSAR Image Processing Using Super Resolution Spectral Estimation with SVD-Periodogram Method

Author IndexCopyrightHelpPrintSearchExit

HistoryItem_V1 AddNumbers Range: all pages Font: Times-Roman 10.0 point Origin: bottom centre Offset: horizontal 0.00 points, vertical 36.00 points Prefix text: '' Suffix text: '' Use registration colour: no

BC 50 TR 1 0 629 187 0 10.0000 Both 10 1 AllDoc

CurrentAVDoc

0.0000 36.0000

QITE_QuiteImposingPlus2 Quite Imposing Plus 2.9b Quite Imposing Plus 2 1

0 10 9 10

1

HistoryList_V1 qi2base

Button1: