Classification of UWB multipath clusters and its distortion effects on positioning error
ABSTRACT Quantification of distortion effects on UWB system performances in terms of positioning error is analysed in this research. UWB multipath distorted channels are simulated in each frequency subband, over 2-11 GHz. Its characteristics are modelled corresponding to multipath clusters along the propagation paths. The classification of clusters and physics-based distortion mechanisms are generalized to be included into the simulation algorithm. Finally, distortion impacts on system performances regarding to frequency dependent characteristics and positioning errors are investigated.
- SourceAvailable from: A.Lee Swindlehurst[show abstract] [hide abstract]
ABSTRACT: Most previously proposed statistical models for the indoor multipath channel include only time of arrival characteristics. However, in order to use statistical models in simulating or analyzing the performance of array processing or diversity combining, it also necessary to know the statistics of the angle of arrival and its correlation with time of arrival. In this paper, a system is described which was used to collect simultaneous time and angle of arrival data at 7 GHz. Data processing...IEEE Journal on Selected Areas in Communications. 01/2000; 18:347-360.
- [show abstract] [hide abstract]
ABSTRACT: Multipath effects in complex environments can result in distortion and time elongation of received UWB pulses. These effects have been analysed using a ray tracing channel model with a PPM-TH modulation scheme and a RAKE receiver architecture. The resultant BER has been calculated using both this model and a generalised statistical model from the literature. Results indicate that the late arrival multipath components allowed for in the ray tracing model have a significant effect on system performance for bit rates of the order 100 Mbit/s. The generalised channel model is shown to be in general agreement with the ray tracing approach for low bit rate systems but somewhat optimistic for high bit rates in a complex multipath scenario.IEE Proceedings - Communications 03/2006; · 0.32 Impact Factor
Conference Proceeding: Multi-target UWB passive ranging with local template uncertainty[show abstract] [hide abstract]
ABSTRACT: Modified phase-only correlator (MPOC) has been proposed for high-resolution multi-target ranging. However, its performance has been found to be deteriorated by Ultra Wideband (UWB) pulse distortion in practical two-way ranging systems, where there are multiple passive reflectors. From experimental measurement, a new received signal model that accounts for the pulse distortion has been proposed, and an MPOC-based ranging algorithm that uses a parametric local template is presented accordingly. The new MPOC algorithm has a higher probability of correct detection of the reflectors. Moreover, it only requires one-time calibration for the local template, which is advantageous for practical applications.Ultra-Wideband, 2008. ICUWB 2008. IEEE International Conference on; 10/2008
Classification of UWB Multipath Clusters and
Its Distortion Effects on Positioning Error?
K. Makaratat, TWC. Brown, S. Stavrou, and B. Evans
Centre for Communication Systems Research, University of Surrey
Guildford, Surrey, GU2 7XH, United Kingdom
K.Makaratat@surrey.ac.uk, T.Brown@surrey.ac.uk, S.Stavrou@surrey.ac.uk, B.Evans@surrey.ac.uk
Abstract—Quantification of distortion effects on UWB system
performances in terms of positioning error is analysed in this
research. UWB multipath distorted channels are simulated in
each frequency subband, over 2-11 GHz. Its characteristics are
modelled corresponding to multipath clusters along the
propagation paths. The classification of clusters and physics-
based distortion mechanisms are generalized to be included into
the simulation algorithm. Finally, distortion impacts on system
performances regarding to frequency dependent characteristics
and positioning errors are investigated.
Ultrawideband (UWB) pulse distortions are inherently
characterised by the extremely large bandwidth. Since
quantifying the impacts of pulse distortion on UWB system
performance appears to be novel, recently there are several
works reported about quantification of UWB distortion effects
in addition to the physics-based pulse distortion issues which
have been addressed in previous research -. Inter symbol
interference (ISI) and probability of bit error rate (BER) due
to distortion effects in various complex propagation
conditions are commonly investigated as researched by -.
Furthermore, Zhou  also reported loss of signal-to-noise
ratio (SNR) as high as 4 dB in template mismatches due to
distortion effects. Moreover, the much larger error range than
the Cramer-Rao lower bound (CRLB) was found in this study.
These lead to errors that can limit the accuracy of times-of-
arrival (TOAs) of received multipath signals. The very high
temporal resolution of UWB pulse signals makes UWB
signals become the ideal candidates for combined
communications and positioning. If the TOAs of incoming
multipath signals are known with little uncertainty, the
computation of propagating distances from the source to the
receivers is still possible with a few errors in estimation. Yuan
et al. reported the deterioration of UWB positioning
performance due to distortions even though the modified
phase-only correlator method for high-resolution multi-target
ranging was employed .
Thus, the effect of UWB distortions becomes significant
and should be investigated to quantify its impact on system
performances and to examine the suitable methods for
compensating its deteriorating effects. Consequently, this
paper reports the quantification of frequency distortion in
terms of positioning error. Characteristics of signal dispersion
due to various multipath distortion conditions are analysed.
Distortion and positioning errors in individual multipath
components subject to specific scenarios and obstruction
materials are analysed herein.
II. CLASSIFICATION OF UWB MULTIPATH CLUSTERS
According to multipath cluster investigations derived from
many indoor UWB measurements -, each multipath
cluster could be identified by a group of MPCs, which are
scattered or reflected from obstructions, with similar angles-
of-arrival (AOAs), angles-of-departure (AODs), and TOAs.
The received multipath clusters from dominant propagation
paths from a transmitter to a receiver are expected to come
from three types of radio propagation paths. The first group
corresponds to scattering nearby the transmitter site. Similarly,
another group of clusters can be observed at the receiver site
due to the scattering objects in the neighbouring area of the
receiver. Finally, line-of-sight (LOS) components between the
transmitter and the receiver are considered. Consequently,
these different propagation clusters can be classified into three
classes as Class-I, Class-II and Class-III type of clusters
respectively  as illustrated in Fig. 1 where the
omnidirectional antenna is considered as the transmitter, and
3x3 planar array antenna is the receiver.
When defining significant clusters specifically related to
propagation scenarios, in the LOS scenario of Case-A
describing a small furnished office room, the channel would
be dominated by Class-III clusters. Next, the non-line-of-sight
(NLOS) scenario Case-B is considered where a transmitter
and a receiver are in the same office room with a-light-wall or
a-cloth-partition separation between both ends; the channel
would be still dominated by Class-III clusters. However, when
propagation paths in a larger furnished office room, NLOS
scenario Case-C, are considered, all three classes of clusters
are presented. Finally, if a transmitter and a receiver are
located in a different furnished room or separated by a thick
wall, the channel would be dominated by Class-I and Class-II
clusters corresponding to the extreme NLOS condition (Case-
D). Diagrams of four channel
corresponding to cluster classifications are depicted in Fig. 1
(a)-(d) for LOS Case-A, NLOS Case-B, NLOS Case-C, and
NLOS Case-D respectively. All individual multipath
components (MPCs) are taken into account corresponding to
dominated clusters and specific environments.
(a) LOS Case-A (b) NLOS Case-B
(c) NLOS Case-C (d) NLOS Case-D
Fig. 1 Configuration of UWB multipath clusters in indoor environments
Consequently, in order to construct realistic UWB
propagation channels, classified MPC clusters and pulse
distortion effects are simulated together. Typically, the total
response, h(τ), from a complex multipath channel can be
modeled by the summation of all impulse responses of local
scattering with the closed form expression of specific
geometric configurations. Unlike typical UWB channel
models, this simulation relies on combining pulse distortion
characteristics into the UWB channel impulse response (CIR)
model. Frequency-dependent effects are included into the
simulation as each CIR model is considered at each frequency
subband, s, , . According to Fig. 1, in the LOS scenario
of Case-A, the channel would be dominated by Class-III
clusters and distorted by multiple reflections and half-plane
diffraction as defined by
) 1 (),,,(
Next, the channel in the NLOS Case-B would be still
dominated by Class- III clusters and distorted by the similar
effects as Case-A. However, thin slab diffraction is
additionally included into the distortion model, hIII
slab(s,τ,θ,φ) to construct hcase-B. When UWB signals
propagating in a larger furnished office room, NLOS Case-C,
all three classes of clusters are presented. All distortion effects
might possibly appear in Class-III similarly to LOS-A, but
distortions due to thin slab diffraction and thick slab reflection
are modeled for Class-II clusters. CIRs of Case-C can be
described by (2) where gI(s,θTx,σTx) is the single-directional
CIR of Class-I clusters at the transmitter site. θTx and φTx are
the elevation and azimuth AOD of the MPCs.
) 2 (),,(),,,(),,,(
Finally, in the severest NLOS Case-D, the channel would
be dominated by Class-I and Class-II clusters corresponding
to room separation by a thick wall. Only thick slab reflection
is considered. Thus only the last two terms in (2) are taken
into account for this hcase-D. All technical terms of hLOS, hGO,
hhalf-plane, hthin-slab, and hthick-slab are not illustrated in this paper,
but their formulas are written in detail as defined in  and .
III. POSITIONING ERROR
In addition to SNR loss and BER due to the pulse shape
mismatch, pulse distortion can also degrade synchronisation
and positioning by time shifting of TOAs as well as an
amplitude error in the correlation peak. Basically, the
accuracy of the TOA estimation or positioning errors caused
by pulse distortion is expressed by the minimum variance of
the TOA estimation error, σ2
τ in terms of the CRLB. This
value is related to signal bandwidths and SNR at the receiver.
For a single path additive white Gaussian noise (AWGN)
channel, the best achievable accuracy of a position estimate,
which is derived from TOA estimation, satisfies the following
inequality , :
This value benchmarks the positioning error caused by
pulse distortion. Thus, the minimum estimated position error
less than dCRLB m can be obtained when considering SNR=0.
c= 3x108 m/s and β is the effective or root mean square signal
bandwidth given by function of the Fourier transform of the
transmitted signal, P( f ). In this study, frequency domain of
pulse position modulation time hopping (PPM-TH-UWB)
transmitted signals is considered to compute β over 2-11 GHz.
In this work, PPM-TH-UWB is generated as the
transmitted signal, W(τ), and transmitted over the synthesised
classified multipath channels; therefore channel impulse
responses including cluster types and distortion effects , 
are taken into account to quantify distortion impacts. Results
of this research are described in the next section.
2 / 1
IV. QUANTIFICATION RESULTS OF DISTORTION EFFECTS
Binary PPM-TH-UWB is generated as the transmitted
signal conveying 3000 bits through 3000 pulses, thus code
repetition coding is not applied. The average pulse repetition
period is 60ns guaranteeing the absence of inter symbol
interference (ISI) in LOS. These generated signals, W(τ),
are simulated and transmitted over the synthesized multipath
channels as specifically classified in each environment;
therefore channel impulse responses (CIRs), h(τ), including
cluster types and distortion effects are taken into account.
Positioning errors due to pulse distortion are computed and
shown in Table 1 regarding classification of scenario cases,
cluster types and obstruction materials. In order to present the
pulse distortion effects on UWB propagation channels,
comparison between lower bound average distance estimation
errors computed by CRLB and estimated ranging errors are
described. Since β, 3.4 GHz, is calculated from the same
transmitted signals over 2-11 GHz, all scenario cases take into
account the same value of β leading to the lower bound
distance error of dCRLB=0.99 mm. in all scenario cases.
Ranging errors are examined by time difference of
correlation peaks between pulse signals as depicted in Fig. 2.
Fig. 2 (a) presents the ideal case of free space LOS scenario
where autocorrelation results of the received pulse R(τ) shown
by the solid line and autocorrelation results of the transmitted
pulse W(τ), the dashed line, present the same peaks of the
waveforms. Both peaks of these symmetric waveforms
correctly indicate the position. Since the impact of pulse
distortion is examined in this study, W(τ) is considered as the
local template to be correlated with the received distorted
signals R(τ) as shown in Fig. 2 (b)-(f). The solid line presents
the correlation result between different templates, W(τ) and
R(τ). These two different templates are normalised to have the
same energy before correlating. Moreover, the dashed curve is
the result of correlation with the same template itself or
autocorrelation of W(τ). The symmetric autocorrelation
waveform can be observed contrary to the different template
correlation whose the asymmetric waveform gives an error in
-70 -350 3570
-2 -10 12
Fig. 2 Correlation of distorted signals with transmitted signals
PULSE DISTORTION IMPACTS ON POSITIONING ERROR
β β β β (GHz)
1 0.083 24.9
Wallboard 3.4 0.99 4
2 -0.415 -124.5
1 0.083 24.9
2 -0.418 -125.4
3 -0.333 -99
Wallboard 3.4 0.99 9
4 -0.50 -150
1 -0.083 -24.9
2 0.083 24.9
Wooden door 3.4 0.99 4
3 0.333 99
Partition 4 -0.417 -125.1
3.4 0.99 6
5 -0.584 -175.2
1 0.584 175.2
2 0.583 174.9
3 0.083 24.9
Concrete Block 3.4 0.99 10
4 0.167 50.1
the position. Distinguishing the difference between these two
peaks can define timing errors as also shown in Table 1.
Negative values of timing errors refer to position shifting
prior to the reference time, which is indicated by the
symmetric waveform peak. When multiplying timing errors
by the light velocity, c=3x108, ranging errors can be defined.
Comparing ranging error distances and the lower bound
distance error, dCRLB=0.99 mm, results significantly show that
apart from SNR, distortion effect can cause performance
degradation in positioning errors much bigger than dCRLB. The
second MPCs scattered from obstruction clusters class-III in
scenarios Case-A, Case-B and Case-C are examined as
illustrated in Fig. 2 (b)-(d) respectively. Since a few difference
in timing errors is rarely observed in the first MPCs
corresponding to LOS components or the-shortest-distance
propagation paths, to present noticeable timing error
characteristics caused by distortion, the later components of
multipath receiving pulses are selected as examples to be
shown. The fifth and the fourth MPCs originated from clusters
class-II in scenarios Case-C and Case-D are also presented in
figure (e) and (f) respectively.
These examples of received pulse distortion are analysed
specifically at the highest energy allocation frequency
subbands, as stated in  and , with various types of
obstruction material. Timing errors of propagation channels in
Case C-III are less than in Case-A and Case-B
notwithstanding its propagation through various obstructions
along the path. This might be due to these pulse distortion
effects are selected as examples taken into account different
frequency subbands and different obstruction. Figure (e)
presents the correlation of dense multipath channels of Case-C
propagation. The later fifth MPC scattered by concrete block
obstruction, cluster Class-II, gives the highest timing error of -
0.584 ns or the ranging error of -175.2 mm. And 0.167 ns
timing error or 50.1 mm. ranging error can be seen for the
fourth MPC of Case D-II due to thick slab reflection from
Furthermore, comparison of pulse distortion effects on
each MPC at each propagation channel is also presented in
Table 1. Generally, positioning errors of the first three MPCs
in propagation channels obstructed by cluster Class-III are
roughly in the same range and less than errors caused by
cluster Class-II. More errors are remarkably estimated in later
MPCs of propagation Case C-II. Moreover, low values of
positioning errors in Case D-II in the fourth and the fifth
MPCs are irregularly observed. This can be assumed that
amplitudes of received MPCs are considerably attenuated
along propagation paths without extremely distorted pulse
shape, hence correlating normalised received pulses, the 4th
and 5th MPCs, with the local template W(τ) possibly gain
small timing and ranging errors.
characteristics and the quantification of distortion effects on
UWB multipath channels were presented. Apart from the
probability of bit error rate, performance degradation in
results of the frequency-dependent
positioning errors was determined to quantify distortion
effects on UWB multipath channels for all frequency
subbands. Correlation between
transmitted pulses determined timing and ranging errors
which these positioning errors extremely exceeded the
benchmark errors estimated by CRLB. In addition, results also
examined that LOS component gained more accuracy of
positioning errors than late-arrival
distorted pulses and
R. C. Qiu, “A generalized time domain multipath channel and its
application in ultra-wideband (UWB) wireless optimal receiver – part
II: physics-based system analysis,” IEEE Trans Wireless Commun., vol.
3, no. 6, pp. 2312-2324, Nov. 2004.
G. Wang, and W. Kong, “Angle-dependent pulse distortion in UWB
radiation and its impact on UWB impulse communications,”
Electronics Letters, vol. 41, no. 25, pp. 1360 – 1362, Dec. 2005.
A. Muqaibel, A. Safaai-Jazi, A. Bayram, A.M. Attiya, and S. M. Riad,
“Ultrawideband through-the-wall propagation,” IEE Proc. Microw.
Antennas Propag., vol. 52, no. 6, pp. 581- 588, Dec. 2005.
R. C. Qiu, “A generalized time domain multipath channel and its
application in ultra-wideband (UWB) wireless optimal receiver – part
III: system performance analysis,” IEEE Trans Wireless Commun., vol.
5, no. 10, pp. 2685-2695, Oct. 2006.
M. Z. Win, and R. A. Scholtz, “Characterization of ultra-wide
bandwidth wireless indoor channels: a communication-theoretic view,”
IEEE J. Sel. Areas. Commun., vol. 20, pp. 1613-1627, 2002.
Y. Zhang, and A. K. Brown, “Complex multipath effects in UWB
communication channels,” Proc. IEE Commun., vol. 153, no. 1, pp.
120-126, Feb. 2006.
K. Makaratat, T.W.C. Brown, and S. Stavrou, “Modified UWB spatio-
temporal channel simulation including pulse distortion and frequency
dependence,” IEEE Antennas and Wireless Propag. Lett., vol. 7, pp.
Z. Chenming, and R. C. Qiu, “Pulse Distortion Caused by Cylinder
Diffraction and Its Impact on UWB Communications,” IEEE Trans.
Veh. Technol., vol. 56, no. 4, pp. 2385-2391, Jul., 2007.
Z. Yuan, C. L. Law, L. G. Yong, and F. Chin, “Multi-target UWB
passive ranging with local template uncertainty,” in IEEE Proc. Int.
Conf. on Ultra-Wideband, 2008 (ICUWB 2008), Hannover, Germany,
Sep. 2008, pp. 233-236.
 A. Saleh and R. A. Valenzuela, “A statistical model for indoor
multipath propagation,” IEEE J. Sel. Areas Commun., vol. 5, no. 2, pp.
128–137, Feb. 1987.
 C. Chong, Y. Kim, and S. S. Lee, “A modified S-V clustering channel
model for the UWB indoor residential environment,” in IEEE Proc. on
Vehicular Technology Conference 2005, Stockholm, Sweden, May
2005, pp 58-62.
 Q. H. Spencer, B. D. Jeffs, M. A. Jenson, and A. L. Swindlehurst,
“Modeling the statistical time and angle of arrival characteristics of an
indoor multipath channel,” IEEE J. Sel. Areas. Commun., vol. 18, no.3,
pp. 347- 359, Mar. 2000.
 K. Haneda, J. Takada, and T. Kobayashi, “Cluster properties
investigated from a series of ultrawideband double directional
propagation measurements in home environments,” IEEE Trans.
Antennas Propag., vol. 54, no. 12, pp. 3778- 3788, Dec. 2006.
 R. J. M. Cramer, R. A. Scholtz, and M. Z. Win, “Evaluation of an
ultra-wide-band propagation channel,” IEEE Antennas Propag., vol. 50,
no. 5, pp. 561- 570, May. 2002.
 Y. Chen and V. K. Dubey, “Dynamic simulation model of indoor
wideband directional channels,” IEEE Trans. Veh. Technol., vol. 55, no.
2, pp. 417- 430, Mar. 2006.
 S. Gezici, A. Tian, G. B. Giannakis, H. Kobayashi, A. F. Molisch, H. V.
Poor, and A. Sahinoglu, “Location via ultra-wideband radios,” IEEE
Signal Process. Mag., vol. 22, pp. 70-84, Jul. 2005.
 H. Urkowitz, Signal Theory and Random Processes. Dedham, MA:
Artech House, 1983.
 K. Makaratat, S. Stavrou, and T.W.C Brown, “Simulation of UWB
distortion combined with indoor spatio-temporal channels,” in Proc.
IET Seminar on Wideband and Ultrawideband Systems and
Technologies, London, 6 Nov. 2008.