International Journal of Engineering Science Invention

ISSN (Online): 2319 6734, ISSN (Print): 2319 6726

www.ijesi.org Volume 3 Issue 11 November 2014 PP.01-05

IR UWB TOA Estimation Techniques and Comparison

1,

Sri Hareendra Bodduluri, 2, Anil Solanki and Mani VV

Department of Electronics and Communication
National Institute Of Technology, Warangal, India

ABSTRACT: In the recent times Impulse Radio-Ultra wide bandwidth (IR-UWB) positioning techniques have
been gaining interest especially in agglomerated environments where signals from satellite navigation systems
are not reliable. Range measurements required by UWB positioning systems are usually obtained from the timeof-arrival (TOA) estimation of the first path, which could be a challenging task in the presence of multipath and
interference. In this paper we discuss various Time of Arrival (TOA) estimation techniques which include Deconvolution, Maximum Likelihood (ML) estimation, Pseudo spectrum multiple signal classification (MUSIC)
and Two stage estimation algorithms, their performances against different signal to noise(Eb/No) ratios and
also their complexities. The performances of the proposed algorithms are investigated via simulations using
IEEE 802.15.4a UWB channel model in Line of Sight (LOS) environment.

KEYWORDS: UWB, TOA, ML estimation, MUSIC Pseudo spectrum, DE convolution, Two Stage Estimation.
I.

INTRODUCTION:

In the recent times the need for accurate location has been gaining interest.Especially in cluttered
environments (inside buildings, in urban areas, tree canopies etc.), where Global Positioning system is
inaccessible.The ultra-wide band-width (UWB) technology offers potential for achieving high accuracy in such
cluttered environments [1]-[4] . Most positioning techniques are based on, the first path is often not the time-ofarrival (ToA) estimation of the first path [2],[5]-[6]. TOA based position estimation systems measure the
distance between transmitter and receiver based on the estimation of propagation delay which is affected by
multi path channel and noise. In dense multipath channels, the first path is often not the strongest making the
TOA estimation challenging [7]-[8]. The Direction of Arrival is also used for the estimation position The
Direction of Arrival (DOA) based positioning technique involve the use of antenna arrays. Due to the large
bandwidth of the UWB signal, the number of paths may be very large, especially in indoor environments.
Therefore, accurate angle estimation becomes very challenging due to scattering from objects in the
environments. Moreover, time-based approaches can provide very precise location estimates, and therefore they
are better motivated for UWB.An UWB signal is defined as signal that possess an absolute bandwidth larger
than 500Mhz or fractional energy bandwidth about 0.2-0.25 .The short pulse duration of UWB offers a lucrative
application which is positon estimation especially indoor positioning. In the following estimation techniques
LOS scenario is assumed. In general different measures are used to evaluate a TOA estimation algorithm in
wireless channel like noise sensitivity, complexity of the algorithm and priori information.
Many high resolution time delay estimating algorithms have been proposed in order to accurately
measure TOA by determining the delay of the first incoming signal path. These techniques can either be
performed in time domain or in frequency domain. Energy detection(ED) based estimations are growing popular
due to their low complexity implementation at sub-NY Quist sampling rates.The main purpose of our work is to
investigate the effects of multipath propagation on ToA estimation using real measurement data by considering
different algorithms with different levels of complexity. The trade-off between estimation accuracy, complexity
and sensitivity to parameter choice for different propagation conditions is discussed.The outline of this paper is
as follows. Section-II introduces the IR UWB signal model. Different estimation techniques like DE
convolution, ML correlation, pseudo spectrum MUSIC and Two step coarse-fine estimation techniques are
discussed in section III. Numerical analysis and conclusion are discussed in section IV and section V
respectively.

II.

SYSTEM MODEL:

We consider an IR UWB system where transmission of symbol happens in repeated pulses N f . The
mathematical model for transmitted signal is given as
(1)

www.ijesi.org

1 | Page

IR UWB TOA Estimation Techniques and Comparison

p(t) refers to a Gaussian pulse or one of its derivatives with a pulse duration of T p. Symbol duration Tsym=Nf :Tf
which is much greater than the pulse duration T p ,where Nf is number of frames and T f is frame duration. T is
Pulse position Modulation (PPM) interval. Nc is the number of chips per frame cj which is the time hopping
sequence which takes values from {0, Nc-1}. An ideal received wideband signal template with T p = 0.77 ns is
shown in Fig.1.

Fig. 1. An ideal UWB pulse template

Mathematical model for the general considered multicluster, multipath fading channel [9] is given by
(2)
In the above equation, L represents the number of multipath components, k and k are the complex
attenuation and propagation delays of the kth path respectively and (t) is the unit impulse response. In general,
due to the mobility of the transmitter, the receiver, and objects in the environment, the parameters k and k are
time varying random variables. A realization of the Amplitude response of IEEE 802.15.4a is shown in Fig.2

Fig. 2.Amplitude response of the channel.

The multipath affected received signal r(t) consists of superimposition of several attenuated delayed
and eventually distorted replicas of a transmitted wave from s(t).When propagation fluctuation with in an
observation time T>>Tsym and path dependent distortions can be neglected.r(t) can be expressed as follows
(3)
Where s(t) is the transmitted signal. L is is the number of multipath components. k and k are the
complex attenuation and propagation delays of the kth path respectively. n(t) is the Gaussian noise present
during the transmission.

III.

TOA ESTIMATION TECHNIQUES:

A. Deconcolution : Deconvolution of the time domain waveforms can be used to determine impulse response of
Linear Time Invariant (LTI) systems. Deconvolution methods are essentially inverse filters [10]. The received
signal r(t) can be represented as a convolution of the transmitted signal and channel impulse response (CIR).
r(t) = s(t) * h(t) + n(t)
(4)
where n(t) is the additive white Gaussian noise. In the frequency domain convolution transforms into
multiplication. So the above equation (4) becomes
R(f) = S(f) H(f) + N(f)

(5)

www.ijesi.org

2 | Page

IR UWB TOA Estimation Techniques and Comparison

In the above equation R(f),S(f), H(f) and N(f) are Fourier transforms of r(t), s(t), h(t), n(t) respectively. In ideal
conditions we divide R(f) by S(f) to obtain the channel response and estimate the TOA of the first path. The
vector form representation of the above equation (5) is R = SH + N
Where R = [R(1) . R(N)]
S = dia[S(1) . S(N)]
N = [N(1) . N(N)]
H = [H(1) . H(N)]
Let Y =
then we have Y = H + W, where W =
. One of the adverse effects of inverse filtering is noise
enhancement. The following constrained inverse filter provides a better performance against noise
(6)
where is taken as 50 [11]
B. ML Estimation: If Gaussian noise is added then ML criterion is equivalent to the minimum mean squared
error (MMSE) criterion. Given an observation vector r of the received signal
r = S() + n
(7)
Where n
r

with elements

for i = 1, 2, , N
for i = 1, 2, , N

with elements

S() =
for l = 1, 2,

where p

with elements

, for i = 1, 2, .. , Z.

p(t) is the transmitted pulse, Z is the number of sample of the pulse shaper,
is the discretized version of time
delay , such that
, TS is the sampling time and Nis the number of samples of the received signal.
The ML estimate of the delay vector is the value that minimize the following mean squared error
where
The ML estimate of in the continuous-time domain, reformulated in the discrete-time domain, is given by [12]
(8)
Where

is the correlation between the received signal and different delayed versions of S() and
is the auto correlation matrix of S(). Hence Eq. (8) can be written as
(9)

C. MUSIC Pseudo Spectrum Algorithm: The MUSIC algorithm is a typical sub spaced based method. The
sub spaced based methods are better than parametric methods like ML algorithm. The MUSIC algorithm is
based on Eigen vector decomposition (EVD) of the covariance matrix of received signal. The vector form of the
signal in frequency domain is written as
R = SH + N
(10)
Where R, S, H, N are vector representation of frequency domain of received signal, transmitted signal, channel
and noise respectively. From vector form (10) of received signal, if we assume that the auto correlation matrix
of the received signal can be obtained as
(11)

www.ijesi.org

3 | Page

IR UWB TOA Estimation Techniques and Comparison

= SE[H
= SP
, where P = E[H
and P is symmetric and positive definite with
rank L. Let the number of samples N be larger than the number of paths L, and rank of SP will be L. Then
from the formulation of received signals and additive noise, we obtain
SP
and

(12)

= SP

(13)

whereei is the eigenvector and associated with corresponding Eigen value . The Eigen decomposition of auto
correlation matrix gives two orthogonal sub spaces known as signal and noise subspace. The principle largest L
Eigen values span the signal subspace and the rest span the noise subspace. The Eigen values of the noise sub
space {
} are equal to the variance of the noise . Let
and
be the noise and signal sub
spaces respectively, then
,
. The MUSIC pseuedospectrum is given by [13]
(14)
Where

The Toa of the first arriving path can be estimated by finding the delay at which pseuedospectrum MUSIC
achieve the maximum value.
D. Two Stage Estimation: In two stage estimation first stage is a simple coarse estimation that provides the
time reference for the symbol synchronization and estimates the threshold used in Toa estimation algorithm
[14]. The two stages estimation process is performed in frequency domain. It is assumed that an all zero training
sequence is used for the estimation. The coarse estimation consists of an energy estimator and a simple search
algorithm that identifies the beginning of the symbol by applying a minimum distance criterion. In coarse
estimation an acquisition time of length equal to the duration of
is considered. Minimum acquisition
window shall be equivalent to two symbol duration in order to perform the fine estimation over a single symbol
. The fine TOA estimation
and a high resolution time delay estimates of the first arriving path with
respect to the time reference obtained in the coarse estimation stage. The TOA estimation resulting from the fine
estimation is given by
(15)
The fine TOA estimator consists of finding the first delay that exceeds the given threshold P th in the power
delay profile summed overall direction [15]. P th is assumed based on coarse estimation
(16)
Where P() =

(17)

The signature vector is represented as

For more robust estimation

, where

is averaged over

symbols.

E. Numerical Results: For numerical evaluation of the above algorithms we consider the channel models
developed within the framework of the IEEE 802.15.4a. In particular it is used the CM1 Residential LOS
channel model.The pulse p(t) is Gaussian second order derivative with a 3dB bandwidth of 1.3 GHz and a centre
frequency f0 = 4:5GHz according to the European ECC. The pulse repetition period is T f = Tsym/Nf = 128ns.All
simulations are given for 100 independent channel realizations.
Simulation Parameters
Parameter
Pulse duration P(t),

Value
0.77 ns

Band Width BW
Number of frames per symbol,

1.3GHz
31

Number of symbols in the acquisition interval,

Symbol duration,

47 symbols
3974.4 ns

Sampling rate, 1/

3GHz

The remaining parameters are chosen according to the algorithm in consideration.

www.ijesi.org

4 | Page

IR UWB TOA Estimation Techniques and Comparison

Fig.3. depicts the normalized root mean square error (RMSE) of the estimated TOA in nanoseconds (ns) of
various algorithms against varying snr

Fig. 3. Error performance of various algorithms as a function of Eb/No in LOS environment.

F. Conclusion: Various TOA estimation techniques for UWB system complaint with 802.15.4a have been presented. It is
observed that music psuedospectrum has higher performance due to separating into signal and noise subspaces, the effect of
noise is reduced but also this method has high complexity due to matrix Eigen decompositionmaking this algorithm quite
difficult to be used in practical applications. Another drawback of this method is that the number of paths L should be known
as priori information of channel. ML has the second best performance against noise; no prior information is needed forthe
implementation of ML estimation method, but it has high complexity due high sampling rate. Two Stage Estimation has a
good performance and also low complexity but is asymptotic at higher SNR in nature due blind resolution. De-convolution is
poor against lower SNR another drawback associated with the De-convolution methods is their increased complexity and
memory requirements (mainly due to matrix inversions), which increase the implementation cost of this method in a real
system. So we can observe a trade-off between complexitiesandaccuracy so far two stage estimation has the best trade-off
between accuracy and complexity.

REFERENCES:
[1]
[2]

[3]

[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]

D. Dardari,"Pseudo-random active UWB reflectors for accurate ranging," IEEE Commun.Lett., vol. 8, no. 10, pp. 608610, Oct
2004.
Sinan Gezici, Zhi Tian, Georgios B. Biannakis, Hisashi Kobayashi, Andreas F. Molisch, H. Vincent Poor, ZaferSahinoglu
"Localization via ultra-wideband radios: a look at positioning aspects for future sensor networks,"IEEE Signal Processing Mag.,vol
22,pp.70-84,July 2005.
C.Falsi,D.Dardari,L.Mucchi
and
M.Z.Win,"Time
of
arrival
estimation
for
UWB
localizers
in
realsticenvironments",EURASIPJ.Appl.Signal Processing(Special Issue on Wireless Location Technologies and Applications),vol2006,pp.Article ID 32082,1-13,2006.
Y.Shen and M.Z.Win,Fundamental limits of wideband localization ac-curacy via fisher information,inProc.IEEE Wireless
Commun. and Net-working Conf., HONG KONG,Mar.2007,pp. 3046-3051.
D. Dardari, C.-C. Chong, and M. Z. Win"Analysis of threshold-based TOA estimator in UWB channels,"inProc. of European Signal
Processing Conf.(EUSIPCO),Florence, ITALY,Sept. 2006.
K.Yu and I.Oppermann,"Performance of UWB position estima-tionbased on time -of-arrival measurements,"inInternational Workshop on Ultra Wideband Systems.Joint UWBST and IWUWBS 2004., Kyoto.Japan,May,2004,pp.400-404.
M. Z. Win and R. A. Scholtz,"On the robustness of the ultra-widebandwidth signals in dense multipath environments,
"IEEECommun. Lett., vol. 2, no. 2,pp.51-53, Feb. 1998.
Characterization of ultra-wide bandwidth wireless indoor commu-nications channel: A communication theoretic view,"IEEE J.
Select.AreasCommun., vol. 20, no. 9, pp. 1613-1627, Dec. 2002,.
lvarez, G. Valera, M. Lobeira, R. Torres, and J. L. Garca,"Ultrawidebandchannel characterization and modeling,"in Proc. Int.
Workshop on Ultra Wideband Systems, (Oulu, Finland), June 2003.
M. D. Hahm , Z. I. Mitrovski , and E. L. Titlebaum ,"Deconvolution in the presence of Doppler with application to specular
multipath parameterestimation ,"IEEE Trans. Signal Process. , vol. 45 ,pp. 2203 2219 , 1997
A Muqaibel, A Safaai-Jazi, B Woerner, S Riad,"UWB channel impulse response characterization using
deconvolutiontechniques",Circuits and Systems,2002. MWSCAS-2002.
M. Z. Win and R. A. Scholtz,"Characterization of ultrawide bandwidth wireless indoor channels: a
communicationtheoreticview,"IEEE Journal on Selected Areas in Communications, vol. 20, no. 9, pp. 16131627, 2002.
Chin-Der Wann ;Sheng-HsiungHsu,"Estimation and Analysis of Signal Arrival Time for UWB Systems",Vehicular Technology
Conference, 2004. VTC2004-Fal, IEEE 60th,vol:5,2004
M. Najar and M. Navarro, "Joint synchronization and demodulation for IR-UWB", in IEEE International Conference on UltraWideband, Hannover, Germany, Sept. 10-12 2008. (ICUWB2008) .
Navarro M.,Prior S.,Najar, M.Low Complexity Frequency Domain TOA Estimation for IR-UWB Communications",IEEE 64th
Vehicular Technology Conference,Sept. 2006.

www.ijesi.org

5 | Page