DVB-T Signal Analysis for Passive Radar Applicationinfocom.uniroma1.it/rrsn/dottoratoTLR/uploads/Dottorand… · · 2010-10-19technique for Ambiguity Function improvement synchronization

DVBDVB--T Signal Analysis for T Signal Analysis for Passive Radar ApplicationPassive Radar Application

Dipartimento INFOCOM – Roma, 22 Ottobre 2009

Diego LangellottiTutor : Prof. Pierfrancesco Lombardo

Outline

IntroductionPassive Bistatic Radarwaveforms of opportunity: DVB-T signalspp y g

DVB-T signal features

DVB-T signal simulation

DVB-T Ambiguity Function Evaluation:technique for Ambiguity Function improvementsynchronization effects on the performance

R l d f iReal data performance comparison

References

22/10/2009 Diego Langellotti, "DVB-T Signal Analysis forPassive Radar Application"

2

PBR Processing scheme

A passive radar exploits existing transmitters as illuminators of opportunity to perform target

detection and localization.

Advantages:- low cost

REFERENCE SURVEILLANCE

- covert operation, low vulnerability - reduced impact on the environment

Drawbacks:

Ref SurvDISTURBANCECANCELLATION

CHANNEL CHANNEL- the transmitted waveform is not under control of the radar designer.

Drawbacks:

2D CROSS-CORRELATION

Surv’

CFAR THRESHOLD

Ref

g

- continuous wave and lower power levels ( long integration time).


Waveform of opportunity: DVB-T

Exploitation of different waveforms of opportunity: digital TV signals

Frequency range:52 88 d 174 230 MH VHF52 ÷88 and 174 ÷ 230 MHz VHF470 ÷ 870 MHz UHF

Wider bandwidth: 7 ÷ 8 MHz increased range resolutionWider bandwidth: 7 8 MHz increased range resolution

High power transmissions

Wide coverage:Switch off of analogical transmissions already started

OFDM M d l tiOFDM Modulation

Presence of deterministic peaks


DVB-T signal (1/3)Super Frame (4-Frames) Super Frame (4-Frames)

Frame I (68 symbols) Frame II (68 symbols) Frame III (68 symbols) Frame IV (68 symbols)

Symbol 0 Symbol 1 Symbol 2 Symbol 67

TS = TU + TG

Guard Interval

Useful Part

TG TU

2 transmission modes:2k (TU = 224µs, carrier spacing 4.464 Hz) ( U µ p g )8k (TU = 896µs, carrier spacing 1.116 Hz)

Guard interval duration:1/32, 1/16, 1/8 or 1/4 of the useful part duration TU1/32, 1/16, 1/8 or 1/4 of the useful part duration TU

DVB-T signal (2/3)TU = 224 µs (2k Mode)

TG TU

U µ ( )

TU = 896 µs (8k Mode)Guard Interval Useful Part

Symbol Carriers

KMIN = 1704 (2k Mode)

KMAX = 6816 (8k Mode)KMIN = 0

OFDM symbol useful part is composed of:d t ith 3 diff t d l ti h QPSK 16 QAM 64 QAMdata, with 3 different modulation schemes: QPSK, 16-QAM, 64-QAMcontinual pilots, transmitted at fixed carriersscattered pilots, transmitted at fixed carriers over four symbolsTransmission Parameter Signalling (TPS) pilots, transmitted at fixed carriersa s ss o a a ete S g a g ( S) p ots, t a s tted at xed ca e s

Modulation of data and TPS is normalized, while the pilots (continual and scattered) are transmitted at boosted power level (the average power EP = 16/9)


g p P / )

DVB-T signal (3/3)

DVB-T signal frame structure

Continual Pilots

Scattered Pilots

s0

s67

s66

s0s1s2

time time (symbol) c0 c1 c2 c3 c4 cmax

frequency (carrier)TPS Pilots


System Block Diagram


DVB-T signal simulation

Mapper Frameadaptation OFDM

Guardinterval

insertionD/A Front endInput sequence

Pilots & TPSsignals

Mapper input sequence:random binary sequence, grouped into symbolsrandom sequence created by applying mpeg2, coding and interleavingassigned binary sequence, grouped into symbolsg y q g p yassigned binary sequence with application of mpeg2, coding and interleaving


DVB-T signal simulation: Mapper (1/2)


Guardinterval


Pilots & TPSsignals

Transformation of M bit packets into complex symbols, according to the selected constellation

QPSK M 2QPSK: M = 216-QAM: M = 464-QAM: M = 6

Possibility to select the α parameter value in order to consider hierarchical constellations


DVB-T signal simulation: Mapper (2/2)

Mapping example:

MODE: 2kGuard interval: 1/32

QPSKnormalized

constellation

/Channel: 8 MHzSymbols: 10Hierarchy: not set

16-QAMnormalized normalized

constellation

64-QAMnormalized

constellation

11

constellation


DVB-T signal simulation: Pilots & TPS signals (1/2)


Guardinterval


Pilots & TPSsignals

Generation of continual and scattered pilots according to a reference sequence PRBS (maximum length sequence)reference sequence PRBS (maximum length sequence)Generation of Transmission Parameter Signalling (TPS)pilots (one TPS bit per symbol), according to 68 bit sequence


DVB-T signal simulation: Pilots & TPS signals (2/2)

QPSKnormalized

constellation

Mapping example:

MODE: 2kGuard interval: 1/32

TPS pilots/

Channel: 8 MHzSymbols: 10Hierarchy: not set Continual / Scattered

pilots

16-QAMnormalized normalized

constellation

64-QAMnormalized

constellation

13

constellation


DVB-T signal simulation: OFDM symbol (1/3)


Guardinterval


Pilots & TPSsignals

Frame adaptation: data, pilots (continual and scattered), and TPS are organized inside a symbol according to the carrier positions suggested by the DVB-T standardgg yOFDM

zero-padding: the number of carriers is increased to the next power of 2 (2048 for 2K mode, 8192 for 8K mode)

lsignal IFFTGuard interval insertion: each OFDM block is extended copying in front of it its own end (cyclic prefix, variable width: 1/4, 1/8, 1/16 or 1/32 of the original block length)or 1/32 of the original block length)


DVB-T signal ambiguity function (1/2)

The presence of specific features of the DVB-T signal (guard interval,pilots, etc.) introduces a number of undesired deterministic peaks in theambiguity function

These peaks can mask the signal reflected from targets and/or

15

introduce false alarms


DVB-T signal ambiguity function (2/2)

The positions of these unwanted peaks are deterministic

-5

0intra-

symbol inter-symbol

peaks

-20

-15

-10

-5

guard interval peak

symbol peaks

peaksSpecifically, in 2k mode:

peak generated by guard intervalτ= 224 µs (TU)

-35

-30

-25

|χ(τ

,0)|,

(dB

)

peaks due to pilot carriersintra-symbol peaks0 ≤ τ≤ TS = TU + TG

55

-50

-45

-40S U G

inter-symbol peaksτ> TS

-0.2 0 0.2 0.4 0.6 0.8 1-55

τ, (ms)

Different techniques have been proposed in order to remove these undesired peaks These techniques require synchronization

16

undesired peaks. These techniques require synchronization.


Ambiguity Function improvement (1/3)TARGET REFERENCE TARGET

CHANNEL

DPI SUPPRESSION

REFERENCE CHANNEL

TIME SYNCHRONIZATION

GUARD INTERVAL BLANKING

Removing Guard Interval Peak

PILOTS EQUALIZING

PILOTS BLANKING

Removing Pilot Peaks

CAF2CAF1

CAF3

R. Saini and M. Cherniakov, “DTV signal ambiguity function analysis for radar application”, IEEProc. on RSN, Vol. 152, No. 3, pp. 133 – 142, June 2005

Z Gao R Tao Y Ma and T Shao “DVB-T Signal Cross-Ambiguity Functions Improvement forZ. Gao, R. Tao, Y. Ma, and T. Shao, DVB-T Signal Cross-Ambiguity Functions Improvement forPassive Radar”, Radar, 2006. CIE '06. International Conference on, pp.1-4, 16-19 Oct. 2006


Ambiguity Function improvement (3/3)

TARGET CHANNELREFERENCE CHANNEL

DPI SUPPRESSION

AF-BASED FILTER

CAF

C. Bongioanni, F. Colone, D. Langellotti, P. Lombardo, T. Bucciarelli, “A New Approach for DVB-TCross-Ambiguity Function Evaluation”, submitted to EuRAD 2009


g y

Synchronization (1/2)

FFTGuard

intervalremoval

A/DFront endReceived signal

CARRIER FREQ. UNCERTAINTY

SAMPLING FREQ. UNCERTAINTY

ARRIVAL TIME UNCERTAINTY

Synchronization uncertainties in a DVB-T receiver:carrier frequency

a difference in the local oscillators in the transmitter and receiver gives rise toga shift of all the subcarriers

sampling frequencya sampling clock offset (with respect to the nominal system samplingp g ( p y p gfrequency) leads to a drift of the OFDM symbol window

arrival time of the OFDM symbolthe transmitter scale is unknown to the receiver


Synchronization (2/2) POST-FFTFine

Synchronization

Received signalFFT

Guardintervalremoval

A/DFront end Frequency Correction

PRE-FFTML

Estimator

Synchronization can be achieved through the following steps:PRE-FFT

coarse timing synchronizationcoarse frequency synchronization

POST-FFT

Maximum Likelihood (ML) estimation (guard interval correlation)

fine carrier frequency estimationsampling frequency acquisitionsampling frequency tracking

Based on the OFDM symbol structure (continual pilot positions)

20

sampling frequency tracking


PRE-FFT synchronization (1/2)l 1 l l 1

Guard Interval

Guard Interval

Guard Interval

l-1 l l+1

Initial obser ation

θ

Initial observation

0 θ+ TG θ+ TU θ+ TS

PRE-FFT synchronization uses the knowledge of the DVB-T signal structureML estimator

{ }

with ( ) ( ) ( )∑−+

=+=

1*GNm

mkUNksksmγ

( ) ( ) ( )∑−+

Φ1 221 GNm

Nkk

( ) ( )( ) ( )θρθγπεθγεθ Φ⋅−∠+=Λ 2cos),(

( )MLML θγε~1~ ∠−=

θ( ) ( ){ }θρθγθ Φ⋅−= maxarg~

ML( ) ( ) ( )∑

=++=Φ

2 mkUNksksm

( ) ( ){ }( ) ( )

+=

22

*U

NkEkE

NksksEρ

21

( )MLML γπ2 ( ) ( )

+

22

UNksEksE


PRE-FFT synchronization (2/2)PRE FFT synchronization examplePRE-FFT synchronization example:

MODE: 2k (2048 samples)Guard interval: 1/4 (512 samples)Samples per symbol: 2560 (2048+512)Samples per symbol: 2560 (2048+512)Channel: 8 MHz (T = 7/64 µs)Delay: 2000 x TCarrier Frequency offset: 0.2/TU

Simulation results

Delay estimation (samples):20004560 (2000 + NS)7120 (2000 + 2NS)

Carrier frequency offset estimation:0.2


POST-FFT synchronization (1/2)

POST-FFT data use the knowledge of the OFDM symbol structurecontinual pilot positions

used to estimate the Integer carrier frequency offsetused to estimate Fractional carrier frequency offsetused to estimate Sampling frequency offset

FIu fnTff '' ∆+=⋅∆=∆

( ) ( )∑∈

+∈+=

IkllImI mksksn 1maxargˆInteger carrier frequency offset

Fractional carrier frequency offsetSampling frequency offset ( )+

=∆ llf '~ ,1,2 ϕϕSa p g eque cy o set

+⋅

=∆

U

G

F

NN

f

122 π( ) 1

122

~ ,1,2

KNGll ⋅

+

−=

π

ϕϕζM. Speth, S. A. Fechtel, G. Fock, and H. Meyr,

“Optimum Receiver Design for Wireless Broad-2122

NUG

+⋅ π

=

= ∑∑ )(arg,)(arg 21 ll ksks ϕϕ

Optimum Receiver Design for Wireless Broad-Band Systems Using OFDM – Part II: A CaseStudy”, IEEE Transactions on Communications,Vol. 49, No. 4, April 2001, pp. 571-578

∆f’F


∑∑∈∈ 1,2,1

)(g,)(g ,2,1Ck

lCk

ll

ϕϕ

POST-FFT synchronization (2/2)POST FFT synchronization example

6889

POST-FFT synchronization example:

MODE: 8k (8192 samples)Guard interval: 1/32 (256 samples)Samples per symbol: 8448 (8192+256)Samples per symbol: 8448 (8192+256)Channel: 8 MHz (T = 7/64 µs)

Delay: 6888 x TCarrier Frequency offset: -0.68/TU

Simulations resultsDelay estimation (samples):

Ca e eque cy o set: 0.68/ USampling offset: 5e-006

Delay estimation (samples):PRE-FFT: 6889

Carrier frequency offset estimation:PRE-FFT: 0 3202PRE-FFT: 0.3202POST-FFT

Integer carrier frequency offset -1/TU

Fractional carrier frequency offset -8.8947e-004/TU


Sampling offset 4.6818e-006

Effect of synchronization errors on the CAF


Real data performance comparison

R l d t tReal data setsDay :25th June 2009Carrier frequency: 714 MHzM d T i i 8kMode Transmission: 8kGuard interval: TG =1/32 TU

Elementary period: T = 7/64 µs

Synchronization resultsDelay estimation (samples):

PRE FFT 5500 TPRE-FFT: 5500 x TCarrier frequency offset estimation:

PRE-FFT: 0.0214 (24 Hz)POST-FFTPOST-FFT

Integer carrier frequency offset -2 (-2.23 KHz)Fractional carrier frequency offset -8.8947e-004(0.05 Hz)S li ff 4 6818 006


Sampling offset 4.6818e-006

ConclusionsConclusioni:Conclusioni:

É stato creato il simulatore di segnale DVB-TÉ stato analizzata la funzione di ambiguità e progettato un filtroper il controllo dei lobi lateraliper il controllo dei lobi laterali.É stato analizzato il problema della sincronizzazione e valutatol’impatto che eventuali errori di sincronizzazione produconosulle prestazioni in particolare sul PSLR della funzione disulle prestazioni, in particolare sul PSLR della funzione diambiguità.

Sviluppi futuri:Estensione del simulatore di segnale alle altre categorie diEstensione del simulatore di segnale alle altre categorie disegnale DVB, in particolare DVB-SH e creare una scenariorealistico con cui effettuare le analisi successive relativeall’elaborazione del segnale.gEstensione del filtro lineare progettato alle frequenze doppler diinteresse.Applicazione gli algoritmi sviluppati attraverso il simulatore suipp g g ppdati reali.


27

ReferencesE T l i ti St d d I tit t “Di it l id b d ti (DVB)European Telecommunications Standard Institute, “Digital video broadcasting (DVB);framing structure, channel coding and modulation for digital terrestrial television”, EN 300744, V1.1.2, 1997.

J.-J. van de Beek, M. Sandell, M. Isaksson, and P. Börjesson, “Low-complex frameh i ti i OFDM t ” i P ICUPC N 6 10 1995 982 986synchronization in OFDM systems”, in Proc. ICUPC, Nov. 6-10, 1995, pp. 982-986

J.-J. van de Beek, M. Sandell, and P. Börjesson, “ML Estimation of Time and Frequency Offsetin OFDM Systems”, IEEE Transactions on Signal Processing, Vol. 45, No. 7, July 1997, pp.1800-1805

M. Speth, S. A. Fechtel, G. Fock, and H. Meyr, “Optimum Receiver Design for Wireless Broad-Band Systems Using OFDM – Part I”, IEEE Transactions on Communications, Vol. 47, No. 11,Nov. 1999, pp. 1668-1676

M. Speth, S. A. Fechtel, G. Fock, and H. Meyr, “Optimum Receiver Design for Wireless Broad-Band Systems Using OFDM – Part II: A Case Study”, IEEE Transactions on Communications,Vol. 49, No. 4, April 2001, pp. 571-578

R. Saini and M. Cherniakov, “DTV signal ambiguity function analysis for radar application”,IEE Proc. on RSN, Vol. 152, No. 3, pp. 133 – 142, June 2005

Z. Gao, R. Tao, Y. Ma, and T. Shao, “DVB-T Signal Cross-Ambiguity Functions Improvementfor Passive Radar”, Radar, 2006. CIE '06. International Conference on, pp.1-4, 16-19 Oct. 2006

C. Bongioanni, F. Colone, D. Langellotti, P. Lombardo, T. Bucciarelli, “A New Approach forDVB-T Cross-Ambiguity Function Evaluation”, submitted to EuRAD 2009DVB T Cross Ambiguity Function Evaluation , submitted to EuRAD 2009


DVB-T Signal Analysis for Passive Radar Applicationinfocom.uniroma1.it/rrsn/dottoratoTLR/uploads/Dottorand… · · 2010-10-19technique for Ambiguity Function improvement synchronization

Documents