enabling techniques and algorithms for integrated

Enabling Techniques and Algorithms for Integrated Communication

and Navigation Satellite Systems

PhD Candidate: Lina Deambrogio

Supervisor: Prof. Giovanni E. Corazza

Global Navigation Satellite Systems (GNSS)

• Provide autonomous geo-spatial positioning with global

coverage

• Active systems

• US Global Positioning System (GPS)

• European System Galileo

• The Russian GLONASS

• The Chinese Compass system

• The Indian Regional Navigational Satellite System (IRNSS)

• The Japanese Quasi-Zenith Satellite System (QZSS)

• All these systems will operate independently and without

interfering with each other

2 24-01-2012 - Enabling Techniques and Algorithms for Integrated Communication and Navigation Satellite Systems

System overview: Galileo

3

• Constellation: 30 satellites (27 operational + 3 active spares),

positioned in three circular Medium Earth Orbit (MEO) planes at

23 222 km altitude above the Earth

• High system reliability is of primary importance

• Especially for operation in critical environments, like for air

navigation and SoL services

• Important dates:

• 21 October 2011: Launch of satellite 1&2

• Summer 2012: Launch of satellites 3&4

• 2014 Initial Operational Capability (IOC)

• 2018 Full Operational Capability (FOC)

24-01-2012 - Enabling Techniques and Algorithms for Integrated Communication and Navigation Satellite Systems

GNSS Positioning

• Estimate the user position • Triangulation based on the knowledge of the distance between user

and known points (satellites)

• Distances are computed as the time needed for the signal transmitted by the satellites to reach the user • Additional time uncertainty due to non ideal synchronization

between clocks • Not ranges but pseudoranges • At least 4 satellites are needed

User Receiver (xR , yR , zR)

i-Satellite (xsi , ysi , zsi)


Receiver Operations

5

Data Demodulation

Code Tracking

Carrier Tracking

Acquisition Pseudorange Calculation

Incoming signal

Code Synchronization Phase: • Acquisition

• Coarse timing and frequency estimation

•Tracking • Fine timing and frequency estimation

Position computation: • Ephemeris extraction • Pseudorange calculation • Navigation system solution

Navigation system


Types of Aiding


Physical Level

Position Level

Aiding

•Assist Acquisition

•Visible satellites

•Doppler frequency

•AIM: Reduce TTFF

• Keep tracking circuits locked

•AIM: help reacquisition

•Provide missing information

during signal outages

•AIM: enhance positioning

availability

•P2P

•INS

CODE ACQUISITION AIDING TECHNIQUES

24-01-2012 - Enabling Techniques and Algorithms for Integrated Communication and Navigation Satellite Systems 7

Code Acquisition in CDMA Systems

• Goal: Identify the Code Epoch (t) and the Frequency Offsets (fe) of a

specific satellite signal

• Uncertainty Region (UR) is discretized into time slots t , and the frequency domain is discretized into frequency bins f:

• Two-dimensional matrix must be scanned to find Correct Hypothesis (H1)

• A large number of Incorrect Hypotheses (H0)

0 1 -1

Peak =1

Normalized Offset

p

Hyp H1 Hyp H0 Hyp H0

0.5 -0.5 d

H1 H0 H0

Peak =1 BOC(1,1)

Normalized Offset


Secondary Peak:

the tracking block can

lock onto this outlier

peak

Zero Value:

Code Acquisition

block

collects only noise

Code Acquisition Techniques


• Two different Search Strategies can be defined: • Serial search (i.e. schemes based on one or more correlators)

• Consecutive time/frequency tests

• Parallel acquisition strategies (i.e. FFT/IFFT schemes) • Simultaneously tests all the possible code phases (frequency tests are required)

Incoming

Signal 2

Conj(Local Code)

FFT

Local Code

Incoming

Signal IFFT

Conj(FFT)

2

Parallel

Scheme

Serial

Scheme

Selection of the correct hypothesis

through different Decision Criteria

MAX • Decides in favor of the cell with the

largest correlation value

• Requires to scan the entire UR

Threshold Crossing (TC) • Compares each decision variable in turn

with an optimized threshold

MAX/TC •Hybrid solution

Aiding techniques

• Motivation: • Speed-up initial synchronization

• Issues: • Very low SNR values and large uncertainty region causes

operations to be performed several times

• Techniques: • A-GNSS

• P2P

• Cross-Band Aiding


Cross-Band Aiding (CBA) Idea

• Context

• Dual-band (E1/E5) receivers

• Motivation

• Fast Acquisition for Dual-Band GNSS Receivers

• Idea

• Signals in the two bands are transmitted using a common time reference

• Information can be exchanged in order to mutually reduce uncertainty regions

• Issues

• Different propagation delays

• Different chip rates

• Different code lenghts


E5aos

E5bos/cs/SoL

E6cs

E1os/cs/SoL

Lower L-Band Upper L-Band

1176.451176.45 1207.141207.14

1191.7951191.795 1278.751278.75 1575.421575.42MHz

E5

~ 24 MHz~ 24 MHz~ 51 MHz ~ 51 MHz

E5aos

E5bos/cs/SoL

E6cs

E1os/cs/SoL

Lower L-Band Upper L-Band

1176.451176.45 1207.141207.14

1191.7951191.795 1278.751278.75 1575.421575.42MHz

E5

~ 24 MHz~ 24 MHz~ 51 MHz ~ 51 MHz

Primary Code E1 E5

Code Lenght 4092 10230

Chip rate 1.023Mcps 10.23Mcps

Code Period 4 ms 1 ms

Chip Duration 977.5 ns 97.75 ns

E1-E5 CBA strategy


fk-E1 FFT MAX

Conj(FFT)

IFFT

Local Code E1c fk-E1

2MF

BOC

fs-E1

First Step - Master Signal E1

CBA E1-E5

Controller

MF BPSK

fs-E5 SSB

Processing

E5a SSB

fk-E5 MAX

Local Code E5aQ* fk-E5

2

Dual Band

Front-End

IF

Processing

Second Step - Slave Signal E5

Circular

Phase-Shift

Entire E5

AltBOC(15,10)

E1b-c

BOC (1,1)

• Acquisition in E1 over the complete uncertainty region (E1-CUR) Master Signal

• Acquisition in E5 with a reduced uncertainty region (E5-RUR) Slave Signal

E5-E1 CBA strategy


• Acquisition in E5 over the complete uncertainty region (E5-CUR) Master Signal

• Acquisition in E1 with a reduced uncertainty region (E1-RUR) Slave Signal

MF BPSK

fs-E5

fk-E1 MAX

Local Code E1c*

Entire E5

AltBOC(15,10)

fk-E1

2

SSB

Processing

MF BOC

fs-E1

E1b-c

BOC (1,1)

E5a SSB

Dual Band

Front-End

IF

Processing

fk-E5 FFT MAX

Conj(FFT)

IFFT

Local Code E5aQ fk-E5

2

CBA E5-E1

Controller

First Step - Master Signal E5

Second Step - Slave Signal E1

Circular

Phase-Shift

Timing Structures


• E1-E5 CBA

• E5-E1 CBA

E5 E5 E5 E5

E1 1° step

2° step

E5 Reduced Uncertainty Region (E5-RUR)

E1 Complete Uncertainty Region (E1-CUR)

E1-E5

SMBA

~ 977.51 ns

E5 E5 E5 E5

E1 1° step

2° step



E1-E5

SMBA

~ 977.51 ns


E5 E5 E5 E5

E1

1° step

2° step


E5-E1

SMBA

H1E5~ 97.751 ns

H1E1~ 977.51 ns

1ms 1ms 1ms 1ms


E5 E5 E5 E5

E1

1° step

2° step


E5-E1

SMBA

H1E5~ 97.751 ns

H1E1~ 977.51 ns


E5 E5 E5 E5

E1

1° step

2° step


E5-E1

SMBA

H1E5~ 97.751 ns

H1E1~ 977.51 ns

1ms 1ms 1ms 1ms

Code epoch of the slave signal

E5 can be searched inside a

RUR of length:

51051

5 chipsERc

RcE

E

ERUR

The Synchronization in E1 slave

signal need distinguish the start of

the primary code between four

macro-hypotheses.

1415

1 chipsET

TE

Ecode

EcodeRUR

Analytical Overall Mean Acquisition Time: Unidirectional Information Flow (UD)


Master

Search

Master

Error

Master

ACQ

URMfURMf TP

DM

TP

EM zPzP

)1(

PMTz

URMf TP

DM zP

Slave

Search

Slave

Error

Slave

ACQ

URSfURSf TP

DS

TP

ES zPzP

)1(

PSTz

URSf TP

DS zP

CBTz

•The acquisition procedure can be modeled as a Markov Chain: • Represented by the associated Flow Graph

• MAX criterion: the entire uncertainty region is scanned by choosing the largest detection variable

•PDM = Master Band Probability of Detection •PDS = Slave Band Probability of Detection •Pf = Processing Factor •TURM = Master Signal Uncertainty Region •TPM= Master Penalty Time •TURS= Slave Band Uncertainty Region •TPS= Slave Penalty Time

)1)(1()(

)(

PSURSfPMURMf

CBURSURMf

TTP

ES

TTP

EM

TTTP

DsDMA

zPzP

zPPzP

DSDM

PSPMDMPMURMfDSPSURSfDM

CB

z

AUD

PP

TTPTTPPTTPPT

dz

zdPOMAT

))(()(

)(

1

Results: UD performance timing


1

10

100

1000

10000

Mean

Ac

qu

isit

ion

Tim

e [

msec]

E1 C/No [dBHz]

CBA E5-E1, Timing Errors (UD)

CBA E5-E1, no Timing Errors (UD)

CBA E1-E5, Timing Errosr (UD)

CBA E1-E5, no Timing Errors (UD)

E1&E5 Timing Errors

E1&E5 no Timing Errors

35 37 40 42 45 50

40 42 45 47 50 55 E5 C/No [dBHz]

4. 000978 ms


Analytical Mean Acquisition Time: Bi-directional Information Flow (BD)

Master

Search

Slave

Search

“Full”

Slave

Error

Master /Slave

ACQ

URSf TP

DS zP

URSf TP

mDS zPP

)1(

Slave

Search

“Empty”

Slave

Error

PSTz

CBURMf TTP

DM zP

)1(

CBURMf TTP

DM zP

URSf TP

r zP

)1(

URSf TP

m zP

URSf TP

r zP

PSTz

• PDM = Master Band Probability of Detection

• PDS = Slave Band Probability of Detection

• Pf = Processing Factor

• TURM = Master Signal Uncertainty Region

• TPM= Master Penalty Time

• TURS= Slave Band Uncertainty Region

• TPS= Slave Penalty Time

DSDM

DSDMPSURSURMfCB

z

ABD

PP

PPTTTPT

dz

zdPOMAT

)1()(

)(

1

Analytical Overall Mean Acquisition Time: Bi-directional Information Flow (BD)

1)1(10 EcodeacqEP TT

5)5(10 EcodeacqEP TT

Results: BD performance timing


1

10

100

1000

10000

Mean

Ac

qu

isit

ion

Tim

e [

msec]

E1 C/No [dBHz]

CBA E1-E5 no Timing Errors (BD)

CBA E1-E5 Timing Errors (BD)



E1&E5 no Timing Errors

E1&E5 Timing Errors

40 45 50 55

E5 C/No [dBHz]

35 40 45 50

4.000978 ms

Performance Comparison: UD Vs BD


1

10

100

1000

10000

Mean

Ac

qu

isit

ion

Tim

e [

msec]

E1 C/No [dBHz]

CBA E1-E5 no Timing Errors (UD)

CBA E1-E5 Timing Errors (UD)



CBA E5-E1 no Timing Errors (UD)

CBA E5-E1 Timing Errors (UD)



40 45 50 55 E5 C/N0 [dBHz]

35 40 45 50

CODE TRACKING AIDING TECHNIQUES


Code Tracking

• Refine the coarse values of code phase and frequency

• Keep track of the signal properties over time

• Necessary for pseudorange estimation

• Code tracking

• usually based on a feedback scheme

identified as Delay Lock Loop (DLL)

• Carrier tracking

• either tracking the phase or the

frequency

21

Local

Oscillator

90°

Primary Code

Generator

Incoming

Signal

E

E

L

L

IE

IL

QE

QL


Tracking Aiding Techniques

• Motivation: • Improve tracking loops ability to stay locked in the presence of

• Multipath

• Ionospheric scintillation

• Noise

• High receiver dynamics

• Improve timing estiation precision

• Goal: Robust Code Tracking for GNSS signals

• Solutions: Vector Tracking

Ultratight Integration


Vector Tracking Functioning

• User position and velocity are used to predict code phase and carrier frequency

• Combines tracking and navigation processing

• All received signals are processed together

• Feedback provided by the navigation processor

• Characteristics: • Noise reduction

• Robustness to temporary blockages

• Better optimization

• Performance improvement with respect to scalar loops

• Errors in one channel can affect other channels


Correlators Discriminator

NCO Code

Generator

Incoming

Signal

Navigation

Processor

...

Correlators Discriminator

NCO Code

Generator feedback

path

feedback

path

VDLL Implemetation


• Additional states in the estimation filter require additional measurements

• Atmospheric error and satellite clock errors predictions can be extracted from the broadcasted navigation data

F H

z-1

Corrections (sat position, clock errors, ionosphere)

S

z-1

atmospheric bias,

sat clock bias

X~

X̂

j

i 1ˆ

t

j

it

1ˆˆˆ

ii XXX

j

i

j

i 1ˆ

tttj

i

j

i 1ˆˆˆ

ttt

Ultra-tight Code Tracking

• Motivation: • improve tracking loops ability to stay locked

• improve timing estiation precision

• Inertial Sensors: • Autonomous Navigation Systems → Robustness against signal outages and jamming

• High update rate

• Precision that degrades with time

• Goal: Robust Code Tracking for GNSS signals

• Solution: Ultra-tight Integration


GAUSS Ultra-tight integration • The Gaussian AUtocorrelation Scaled Sum (GAUSS) scheme:

• An artificial autocorrelation peak is generated from INS information

• The two sets of autocorrelation are summed non-coherently each scaled by its Mean Square Error (MSE)

2

2

2

)(

22

1)( INS

INSt

INS

etG

delay estimated by the INS

variance selected according to the early-late spacing

Local

Oscillator

90°

Primary Code

Generator

Incoming

Signal

E

E

L

L

IE

IL

QE

QL

Estimated Delay

From INSGaussian

Generator

GE

GL

Discriminator

σ2GNSS σ2

INS

DGAUSS

INSGNSS

INSGNSSGNSSINSGAUSS

MSEMSE

DMSEDMSED


GAUSS concept


GAUSS results

28

-0.025

-0.02

-0.015

-0.01

-0.005

0

0.005

0 20 40 60 80 100 120 140 160 180 200

Iterations

Ch

ips

Tracking Estimation GNSS stand alone

GAUSS Ultra-Tight Tracking Estimation Te=20 s

GAUSS Ultra-Tight Tracking Estimation Te=200 s


International Projects

• ESA-TT&C-I (Spread Spectrum Codes suitable for Mono-mode TT&C Transponders)

• ESA-P2P (Distributed and cooperative positioning innovative GNSS applications)

• EU-GAMMA-A (Galileo Receiver for Mass Market Applications in the Automotive Area)

• TAS-I (Study of GNSS Software Radio Receivers)


Publications

• [1] F. Bastia, L. Deambrogio, C. Palestini, M. Villanti, R. Pedone, and G.E. Corazza - “Hierarchical Code Acquisition for Dual Band GNSS Receivers”, Position Location and Navigation Symposium (PLANS2010), May 3-6, 2010, Palm Springs, California.

• [2] C. Palestini, L. Deambrogio, F. Bastia, and G. E. Corazza - “An Insider View on Tracking Loops: a Novel Ultra-Tight GNSS/INS Hybridization Approach”, Position Location and Navigation Symposium (PLANS2010), May 3-6, 2010, Palm Springs, California.

• [3] I. Thibault, G. E. Corazza and L. Deambrogio - “Random, Deterministic, and Hybrid Algorithms for Distributed Beamforming”, Advanced Satellite Multimedia Systems Conference (ASMS/SPSC 2010), September 13-15, 2010, Cagliari, Italy.

• [4] L. Deambrogio, C. Palestini, F. Bastia, G. Gabelli and G. E. Corazza “Impact of High-End Receivers in a Peer-To-Peer Cooperative Localization System”, Ubiquitous Positioning, Indoor Navigation and Location-Based Service (UPINLBS 2010), October 14-15, 2010, Kirkkonummi, Finland.

• [5] I. Thibault, G. E. Corazza and L. Deambrogio – “Phase Synchronization Algorithms for Distributed Beamforming with Time Varying Channels in Wireless Sensor Networks”, International Wireless Communications and Mobile Computing Conference (IWCMC 2011), July 5-8, 2011, Istanbul, Turkey.

• [6] L. Deambrogio, F. Bastia, C. Palestini, M. Villanti, R. Pedone and G. E. Corazza – “Cross-Band Aided Code Acquisition in Dual Band GNSS Receivers” submitted to IEEE Trans. on Aerospace and Electronic Systems, November 2011.

• [7] G. Gabelli, L. Deambrogio, C. Palestini, F. Bastia, G. E. Corazza and J. Samson – “Cooperative Code Acquisition based on the P2P Paradigm”, ION International Technical Meeting (ION ITM 2012), January 30- February 1, 2012, Newport Beach, California.

• [8] L. Deambrogio, O. Julien, C. Macabiau, G. E. Corazza – “Improvement of Carrier Tracking Robustness in the Presence of Degraded GNSS Signals using Vector Tracking Loops”, Abstract accepted to Position Location and Navigation Symposium (PLANS2012), April 23-26, 2012, Myrtle Beach, South Carolina.


Short Courses

• NEWCOM++ Winter School, 2-6 Febbraio 2009, Aachen, Germania

• Short-Range Positioning Systems: Fundamentals and Advanced Research Results with Case Studies, 5-6 March 2009, Bologna

• GPS/INS Multisensor Kalman Filter Navigation, 10-14 Maggio 2009, Parigi, Francia

• SatNEx Summer School, 27-31 Luglio 2009, Salisburgo, Austria

• Heterogeneous wireless networks: architectures, Qos performance and applications, 9 e 11 Settembre 2009, Bologna

• Aster DOC, 5-10 Luglio 2010, Bologna


Enabling Techniques and Algorithms for Integrated Communication

and Navigation Satellite Systems

PhD Candidate: Lina Deambrogio

Supervisor: Prof. Giovanni E. Corazza

POSITION LEVEL TECHNIQUES


Position Level Aiding

• Motivation: • improve positioning solution availability

• improve accuracy

• Inertial Sensors: • Autonomous Navigation Systems → Robustness against signal outages and jamming

• High update rate

• Precision that degrades with time

• P2P cooperation: • Peers in a network have positions correlated to each other

• Goal: Availability and Accuracy improvements in positioning

• Solution: Loose/Tight Integration

P2P positioning


Motivation

• The P2P paradigm: • Peers belonging to the same network

have positions and velocities that are correlated to each other

• Exploitation of direct communication links among nodes in a network used to exchange aiding information

• Increased positioning capabilities especially in those scenarios where triangulation would be impossible for a single user


• Different solutions • GNSS-data only Algorithm (exchange of pseudorange measurements)

• Hybrid Algorithm (exchange of terrestrial ranging information)

• Method:

• Extended Kalman Filter

• Goal: • GNSS-data only → Improve service availability

• Hybrid → Improve accuracy

P2P Positioning


Project position

estimates

ahead of time to the

next time step

Accept measurements

and update estimates

accordingly

Input

measurements

Output

estimations

•Position

• Velocity

• Clock bias

• Direct measurements:

pseudorange to visible

satellites

• Aiding information:

additional pseudoranges

of neighboring peers to

non visible satellites

GNSS-data only

Kalman filter

P2P Positioning Results

895690

895700

895710

895720

895730

895740

895750

895760

4469140 4469150 4469160 4469170 4469180 4469190

Y-e

cef

[me

ter]

X-ecef [meter]

GNSS-Data Only Exchange

Real Mass-Market AidingPeer Position

Calculated UserPosition

Real User Position


enabling techniques and algorithms for integrated

Documents