NICTA/ANU
Seminars Page
Presenter
: Dr. Jian (Andrew) Zhang, ANU
Date: Mon. 17th Oct. 2005, 1pm - 2pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Synchronization and Channel Estimation in Single-carrier UWB
Systems
Abstract:
Ultra Wideband (UWB) signals have very high temporal resolution
ability. This implies a frequency selective channel with rich multipath
in practice. Synchronization and channel estimation in such dense
multipath channels are very challenging problems. Complexity and
performance are generally hard to be balanced.
This talk will first give an overview of current synchronization and
channel estimation techniques for UWB systems. Limitation of these
methods will be highlighted. The Principal Components Tracking (PCT)
Algorithms will then be introduced, which are extensions of subspace
based algorithms. The PCT algorithms can track and identify $p$
principal multipaths out of $L$ multipath signals with significantly
reduced complexity, while retaining the high resolution and one-shot
realization abilities of subspace methods.
More Info:
http://rsise.anu.edu.au/~jian
Presenter
: Parastoo Sadeghi, ANU
Date: Wed. 12th Oct. 2005, 11am-12noon
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Information Capacity Analysis of Time-Varying Fading Channels
Using Finite-State Markov Channel Models
Summary: Finite-state Markov channels (FSMC) are often used to
model time-varying fading channels. In this talk, we present a brief
overview on the applications and applicability of FSMC models for
time-varying fading channels. Then, we analyse the information capacity
of time-varying fading channels using finite-state Markov channel
models. We discuss the effect of the number of fading channel gain
partitions on the FSMC capacity and study the FSMC capacity saturation
by increasing the number of fading channel gain partitions.
Furthermore, we consider the effect of fading channel memory order on
the capacity. It is observed that increasing the assumed channel memory
order may not result in monotonically increasing lower estimates of the
fading channel capacity. We explain this phenomenon by analytical
comparison of the capacities of high-order and low-order FSMC models
and attribute this phenomenon to the hidden nature of FSMC states in
the presence of channel noise.
Biography:
Parastoo Sadeghi received her B.Sc and M.Sc degrees from Sharif
University of Technology, Tehran, Iran in 1995 and 1997, respectively.
She has recently finished her PhD studies at the University of New
South Wales, Sydney. She is currently a research postdoctoral fellow at
the Australian National University, Canberra. Her research interests
include capacity, coding, and channel estimation for time-varying
fading channels, adaptive signal processing, and the information theory
of multiple antenna and multiuser systems.
More Info:
parastoo.sadeghi at rsise.anu.edu.au
Presenter
: Dr Fredrik Braennstroem (Chalmers
University of Technology)
Date: Tue. 4th Oct. 2005, 1pm - 2pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Multiple parallel concatenated codes with optimal puncturing and
energy distribution
Summary:
In this paper we show how to find optimal energy distribution together
with optimal puncturing ratios for parallel concatenated codes with two
or more constituent codes. The energy distribution and the puncturing
ratios are optimal in terms of minimizing the average signal-to-noise
ratio convergence threshold. The extrinsic information transfer
functions of the constituent codes are used for the optimization. Using
this technique we obtain additional degrees of freedom for constructing
codes with low convergence thresholds over a large range of code rates.
BIO:
See: http://www.s2.chalmers.se/~fredrikb
Presenter
: John M. Walsh, Cornell University.
Date: Mon. 19th Sept. 2005, 2pm-3pm
Location: Seminar room, ground floor, RSISE, building
115 cnr. North and Daley Roads, ANU
Title: The Where, When, and Why of Turbo Decoder Convergence
Summary:
Along with being one of the most prominent communications inventions of
the past decade, the introduction of turbo codes in [1] began a new era
in communications systems which brought them closer than ever to
theoretical performance limits. The creation of turbo codes introduced
a new method of decoding these codes which brought the decoding of
complex codes within
the reach of computationally practical algorithms. The iterative
decoding algorithm, while being suboptimal, performs well enough to
bring turbo
codes very close to theoretically attainable limits. An accurate
justification for why the decoding strategy performs as well as it does
is still lacking. This is exacerbated by the fact that the turbo
decoder, unlike most of the designs in modern communications systems
engineering, was not originally introduced as a solution to an
optimization problem.
This has made explaining just why the turbo decoder performs as well as
it does very di cult. Significant progress has been made with EXIT
style analysis [2] and density evolution [3], but these techniques
ultimately
appeal to results which become valid only when the block length grows
rather large. Other attempts, such as connections to factor graphs [4]
and belief propagation [5], have been largely unsuccessful at showing
convergence due to loops in the turbo coding graph. The information
geometric attempts [6], [7], [8], and [9], in turn have been inhibited
by inability to e ciently describe intrinsic information extraction as
an
information projection. None of these convergence frameworks, so far,
have identified the optimization problem that the decoder is attempting
to
solve. In this presentation I will show that the turbo decoder admits an
exact interpretation as a well known iterative method [10] attempting
to find a solution to a particular intuitively pleasing constrained
optimization problem. I will use this framework to give some conditions
for convergence of the turbo decoder. After explaining this framework
for the turbo decoder, I will briefly discuss extensions to belief
propagation in general factor graphs, contrasting my results with those
from the analogy to statistical physics [11], which are prominent in
that arena.
Biography:
John M. Walsh was born in Carbondale, IL on September 23, 1981. He
received the B.S. degree magna cum laude and the M.S. degree from
Cornell University, Ithaca, NY, in 2002 and 2004, respectively, where
he is
currently pursuing the Ph.D. degree. His research interests include the
convergence and performance analysis of turbo decoding and belief
propagation, interactions between interconnected adaptive devices, and
channel shortening. Mr. Walsh is a member of Tau Beta Pi and Eta Kappa
Nu.
References:
[1] C. Berrou, A. Glavieux, and P. Thitimajshima, Near shannon
limit error-correcting coding and decoding: Turbo-codes., in ICC
93, Geneva, May 1993, vol. 2, pp. 1064 1070.
[2] S. ten Brink, Convergence behavior of iteratively decoded
parallel concatenated codes., IEEE Trans. Commun., vol. 49, pp.
1727 1737, Oct.2001.
[3] H. El Gamal and A. R. Hammons, Jr., Analyzing the turbo
decoder using the gaussian approximation, IEEE Trans. Inform.
Theory, vol. 47, pp. 671 686, Feb. 2001.
[4] F. R. Kshischang, B. J. Frey, and H.-A. Loeliger, Factor
graphs and the sum-product algorithm, IEEE Trans. Inform. Theory,
vol. 47, pp. 498 519, Feb. 2001.
[5] R. J. McEliece, D. J. C. MacKay, and J.-F. Cheng, Turbo
decoding as an instance of pearls belief propagation algorithm.,
IEEE J. Select. Areas Commun., vol. 16, pp. 140 152, Feb. 1998.
[6] M. Moher and T. A. Gulliver, Cross-entropy and interative
decoding., IEEE Trans. Inform. Theory, vol. 44, pp. 3097 3104, Nov.
1998.
[7] T. Richardson, The geometry of turbo-decoding
dynamics., IEEE Trans. Inform. Theory, vol. 46, pp. 9 23, Jan.
2000.
[8] S. Ikeda, T. Tanaka, and S. Amari, Information geometry of
turbo and low-density parity-check codes, IEEE Trans. Inform.
Theory, vol. 50, pp.1097 1114, June 2004.
[9] J. Walsh, P. Regalia, and C. R. Johnson, Jr., A refined
information geometric interpretation of turbo decoding, in
International Conference on Acoustics, Speech, and Signal Processing
(ICASSP), Philadelpha, PA,
Mar. 2005.
[10] J. Walsh, P. Regalia, and C. R. Johnson, Jr., A convergence
proof for the turbo decoder as an instance of the Gauss-Seidel
iteration, in IEEE International Symposium on Information Theory,
Adelaide,
Australia, Sept. 2005.
[11] J.S. Yedidia, W.T. Freeman, and Y.Weiss, Constructing
free-energy approximations and generalized belief propagation
algorithms, IEEE Trans. Inform. Theory, pp. 2282 2312, July 2005
More Info:
Presenter : Dr.
Bijan Rohani, WATRI
Date: Wed. 24th Aug. 2005, 1pm-2pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Frame Quality Feedback for Real-Time Perceptual Quality
Estimation
Summary:
In many modern mobile radio networks, accurate speech quality
measurements are required for a variety of reasons. These range from
daily network maintenance to resource management through power control
and link adaptation. While objective perceptual speech quality
measurement techniques have been improved significantly over the last
decade, their usage have been restricted to non real-time applications
such as codec tests. The main reason for this has been the dependence
of these speech quality measurement techniques on availability of the
original speech signal - in addition to its received version - so it
can be used as reference for accurate quality estimation. However, the
reference signal cannot be provided in the receiver in real-time
applications such as power control in a modern mobile communication
system. Consequently, metrics such as Carrier-to-Interference ratio
(C/I), average Bit Error Rate (BER), average Frame Erasure Rate (FER)
and a variety of other channel quality metrics have commonly been used
for estimation of speech quality. It should be noted that speech
quality estimation techniques based on channel metrics are not truly
perceptual as they do not utilize any acoustic information for quality
estimation. As such, these techniques give a statistical expectation of
the speech quality rather than the actual perceptual speech quality.
In this seminar, a real-time speech quality estimation technique for
wireless applications is described. Unlike its commonly used
counterparts, this technique utilizes the acoustic information present
in the received signal. Therefore, the estimated speech quality is a
true measure of the perceptual quality rather than its statistical
expectation. This technique is based on the quality of the encoded
speech frames at the receiver.
Biography
Bijan received his BEng and PhD from Curtin University of Technology in
1992 and 1999, respectively. From 2000 to 2002, he was with the Genista
Research in Singapore where he worked on algorithms for monitoring and
control of voice quality in GSM and UMTS WCDMA systems. His focus was
to develop and use perceptual metrics to replace conventional so-called
channel quality metrics. From 2003 to July 2005 he was working at
Curtin University in Malaysia before joining WATRI recently. Bijan’s
interests are in wireless communications and communications signal
processing with particular interest in applications of perceptual
quality measures in communications.
More Info:
http://www.watri.org.au/home-watri/home.html
Presenter : Prof. Li
Ping, City University, Hong Kong
Date: Wed. 24th Aug. 2005, 3pm-4pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Interleave-Division Multiple-Access (IDMA) for Future
Wireless Systems
Summary:
This talk will outline a new multiple access scheme known as
interleave-division multiple-access (IDMA) that employs interleaving as
the only mechanism to distinguish users. IDMA possesses many desired
features for future wireless systems which are difficult to achieve
simultaneously with the current technologies such as FDMA, TDMA, CDMA
and OFDMA. These include
- very low receiver cost (near single-user turbo-receiver complexity),
- near optimal multi-user performance,
- de-centralized (i.e., asynchronous) control,
- simple treatment of ISI,
- cross-cell interference mitigation,
- wide-band diversity against fading,
- flexibility for multi-rate services (e.g., mixed voice and IP),
- high capacity (e.g., supporting 64 users with a spreading ratio of
only 8),
- high throughput (e.g., rate > 8 bits/Hz),
- excellent power efficiency (close to limit),
- straightforward extension to multiple antenna systems.
We will provide both analytical and simulation results to confirm the
features listed above.
About the Speaker
Dr. Li Ping received his Ph.D. degree at Glasgow University in 1990. He
lectured at Department of Electronic Engineering, Melbourne University,
from 1990 to 1992, and worked as a research staff at Telecom Australia
Research Laboratories from 1993 to 1995. He has been with the
Department of
Electronic Engineering, City University of Hong Kong, since January
1996 where he is now a professor. Dr. Li Ping was awarded a
British Telecom -
Royal Society Fellowship in 1986, IEE J J Thomson premium in 1993 and
the Croucher Foundation Award in 2005.
More Info:
http://www.cityu.edu.hk/cityu/about/professors/fse-ee-li.htm
Presenter : Prof.
Robert Heath, University of Texas at Austin
Date: Mod. 22th Aug. 2005, 2pm - 3pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Grassmann Quantization and MIMO Beamforming Systems
Summary:
Antenna arrays at the transmitter and at the receiver can be used to
create a multiple input multiple output (MIMO) wireless channel that can
be used for data transmissions at high rate with high reliability. It
is known that the knowledge of the matrix channel between the multiple
transmit and receive antennas is crucial to the overall system
performance. In practical systems, however, quantized channel state
information (CSI) may be available at the transmitter through a
feedback link. It is observed that the problem of quantization of CSI is
appropriately formulated as a quantization problem on a non-Euclidean
space, the complex Grassmann manifold. This motivates the
characterization
of distortion rate functions of a `Grassmann Quantizer'.
This talk will introduce a class of quantization problems on the
Grassmann manifold and present high resolution distortion rate results
for
the particular case of a complex projective space. Subsequently, a MIMO
beamforming system will be investigated with quantized CSI at the
transmitter. Specifically the performance of such a system will be
characterized as a function of the feedback rate. This talk will touch
upon
the necessity of analysis and algorithm design in non-linear manifolds
to account for the constraints that arise in MIMO wireless communcation.
BIO
--------
Robert W. Heath, Jr. is an Assistant Professor in the Department of
Electrical and Computer Engineering at The University of Texas at Austin
and is a member of the Wireless Networking and Communications Group
where he directs the Wireless Systems Innovations Laboratory. He
received his
B.S.E.E. (1996) and his M.S.E.E. (1997) degrees from the University of
Virginia, and the Ph.D.E.E. (2002) degree from Stanford University. He
is
also president of MIMO Wireless Inc, a consulting company dedicated to
the advancement of MIMO technology. Dr. Heath's research interests
include
wireless communication theory and signal processing. Currently he is
focusing on all aspects of MIMO communication including antenna design,
practical receiver architectures, limited feedback techniques, mobility
management, and scheduling algorithms. He is an Associate Editor for the
IEEE Transactions for Communications and the IEEE Transactions on
Vehicular Technology.
More Info:
http://www.ece.utexas.edu/~rheath/
Presenter : Prof.
Mike Faulkner, Victoria University, Melbourne
Date: Friday 12th Aug. 2005, 11am-12noon
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Wireless Local Area Networks – The Next Generation
Summary:
The IEEE 802.11 community is currently standardizing the physical layer
for the fourth generation of wireless local area networks
(W-LAN).
Some 32 partial and complete proposals were presented by various
organizations at the September 04 standardization meeting. These have
now coagulated into two groups. The new 802.11n standard requires an
enhanced throughput of at least 100Mbit/s as seen on top of the MAC
layer. This is over three times the current rate. The
standard must
be backward compatible to the existing ‘b’, ‘g’ and ‘a’ variants. It is
expected to incorporate both OFDM and MIMO (Multiple Input Multiple
Output) features.
The talk will give a brief technical overview of the current WLAN based
standard (802.11a) and then present the MIMO channel model to be used
for comparing the different proposals. One of the 802.11n technical
proposals will then be described. The talk will end by
summarizing the
key technical aspects and the more interesting features from some of
the other consortiums.
Michael Faulkner, is Professor in the Centre for Telecommunications and
Micro-Electronics at Victoria University in Australia. His
research
group is part of the Australian Telecommunications Co-operative
Research Centre (ATcrc), a mixture of academic and industrial
organizations. He leads the multiple antenna project which
was
responsible for 802.11n activities. His research interests
include
Wireless Systems, Wireless Implementation and Signal Processing.
More Info:
http://www.ctme.vu.edu.au/People.htm
Presenter : Assoc.
Prof. Leandro de Haro,
Radiation Group, Dept. of Signals, Systems and Radio-Communications
School of Telecommunication Engineering
Technical University of Madrid
Tadeus Wysocki, University of Wollongong.
Date: Wed. 10th Aug. 2005, 2pm-3pm
Location: Fishbowl, NICTA Northbourne Ave,
Canberra (map)
Title: Research Activities at University of Madrid in Antenna Design
and Signal Processing for Communications
Summary:
The presentation will be focused on the research activity that is been
performed by the Radiation Group of Dept. Signals, Systems and
Radio-communications of the Technical University of Madrid. Initially,
a review of general research activates of the Radiation Group will be
done. The Radiation Group has mounted and manages the three anechoic
chambers at the premises of the Telecom Faculty. Some details of these
anechoic chambers will be shown. Following the presentation, smart
antenna concept will be reviewed. Smart antenna covers several advanced
antenna concepts: phase arrays (active or not), reconfigurable
antennas, beam-switched antennas, adaptive arrays and finally MIMO
(Multiple-input-Multiple-output) antennas. Smart antenna may be studied
from a signal processing point of view, dealing with beam-forming
algorithms or coding aspects. An antenna theory point of view should
also be taken into account. However, the prototyping needs to take to
account considerations of the technological aspects from both signal
processing and antenna theory. The presentation will show the research
activities performed in the Radiation Groups on such kind of antennas
emphasising the prototyping aspects.
Leandro de Haro
He achieved the engineer degree in 1986 and the PhD degree in 1992 both
from the Telecommunication Faculty of Technical University of Madrid
(UPM). Since 1990 he has been with the Dept. of Signal, Systems and
Radio-communications (SSR) of the UPM first as an Assistant and then as
Associate Professor in the Signal Theory and Communications area. He is
also an IEEE member and a Chairman of the AP/MTT Spanish Chapter.
His research activity has been developed within the Radiation Group
(GR) of the Dept. SSR in the area of antennas and propagation and their
impact on the system design. He has collaborated in the design of
advanced antennas for satellite communications (for earth stations and
satellite on board), Digital TV (relays and handheld terminals) and
wireless systems (base stations and user terminals). He has also
participated in planning activities both for Digital TV and wireless
deployment.
He has been actively involved in several projects official and with
private companies (national and international). He has also been
involved in several European projects (RACE, ACTS, IST (IPs and NoEs)
and COST).
The results of his research activity may be found in several
presentations in national and international conferences as well as in
the published papers.
More Info:
http://www.gr.ssr.upm.es/staff/ldh_uk.htm
Presenter : Rasika
Perera, PhD Student, RSISE/NICTA, Mid-term
Presentation
Date: Thur. 7th July 2005, 10am-11am
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: What can we get without channel state information in SISO
and MIMO Rayleigh fading channels
Summary:
The capacity and optimal input distribution of non-coherent Rayleigh
fading multiple input multiple output (MIMO) channels has been an open
problem for some time. Proper knowledge of channel capacity and the
optimal input distribution motivates researchers to develop powerful
code such as Turbo, which enable us to operate near Shannon limits.
Unlike non-fading channels, coherent Rayleigh and Rician fading
channels, finding capacity and the optimal input distribution of
non-coherent Rayleigh fading channels is considered as a difficult
problem in the literature.
In the first part of this talk, we present the performance of Gaussian
signaling in non-coherent single input single output (SISO) and
MIMO Rayleigh fading channels. Using the closed form solutions
established for mutual information, the performance at any signal to
noise ratio (SNR) for any transmit and receive antenna numbers is
emphasized.
The second part is mainly on the achievable capacity in non-coherent
Rayleigh fading MIMO channels with some significant properties of the
optimal input distribution. The supremum of the capacity for any
transmit and receive antenna numbers is shown using convex
optimisation. The actual capacity and the optimal input distribution
found based on our recent work is discussed at the end.
More Info:
Rasika.Perera AT anu.edu.au
Presenter : Tharaka
Lamahewa, PhD Student, RSISE/NICTA, Mid-term
Presentation
Date: Wed. 22nd June 2005, 10am-11am
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Space-time coding: Perfromance analysis and spatial pre-coder
designs for optimal antenna placement.
Space-time coding combines channel coding with multiple transmit and
multiple receive antennas to achieve bandwidth and power efficient high
data rate transmission over multipath fading channels. Space-time
coding is considered to be the first systematic treatment of coding for
achieving transmit diversity. Most of existing space-time code
construction methods have assumed that channel gains between transmit
and receive antennas are spatially uncorrelated. However, in a
realistic multipath fading channel environment, the assumption of
uncorrelated fading is far too unrealistic due to the non-isotropic
nature of the surrounding scattering environment and insufficient
spacing between antenna elements.
In the first part of the talk, we present analytical performance
results of several space-time codes for different spatial scenarios
using a general spatial channel model which fully accounts for: i)
antenna placement and separation ii) scattering distribution parameters.
The spatial channel model considered in the first part of the talk
decomposes the underlying channel between transmitter and receiver
arrays into deterministic and random parts, where the deterministic
part is related to the antenna configurations (spacing and geometry,
which are fixed and a priori known to the transmitter and receiver) and
the random part is related to the scattering environment surrounding
the receiver and transmitter antenna arrays. In the second part of the
talk, we present two linear spatial pre-coders which reduce the effect
of insufficient antenna spacing on the performance of multi-antenna
communication systems, where the pre-coders are derived based on the
deterministic part of the channel.
More Info:
Tharaka.Lamahewa AT anu.edu.au
Web: http://web.rsise.anu.edu.au/~tharaka/
Presenter : Pawel
Dmochowski, PhD Candidate, Queen's University at
Kingston, Canada
Date: Mon. 6th June 2005, 1pm-2pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Robust Timing Epoch Tracking for OSTBC Receivers in Rayleigh
Fading MIMO Channels
As in single-antenna synchronous communication systems, timing
synchronization is paramount to the performance of MIMO
receivers. Timing synchronization methods recently developed
utilize a training preamble
to compute the likelihood function which is in turn optimized with
respect to the timing parameter to find the optimum sampling
point. In
addition to the data overhead, the computational complexity of such
methods calls for approximations. In the case of single antenna
systems, such
approximations led to the development of Timing Error Detectors (TED's)
which provide timing information with little computational complexity.
These
estimators, however, are not suitable for transmit diversity
systems. The talk will focus on a low-complexity timing error
detector for Alamouti OSTBC
receivers in Rayleigh fading environment. The detector is derived
by modifying a well known Mueller and Muller detector originally
proposed for SISO ISI channels. Theoretical performance metrics are
evaluated, specifically the S-curve and the estimator variance.
The
former is shown to be independent of channel fading, thus demonstrating
a key property of robustness to poor channel conditions. The
decision-directed version of the TED is then used
to evaluate the receiver timing tracking performance by means of
simulations, including those incorporating pilot-based channel
estimation. In
addition to the symbol error rate curves, the range of timing drift
bandwidth tracked as well as the effects of PSAM channel estimation are
studied.
Finally, extensions to higher diversity orders are discussed.
More Info:
www.paweldmochowski.com
Presenter : Michael
Williams, NICTA, Wireless Signal Processing
Program
Date: Fri. 3rd June 2005, 11am-12noon
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Fundamental Limits to Spatial Signal Processing
Abstract:
The conventional approach to temporal signal processing is based on an
understanding of signals in continuous time. Any sampling in the
system is then treated as an artifact of implementation. By contrast,
the conventional approach to spatial signal processing in
communications research is to deal directly with signals that have been
spatially sampled by point antennas. The weakness of this
approach is that the underlying properties of the continuous field in
space have been ignored. In this talk we introduce a modal approach for
dealing with the properties of continuous wavefields in space and
time. This approach provides a fresh persepecitve on common
spatial signal processing problems such as direction of arrival
estimation, MIMO capacity, beamforming, and wavefield extrapolation.
More Info:
Michael is a PhD student in NICTA's Wireless Signal Processing (WSP)
program
http://online.anu.edu.au/RSISE/teleng/teleng2004/people/students/michaelw.php
Presenter : Alex
Smola, NICTA, Statistical Machine Learning Program
Date: Fri. 22nd April 2005, 10-11am
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Exponential Families for Estimation
More Info:
Alex.Smola at nicta.com.au (http://www.nicta.com.au)
http://users.rsise.anu.edu.au/~smola/home.html
Presentation
Presenter : Glenn
Dickins
Date: Thursday
April 21 1:00 pm
Location:
NICTA Board Room, 216 Northbourne Ave
Title:
Lost Space - Finding the Spatial Dimension
The demand for mobile data is increasing at an astonishing rate.
The area of Multiple-Input Multiple-Ouptut (MIMO) communications
systems has generated significant interest interest in recent years as
a way of increasing spectral efficiency of wireless communications by
'exploting the spatial dimension'. This seminar will review some
of the models that have been developed to analyse and design MIMO
systems and present some work aimed at improving the extent to which
these models capture the spatial dimension.
There are two main classes of model used for MIMO systems. The
discrete class uses a vector channel with the characteristics of the
spatial environment represented by a channel matrix. To simplify
the analysis, the complexity of space is discarded in
favour of correlated random matrices. More recently, models of
continuous spatial fields have been used to describe the multipath MIMO
environment. Whilst such models present fundamental limits for
spatial fields, it is difficult to apply an appropriate normalization
and noise model to link such models to the practical case of finite
antennae.
A useful model allows the measurement and incorporation of
a set of parameters to create a suitable prediction of the
behaviour a system without being burdened by unneccessary
detail. In this work we are seeking an appropriate model and
parameters for the spatial communications channel. Whilst
properties such as richness, diversity, dimensionality and spatial
noise have been discussed in the literature, the problem of defining
and applying these concepts in a consistent manner remains an open
research problem. Some intial results and steps towards solving
these problems will be presented.
More Info:
Presenter : Markus
Dangl, University of Ulm, Germany
Date: Wed. April 20th 2005, 1pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: "Markov Chain Monte Carlo Methods for Iterative (Turbo)
Multiuser Detection"
Abstract:
This work is being performed in co-operation with Mark Reed and
Zhenning Shi.
Recently, Markov Chain Monte Carlo (MCMC) methods have attracted a lot
of interest as promising solutions to both multiuser and MIMO detection
problems. Among them, approaches based on Gibbs sampling are especially
well suited due to their good trade-off
between performance and complexity. However, it is known that detection
methods using Gibbs sampling may suffer from a performance degradation
in the high SNR regime. This problem becomes severe, when the number of
samples is small and the system is overloaded, e.g., when the number of
users exceeds the spreading gain. In this talk, we discuss approaches
to overcome this degradation effect and present a comparison of MCMC
methods with known, standard multiuser detection techniques in terms of
both performance and complexity.
Talks from WATRI
Date: Thursday 14th April 2005, 2pm - 3pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
-------
Title: Techniques to enhance speech and suppress echoes in PDA:s and
Mobile
Phones or how to get working solutions out of array processing and
blind signal separation.
Speaker: Prof. Sven Nordholm, Director Signal Processing Laboratory,
WATRI, Perth
Techniques to enhance speech and suppress echoes in PDA:s and Mobile
Phones or how to get working solutions out of array processing and
blind signal separation.
The flexibility provided by hands-free communication devices have
revolutionized the way humans communicate with each other. Hand-held
communication devices will soon become outdated with everyday
essentials such as personal digital assistants (PDA) and mobile phones
becoming hands-free compliant. Nonetheless, when it comes to hands-free
system, there are several disadvantages. Since the user is at a distant
from the microphone, the microphone will also capture the background
noise (such as babble, etc.) as well as the interference due to the
hands-free loudspeaker. Therefore, processing techniques with both
noise and acoustic echo cancellation capability is instrumental in this
application. This paper presents array processing techniques that
jointly suppress the background noise and acoustic echoes.
Sven Nordholm(Professor, Research Director Signal Processing
Laboratory, Western Australian Telecommunication Research Institute,A
Joint Institute between The University of Western Australia and
Curtin University of Technology)
Dr. Sven Nordholm was born in 1960. He got his Ph.D. in Signal
Processing from Lund University in 1992, Licentiate of engineering 1989
and MscEE (Civilingenjör) 1983. He was one of the founders of the
Department of Signal Processing, Blekinge Institute of Technology in
Ronneby in 1990. Where he held positions as Lecturer, Senior Lecturer,
Associate Professor and Professor. Since 199 he has been in Perth,
Western Australia. From 1999-2002 he was director of ATRI and Professor
at Curtin University of Technology . Currently he is professor
and head of Signal Processing research WATRI, Western Australian
Telecommunication Research Institute, a joint institute between The
University of Western Australia and Curtin University of Technology. He
is also Research Executive of the Wireless Program, ATcrc.
His main research efforts have been spent in the fields of Speech
Enhancement, Adaptive and Optimum Microphone Arrays, Acoustic Echo
Cancellation, Adaptive Signal Processing, Sub-band Adaptive Filtering
and Filter Design.
Talk-2:
--------
Title: Optimization with application in signal processing and
communication
Speaker: Dr. Heidi Dam, Signal Processing Laboratory, WARI, Perth
Multi-rate signal processing is gaining more and more importance in
signal processing applications such as echo cancellation, microphone
arrays, speech enhancement and equalisation. The first part of the talk
is aimed at obtaining efficiency multi-rate processing designs by
developing appropriate problem formulations based on mathematical
foundations so that optimisation techniques can be applied. The
developed multi-rate systems will result in low complexity
implementations without significant compromise in performance in each
application. For the second part, we focus on the use of heuristic
optimization technique such as genetic algorithm in applications such
as DFE MIMO and CDMA systems.
%-----------------------------------------------------------------------------------------------------
Dr. Hai Huyen Dam received the Bachelor degree (first class Honours)
and PhD degree (with distinction) from Curtin University of Technology,
Perth, Australia in 1996 and 2001, respectively. From 1999 to 2000, she
spent one year at the Blekinge Institute of Technology, Sweden as a
Visiting Research Associate. Since 2001, she has been working as a
Research Fellow at the Western Australian Telecommunications Research
Institute (WATRI), Curtin University of Technology, Australia. Her
research interests are
array processing, optimization methods, equalization and filter design.
Links : http://www.watri.org.au/home-watri/home.html
Presenter : Elias
Aboutanios, Institute of Digital Communications,
School of Eng and Electronics, The University of Edinburgh
Date: Tue. 29th March 2005, 2:30-3:30pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Space-Time Adaptive Processing for Radar Target Detection
Abstract:
Space-time adaptive processing in the context of radar has been a
subject of research for more than 30 years. In the 1990, the increase
in computational power
provided additional impetus to the research. The problem is that of
detecting a signal in interference. The radar system comprises an
antenna array that collects
a data cube of spatial and temporal samples. The spatial-temporal data
allows the spectrum to be resolved both in range and cross range. This
provides the ability
of resolving low velocity targets which would be, in the case of a
single sensor radar, buried in the clutter.
The detection process is usually implemented as a filter whose output
power is compared to a suitably chosen threshold. The optimal filter
assumes a-priori
knowledge of the interference covariance matrix. In practice however,
the interference covariance matrix is not known and must be estimated.
Traditional
approaches such as the generalised likelihood ratio test assume the
availability of target free training data that is statistically
homogeneous with the test data. This
training data is usually drawn from neighbouring range gates to the
test gate.
It has been recognised for some time now that radar clutter returns can
be in-homogeneous resulting in a degradation in the performance of the
traditional detectors. This has lead to research into approaches such
as knowledge-based STAP which aims at incorporating a priori
information on the terrain into the data processing. At the University
of Edinburgh, we are examining alternative detectors are do not require
independent training data. These detectors can act as standalone
processors or within a KB-STAP framework. This talk will provide an
overview of space-time adaptive processing as well as report on our
research results in the radar signal processing group.
Links : elias at ieee.org
Presenter : Leif
Hanlen
Date: Wed. 16th March 2005, 1-2pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Continuous Gaussian Channels
Abstract:
What is the information content of space? How does this relate to
typical MIMO results?
We present information theoretic capacity results for information
transfer in space, under the general assumption that the receiver
applies a spatial filter to receive signals. Our work presents
novel applications of linear, bounded, invariant operators to
communication channels. We outline the procedure for calculating
the information theoretic capacity for this restricted class of
operator channels and apply this result to wireless information
transfer. This seminar presents ongoing work within the WSP group.
Links : Leif.hanlen at nicta.com.au
Presenter :
Deepshikha Garg, Tohuku University, Japan
Date: Tue. 15th Feb. 2005, 2pm
Location: Meeting Room (Fish Bowl), 216 Northbourne Ave,
Canberra (map)
Title: DS-CDMA with Frequency-domain Equalization and Adaptive
Modulation and Coding for High Speed Downlink Packet Access
Abstract:
Third generation (3G) mobile communications networks based on direct
sequence code division multiple access (DS-CDMA) have been successfully
launched. Data services, as anticipated, have become the dominating
source of traffic load in the 3G networks. High speed data services
will be supported by the high speed downlink packet access (HSDPA)
technique, which allows peak data rates up to around 10Mbps. Adaptive
modulation and coding (AMC), hybrid automatic repeat request (HARQ) and
multicode operation are the enabling technologies used for HSDPA. The
conventional DS-CDMA receivers consist of a rake combiner that can take
advantage of the path diversity. The next generation mobile
communications system is anticipated to support even higher data rates
up to and exceeding 100Mbps. With such high speed data transmissions,
the wireless channel becomes severely frequency-selective. AMC, HARQ
and multicode operation will still be inevitable. However, when the
number of propagation paths in the channel increases, the receiver
complexity increases due to the increase in the number of rake fingers.
Moreover, in a frequency-selective channel, multicode operation
severely suffers from the loss of orthogonality among the orthogonal
spreading codes and the performance with coherent rake combining
severely degrades.
Recently, it was shown that DS-CDMA using frequency-domain equalization
based on minimum mean square error criterion (MMSE-FDE) can partially
restore the orthogonality and provide a better bit error rate (BER)
performance than conventional rake combining. The BER performance of
multicode DS-CDMA with MMSE-FDE improves with the increase in the
frequency selectivity of the channel. For packet transmissions, the
frequency selectivity of the channel is not always desirable. With
higher frequency selectivity, the errors are randomized; however, for
packet transmissions burst errors are preferable to random errors.
Hence, there is a need to evaluate the performance of packet
transmissions in a frequency selective channel. We apply MMSE-FDE for
receiving the DS-CDMA packet signals transmitted with AMC, multicode
operation and HARQ but over a severe frequency-selective channel. When
the same packet is transmitted more than once, time diversity gain is
obtained similar to antenna diversity. Hence, the MMSE weight needs to
be updated with each retransmission to benefit from the time diversity
gain. We propose MMSE weight for packet combining. We consider HARQ
schemes using both Chase combining (CC) and incremental redundancy
(IR). In CC, the retransmitted packets are combined to increase the
received signal power. In IR, the redundancy is increased with each
retransmission, resulting in a decreased coding rate and thus a better
error correction capability. It is found that the use of FDE for the
reception of multicode DS-CDMA packet gives an improved throughput much
higher than that with rake combining in frequency selective channels.
It was also found that with adaptive modulation and coding, there is a
very small advantage of incremental redundancy over Chase combining
employing frequency-domain packet combining based on MMSE criterion.
Links : http://www.mobile.ecei.tohoku.ac.jp/adachi-e/index.html
Presenter : Markus
Dangl, University of Ulm, Germany
Date: Tue. 8th Feb. 2005, 1pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Block Turbo Equalization for Imperfect Channel State Information
Abstract:
We consider a coherent, coded data transmission with perfect
synchronization over an intersymbol interference (ISI) channel, where
the channel impulse
response is a-priori unknown at the receiver. The optimum receiver with
respect to minimum bit error rate requires a tree search and performs
joint channel estimation, equalization, and decoding. In most cases
this optimum joint receiver is impractical due to its tremendous
computational complexity. In order to reduce complexity we focus on an
iterative receiver design, including joint iterative equalization and
decoding (turbo equalization), as well as optionally, iterative
data-aided (decision-directed) estimation. As key part of such a
receiver we propose a soft-input soft-output block equalizer based on
constrained minimum variance filters that is suitable for arbitrary
signal constellations. In contrast to other schemes, our design
furthermore includes the case of having imperfect channel state
information at the receiver by incorporating the statistics of the
channel estimation error. Thereby, we distinguish between data
independent and data dependent evaluation of the covariance matrix of
the estimation error.
We show that the performance of our block equalizer is significantly
improved when considering the influence of the actual data symbols on
the estimation error. The additional complexity compared to the case
when ignoring the data dependency of the estimation error increases
quadratically with the block length. We assess the performance of our
equalizer for block Rayleigh fading channels via Monte Carlo simulation
and EXIT chart analysis, employing turbo equalization with both
one-shot training-based and iterative data-aided channel estimation.
Links : http://it.e-technik.uni-ulm.de/~dangl/
Presenter : Zhenning
Shi
Date: Wed. 15th Dec. 2004, 1pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Efficient Markov Chain Monte Carlo Algorithms for Iterative
Detection of Multiple Access Channels
Abstract:
Markov Chain Monte Carlo (MCMC) are known to address optimization
problems involving high-dimension summations and integrals with
execellent efficiency. We look into the application of MCMC algorithms,
notably Gibbs samplers, for iterative detection of multiple-access
channels. Starting from the Monte-Carlo intergration approach, we
propose a few algorithms that can support very heavy system loads and
advanced modulation schemes for CDMA/MIMO channels. Special procedures
are put forward in the algorithm to improve robustness at high SNRs, a
scenario that typically leads to slow convergence of Markov
chains. MCMC methods are compared to other competing techniques
such as the MMSE filter and (list) sphere decoding for performance and
complexity. Recent progress on hardware design, coding design is
also briefed.
Links : Presentation
Zhenning.Shi at nicta.com.au
Presenter : Rasika
Perera
Date: Wed. 8th Dec. 2004, 1pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Bounds on Mutual Information of Rayleigh Fading Channels with
Gaussian Input
Abstract:
An independent and identically distributed (iid) Gaussian input
maximises the capacity of an additive white Gaussian noise (non-fading)
channel, a Rayleigh fading channel when the Channel State Information
(CSI) is perfectly known at the receiver, and when the CSI is known to
both the transmitter and the receiver.
However, when CSI is not known by neither the transmitter nor the
receiver, capacity achieving distribution is not Gaussian. Therefore,
it is a question of practical interest to find the achievable
information rate of Rayleigh fading channels when the input is Gaussian
distributed.
We consider the mutual information of a discrete time Rayleigh fading
channel, where neither the transmitter nor the receiver has the
knowledge of the channel state information. We specifically derive a
lower bound for the mutual information of this channel when the input
distribution is Gaussian. The bound is expressed in terms of the
capacity of the corresponding non fading channel and the capacity when
the perfect channel state information is known at the receiver.
Links :Presentation (2Mbytes)
Rasika.Perera at anu.edu.au
Presenter :
Kandeepan Sithamparanathan
Date: Mon. 6th Dec. 2004, 1pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Single Frequency Carrier Synchronisation
Abstract:
The problem of single-frequency carrier synchronisation is a well
treated issue in the literature. Here, we revisit the problem, to
enhance the performance of the synchronisation system, given the
requirements, on an application specific basis. The existing techniques
on feedback single frequency synchronisation have trade-offs in terms
of acquisition and tracking performances. In order to address the above
problem some nonlinear signal processing techniques are adopted to
achieve wider acquisition ranges and extended jitter-threshold points
in feedback loops. The loops are analysed both in its acquisition and
tracking modes and signal processing solutions are proposed to enhance
the performances.
Links :Kandeepan.Sithamparanathan at nicta.com.au
Presenter: Dr.
Chandra Athaudage
Date: Tue. 23rd Nov. 2004, 2pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Enhanced MMSE Channel Estimation Using Timing Error
Statistics for Wireless OFDM Systems
----------------------------------------------------------------------------
Abstract:
Estimation and tracking of the frequency-selective time-varying channel
response is a challenging task for wireless communication systems
incorporating coherent OFDM. In pilot-symbol-assisted (PSA) OFDM
systems, the minimum mean-square-error (MMSE) estimator provides the
optimum performance based on the channel statistics (channel
correlation function and SNR). In OFDM systems, FFT-block timing error
introduces a
linear phase rotation to data modulated on individual subcarriers. An
MMSE channel estimator designed only using the wireless channel
statistics performs only sub-optimally when subcarrier phase rotations
due to block timing errors are present. In this paper, we show that by
using the block timing error statistics of the OFDM time-synchronizer
the performance of the MMSE channel estimation can be significantly
improved. Numerical results show that both the channel estimation error
and the BER performance degradation due to timing errors can be almost
completely recovered by the proposed technique.
Links: http://www2.ee.unimelb.edu.au/people/Athaudage.html
Presenter : Leif
Hanlen
Date: Mon. 15th Nov. 2004, 1pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Messages from the frontier: Wireless Futures, a
brief and hopefully controversial report.
Abstract:
"Wireless research in the US is dead, physical layer research is no
longer sexy."
-- quote from wireless researcher
So what is sexy in telecommunications research? This talk is an open
ended non-technical discussion about what is happening, and predicted
to happen in the US markets over the next 6-12months. What are the
markets doing? What's taking interest in the US, and how is general the
environment? In particular, I will bring together disparate viewpoints,
from three separate parts of my recent 3-week trip to the US, including
an industry forum for wireless future research, the information theory
workshop and a visit to the Wireless Communication & Networking
Group
at University of Texas, Austin.
Links : leif.hanlen at nicta.com.au
Presenter : Zhenning
Shi
Date: Thu. 10th Sept. 2004, 2pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: Performance Analysis of Power Allocation for Iterative
CDMA Reception
Abstract:
The combination of forward error control coding (FEC) with code
division multiple access (CDMA) using random spreading sequences is
considered. Through tracking the input-output variance evolutions of
the decoding components, which are the CDMA interference resolution
function and the individual error control decoders, an analysis of the
iterative joint detection was
previously presented. It was shown this variance transfer (VT) analysis
between component decoding devices gives very accurate description of
the convergence process of the iterative joint detector for a CDMA
system with all users received at equal energy. We extend the VT
analysis to the more realistic cases where the different users have
unequal power levels. It is shown that the effective VT functions
modified w.r.t. the unequal power scenario can be used to analyze the
iterative reception employing a wide variety of linear filters for
interference resolution, in a manner very similar to that in the equal
power case. Some results for power allocation are shown via this
analysis approach.
Links : shizn at mail.rsise.anu.edu.au
Presenter : Chrisitan
Pietsch - University of Ulm
Date: Mon. 6th Sept. 2004, 2pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title; Space Time Block Coding and related topics
Abstract:
Links : http://it.e-technik.uni-ulm.de/~pietsch/
Presenter :
Prof. G. Corazza
Date: Fri. 3rd Sept. 2004, 3pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: "Challenging communications paradigms in the information society"
Abstract:
The talk will start by describing the most distinct trends in the
information society, showing how telecommunications networks are
intertwined with the epochal transformation from industrial to
post-industrial society. Then, a discussion on
the necessary innovation methodologies for progress in the information
society will lead to application examples
in the field of communications theory. In particular, novel
post-detection integration techniques will be
presented for spread spectrum communications.
Links : gecorazza at deis.unibo.it
Presenter :
Prof. Lars Rasmussen
Date: Thu. 25th August. 2004, 3pm
Location: Board Room, 216 Northbourne Ave,
Canberra (map)
Title: QoS-Based Adaptive Coding for Concatenated Wireless
Network Design
Abstract:
With the increasing demands of advanced wireless data services, the
ongoing challenge is to design new wireless networks able to
accommodate high data rate applications with a wide range of
Quality-of-Service (QoS) requirements. Furthermore, future wireless
networks must be able to provide such services in a harsh transmission
environment with large densities of highly mobile users. These
challenges call for vast improvements in the efficient use of resources
such as bandwidth, power, space and complexity across all layers in a
network design.
In the physical layer, joint designs of concatenated
transmitter/receiver components have provided significant improvements.
The invention of sub-optimal iterative receiver structures, exchanging
soft information between concatenated components has been particularly
successful, leading to near-optimal solutions with practical
computational complexity. The concepts of concatenated systems and
iterative soft information exchange can be adopted for wireless network
design, moving beyond current design paradigms. The iterative exchange
of information enables a high level of flexibility, and a framework for
cross-layer design and optimisation.
Our first aim is to provide a high level of flexibility for resource
allocation, by adopting adaptive principles in the underlying
communications protocols. In this talk, a novel QoS-based adaptive
coding scheme is proposed, integrating physical layer error control
coding, data link layer retransmission schemes, and parts of network
layer QoS responsibilities. This component is a core element in a
concatenated design, interfacing to scheduling/multiple access
functionalities in higher layers and to multiple antenna systems in
lower layers.
The talk will focus on design principles and tools, illustrated by
examples taken from ongoing research.
Links: www.itr.unisa.edu.au/~lars,
Lars.Rasmussen at unisa.edu.au
Presenter : James
McGowan
Date: Thu. 29th April 2004, (Time to be confirmed)
Location: (To Be Confirmed) Board Room, 216 Northbourne Ave,
Canberra (map)
Title: The Graphical Nature of LDPC Codes
Abstract:
Most of the currently used coding systems are too complicated to be
decoded exactly, and so instead iterative decoding techniques that
(usually) converge to the correct codeword are used. The codes are
represented as a graph made of lines and dots, and the decoding
involves passing belief probabilities along the lines until a stable
solution is reached.
Whether a code is good or bad depends principally on the structure of
the underlying graph. The number of loops, as well as their locations
relative to each other, affect the overall performance. Little has been
studied beyond the basics in this area. I will present my algorithm for
removing loops from codes, as well as some more recent research
concentrating on the individual nodes and their immediate neighbours.
Links : James McGowan (james.mcgowan at anu.edu.au)
Presentation (pdf)
Presenter : David
Smith
Date: Wed. 14th April 2004, 11am-12noon
Location: Board Room, 216 Northbourne Ave, Canberra (map)
TITLE: On Differential/Non-coherent Space-Time Modulation across MIMO
radio channels
Differential Space-Time Modulation (DSTM) schemes across multiple-input
multiple-output (MIMO) radio channels do not require explicit
estimation of the channel at the receiver, and with such DSTM schemes
there is generally less overhead. This presentation will contain a
brief overview of non-coherent space-time modulation across MIMO radio
channels, and concentrate on the presenters research during PhD.
candidature into modulation and detection of DSTM, principally based on
his research into differential unitary space-time modulation (DUSTM).
Links : Presentation
(pdf)
David Smith (David.Smith at nicta.com.au)
Presenter : Jeremy
Roberson
Date: Fri. 2nd April 2004, 2pm - 3pm
Location: Board Room, 216 Northbourne Ave, Canberra (map)
TITLE: Blind Subspace Channel Estimation and Equalization of Punctured
Retransmissions
In this work I will talk about the simple blind subspace algorithm, and
modifications to it. To use a Second Order Statistics (SOS)
channel estimation method, the row rank of the channel matrix must
exceed the column rank. This is done via temporal or spatial
diversity. In this work, we consider temporal diversity within a
network via ARQ. Here we use the retransmission as our necessary
temporal diversity, but we puncture the retransmission to save
bandwidth. By puncturing the retransmission, we no longer need
the coprime property for blind estimation, and we can achieve similar
channel estimation performance to a full retransmission. My
current work is using OFDM, MAP, linear, and DFE equalizers to receive
a punctured retransmission.
Links : Jeremy Roberson (jlrobe at ece.ucdavis.edu)
Presenter : Tharaka
Lamahewa
Date: Thur. 1st April 2004, 2:30pm - 3:30pm
Location: Board Room, 216 Northbourne Ave, Canberra (map)
TITLE: Fading Resistance of Orthogonal Space-Time Block Codes Under
Spatial Correlation
In this talk, the impact on the bit-error rate performance of
orthogonal space-time block codes STBC) due to spatial correlation is
investigated. An analytic model for spatial correlation is used which
fully accounts for antenna separation, antenna placement and
surrounding scattering distributions (Isotropic, Uniform-limited,
von-Mises, etc). The effect of antenna separation on orthogonal STBC is
examined for various non-isotropic azimuth power distributions in the
presence of spatial
correlation. We show that the impact of the space is limited on the
bit-error rate performance of orthogonal STBC, that is, most of the
bit-error rate improvement can be attributed to 'time-coding' rather
than to 'space-coding'. We also investigate how the non-isotropic
parameters of an azimuth power distribution, including the angular
spread and the mean angle of arrival (AOA) of an impinging signal
effect the bit-error rate performance of orthogonal STBC. Finally, an
expression for the antenna separation distance is given where the
performance of orthogonal STBC code is sufficiently close to optimal
under a given scattering environment.
Links : Tharaka.Lamahewa at anu.edu.au
Presenter : Kim
Holburn RSISE
Date: Wed 24 March 2004, 2pm - 3pm
Location: RSISE Seminar Room, ground floor, Building 115 cnr.
North and Daley Roads, ANU
Title: "Towards a single sign-on network in a hostile environment"
Setting up a linux-based LDAP/kerberos system that works with Linux,
Windows and MacOS X desktops from the point of view of a
sysadmin. A work in progress.
Outline:
. Authentication and authorisation
. LDAP
. Kerberos
. How kerberos ties in with LDAP
. How this compares with ADS (MS Active Directory Service)
. Some other security features of our network.
. Where we are up to.
About the speaker: Kim Holburn is a Systems and Network Administrator
at RSISE
Links : kim.holburn at anu.edu.au
Presenter : Jian
(Andrew) Zhang
Date : TBC
Location : Seminar Room, Ground Floor, RSISE
Title: short-range High-speed Ultra wideband Communications --whose
time has come
Abstract:
Ultra wideband (UWB) is an innovation in communications, and it is very
promising in short range high-speed communication
applications such as WPAN and consumer electronics.
In this talk, an overview of UWB systems will be given, including the
technical background, open research problems, Hardware development,
international industry and academic activities, and potential market.
Existing solutions and challenges will be highlighted. And crucial
research directions will be discussed.
As part of this talk, some recently develped novel schemes on the
synchronization and channel estimation in UWB systems will be
introduced. These schemes are improvements of the well-known shift
invariant techniques such as ESPRIT and the pencil methods. In these
improved schemes, principle multipath signals can be automatically
tracked and identified, while the complexity is only roughly associated
with the number of these principle signals.
Links :
jian at syseng.anu.edu.au
http://syseng.anu.edu.au/~jian
Presentation
(3.5MBbytes ppt file!)
Presenter : Prof
M. Amin, IEEE Distinguished Lecturer
Date : Wed. 10th March 2004, 11am
Location : Seminar Room, Ground Floor, RSISE
Title: Adaptive Blind Equalization for Indoor Dynamic Channels
Abstract:
Wireless communication is established using signals that travel in
multipath. The signals within a building in particular, cause certain
effects that can hamper communication. This can result in extreme
variation in signal conditions that affect performance dramatically.
A device known as an equalizer, processing the outputs of
wireless receivers, is used to minimize these effects and improve
performance. Applications are also considered for high-definition
TV, wireless laptop overhead projectors, and wireless indoor networking.
The talk presents a new adaptive channel equalization technique
to combat multi-path effects in an indoor environment using single and
multi-antenna receivers. In the presentation, a modified constant
modulus algorithm (MCMA) for adaptive channel equalization for QAM
signals is proposed. The proposed algorithm minimizes an error cost
function that includes both the signal amplitude and phase at the
equalizer output. In addition to the amplitude-dependent term, which is
provided by the conventional constant modulus algorithm (CMA), the cost
function includes a signal constellation matched error (CME) term. This
term can be designed using finite or infinite order polynomials and
should satisfy a set of desirable properties. The MCMA is compared with
the CMA and several existing algorithms that use amplitude and phase
information for blind equalizations. Both analysis and simulations
demonstrate that the proposed technique leads to faster convergence and
smaller misadjustment as compared to the conventional CMA. It is
also superior in performance or/and in computations to competitive
techniques.
Links :
Presenter : Sarah
Johnson, NICTA Sydney
Date : Thur. 4th March 2004, 9am
Location : Seminar Room, Ground Floor, RSISE
Title: Designing Error Correction Codes for Iterative Decoding
Abstract:
In this presentation low-density parity-check (LDPC) codes are
introduced and techniques for designing them discussed. The sum-product
decoding algorithm is presented before focusing on algebraic LDPC code
design. I will also talk about my recent work in constructing
non-binary and dual-field LDPC codes, and considering applications to
channels with memory.
Links :
Presentation (pdf)
Presenter
Presenter: Monika Trench, Summer Scholar
Date :Tue. 17th Feb. 2004, 2-2:30pm
Location : Seminar Room, Ground Floor, RSISE
Title: Image Processing
Abstract:
An image may be decomposed into information at different scales.
This is the basis for the multi-resolution encoding algorithm -
Laplacian Pyramid. This representation is originally introduced
for image progressive transmission. It also has applications
in image manipulation. Other relevant methods are also
considered. They include wavelet transform, median pyramid and "A
Trous" Algorithm.
Links :
presentation (pdf - 8.8
Mbytes!)
Presenter : Kandeepan Sithamparanathan, University of
Technology, Sydney
Date : 13th Feb. 2004, 2pm
Location : Seminar Room, Ground Floor, RSISE
Title: Signal Processing and Application Specific Synchronisation
Techniques for Digital Receivers : Applications on Mobile, Satellite
and Multicarrier - Multichannel Based Wireless Terminals
Abstract:
The evolution of and advances in high-speed wireless digital
communications have influenced the life of many people during the past
few decades. The applications vary from live video conferencing to
wireless voice and data communications. The high demand for such
applications motivates researchers and engineers to develop highly
sophisticated communication systems by utilising the available
resources at a relatively lower cost.
Here we present a few application specific synchronisation techniques
to be able to operate under rigorous conditions. We also take into
consideration the computational complexity associated with the
techniques. We look into application specific frequency, phase and
timing estimation techniques by modeling them statistically. The
concept of phase locked loop as a means of feedback frequency and
timing estimation is also
treated. We revisit the concept of the tan-lock loop, which we refer to
as the 4-Quadrant arctan-based phase-locked loop, whilst also exploring
many nonlinear loop models based on the 4-Quadrant arctan-based
phase-locked loop. We look into two specific applications, firstly, a
QPSK Earth Station Satellite receiver for LEO applications, and
secondly, an OFDM based WLAN receiver.
Links :
Presenter : Mr. Zhuo Chen
Date : 12th Jan. 2004, 2pm
Location : Seminar Room, Ground Floor, RSISE
Title: Performance analysis of transmit antenna selection
Abstract:
In this presentation, the error performance of a
multiple-input-multiple-output (MIMO) system with transmit antenna
selection (TAS) will be analyzed and presented. In particular, two
schemes will be investigated. The first scheme combines single TAS and
receiver maximal-ratio combining (MRC). And the second one combines TAS
with space-time codes designed for two transmit antennas. Based on the
exact bit or pairwise error probability expressions that we developed,
it is shown that a MIMO system with TAS can achieve a full diversity
order, as if all the transmit antennas were used, although at any time
only one or two transmit antennas are activated for transmission.
The impact of channel estimation error and antenna selection
error
on the system performance will also be discussed.
Links :
presentation (pdf)
Presenter : Dr. Jinho Choi
Date : 12th Jan. 2004, 2pm
Location : Seminar Room, Ground Floor, RSISE
Title: Beamforming methods for the downlink channel
Abstract:
In this talk, beamforming methods for the downlink channel are
discussed with their performance and implementation issues.
As a downlink beamforming method, the eigenbeamforming is extensively
addressed with its performance. We will show
that the eigenbeamforming is very promising, since it can provide the
diversity gain as well as the beamforming gain.
The application of the eigenbeamforming for OFDM systems is also
considered with its implementation issues.
Links :
presentation (pdf)
Presenter : Dr. Jinhong Yuan
Date : 12th Jan. 2004, 2pm
Location : Seminar Room, Ground Floor, RSISE
Title: Adaptive Transmit Antenna Selection with Pragmatic Space-Time
Trellis Codes
Abstract:
We consider the problem of selecting a subset of transmit antennas in
MIMO systems to minimize error probability
when only partial channel information is available at the transmitter.
An analytical error probability upper bound for space-time codes over
slow Rayleigh fading channels is derived. Based on the performance
analysis, two sets of criteria of selecting a subset of available
transmit antennas to minimize the space-time code error probability are
proposed. Furthermore, we present pragmatic space-time trellis
coding schemes for slow Rayleigh fading channels. The principle
advantages of the schemes is that a single encoder and decoder can be
used for systems with a variable number of transmit antennas.
Therefore, they are very suitable for adaptive transmit antenna
selection systems. The performance of the pragmatic space-time codes
with adaptive antenna selection is also evaluated by simulations. It is
shown the adaptive selection offers considerable antenna selection gain
relative to the system with a fixed selection.
Links :
presentation (pdf)
Presenter : Dr. Mark Reed
Authors: Mark Reed, Leif Hanlen
Date : 26th Nov. 2003, 2:30pm
Location : Seminar Room, Ground Floor, RSISE
Title: Return Link Code Acq. for 1-D and 2-D with DS-CDMA for High
Capacity Multiuser Systems
Abstract:
Acquisition of the code timing in a direct-sequence code-division
multiple-access system at the base station must take place before
signal detection and
decoding is possible. Code acquisition under severe multiple
access
interference conditions with time varying codes makes the task even
more
difficult. Inefficient designs lead to large number of false alarms
and/or
missed detections. This requirement is needed for conventional single
antenna
(one dimensional) designs and also for multi-element antenna (two
dimensional)
designs. This paper details a powerful code acquisition technique
for
the uplink of direct-sequence code-division multiple-access systems
under
high loaded situations for both 1-D and 2-D schemes, where the number
of
users is greater than the processing gain. Under this high
multiple
access interference condition the DS-CDMA acquisition problem becomes
very
difficult and conventional search methods simply fail. The method
discussed utilises soft data from a multiuser detector to reduce the
interference
received by the acquisition unit. Analytical performance is compared to
simulation results in terms of the number of users, processing gain,
cancellation
factor, antenna configuration, and noise variance.
Links :
presentation (pdf)
Presenter : Dr. Alex Grant
Authors: Alex Grant, Mohammad Rezaeian
Date : 4th Nov. 2003, 3pm
Location : Seminar Room, Ground Floor, RSISE
Title:Capacity Computation and Symmetry for Discrete Memoryless
Multiple-Access Channels
Abstract:
The Blahut-Arimoto algorithm is generalized for computation of the
total capacity of discrete memoryless multiple access channels. In
addition, a class of multiple access channels is defined with the
property that the uniform distribution achieves the total capacity.
These results are based on the specialization of the Kuhn-Tucker
condition for the total capacity of the multiple access channel, and an
extension of a known symmetry property for single user channels.
Links :
More information contact: alex.grant at unisa.edu.au
Presenter : Prof. Rod Kennedy
Authors: Thushara Abhayapala, Tony Pollock and Rod Kennedy
Date : 28th Oct. 2003, 2pm
Location : Seminar Room, Ground Floor, RSISE
Title: Characterization of 3D Spatial Wireless Channels
Abstract:
In this paper a novel three dimensional spatial channels model is
developed to provide insight into spatial aspects of multiple
antenna communication systems. The spherical harmonic representation of
wavefields is used to decompose the spatial channel matrix into a
product of known and random matrices where the known portion shows the
effects of the physical configuration of antenna elements. The model
supports any arbitrary antenna array configurations as well as any
distribution of scatterers. Possible applications of the model and its
usefulness are outlined.
Links :
presentation (pdf)
VTC Fall 03 paper (pdf)
Presenter : Prof. Rod Kennedy
Date : 24th Sept. 2003
Location : SAS Visions Theatre, National Museam of Australia, Canberra
Title: 'Back to the Future' with Wireless Communications
Abstract:
Some of us still do an about-take when we hear someone quietly mumbling
away to themselves in public places when what they are actually doing
is using a hands-free kit on a mobile phone. This phenomenon would have
been unthinkable only a few years ago when the world was introduced to
first
generation mobile phones — thick as a brick, and almost as large and
dense.
If we continue the trend of better, faster, cheaper and smaller,
we could assume that next generation mobile phones would be as powerful
as a desktop computer, and require no power because they run on fresh
air. They’d be cheap enough to give away with Happy Meals, and probably
small enough to present a serious choking hazard to the kids eating the
burgers.
Size matters. But the reality is that there probably is a limit to how
small (and powerful) mobile phones can become. Basic research into
wireless communications attempts to answer the question on what the
limits are, and how close current technologies are to those limits.
What will this mean for the future of telecommunications and new
products, new services, new functions and those annoying new ringtones?
This lecture will provide an overview of current research into wireless
communication systems at ANU, and highlight the ugly future of
ubiquitous
wireless communication services you don’t want or need, but which are
heading your way anyway!
Professor Rod Kennedy and his team at The Australian National
University have been probing the limits to the performance of wireless
communications by going back to the future, and viewing communication
as information
conveyed on radio waves.
Links :
presentation (pdf)
Presenter : Glenn Dickins
Date : 17th Sept. 2003
Location :
Title: Commercialising Australian Technology
Abstract:
Glenn will talk about his experience in commercialising Australian
technology, measures of success, harsh realities, obstacles,
suggestions and a comparison of the Australian and US approach to
start-ups.
Links :
presentation
(pdf)
Presenter : Dr
Leif Hanlen
Date : Tue 29th July
Location : Seminar Room, Ground Floor, RSISE
Title: Capacity of correlated finite-dimensions MIMO channels
Abstract:
The capacity of correlated finite-dimensions MIMO channels, where the
channel gains have a generalized Wishart distribution is found.
Asymptotic expressions are given where one dimension is much larger
than the other. For many transmitters, the asymptotic capacity
can be divided into two components: one arising from the dominant
eigenvalues of the correlation matrix,
and the other from the remaining eigenvalues.
MIMO models which use random matrices and arbitrary fixed-rank
correlation matrices still have the (unrealistic) infinite linear
growth
for which MIMO is famous.
Links :
presentation
(pdf)
home_page
Presenter : Dr
Mark Reed
Date : Friday 2nd May
Location : Seminar Room, Ground Floor, RSISE
Title: Iterative Multiuser Receivers: Bits to System Design
Abstract:
This presentation discusses high performance multiuser detectors for
DS/CDMA systems, specifically for 3G/UMTS systems. The problem is
introduced before the basic channel model and typical receiver
design is shown. We consider implementation aspects and present
a modular design which could be realised in hardware.
We show BER performance of a 3GPP compliant system both with single and
multiple antennas at the receiver. Network level considerations
are discussed to determine the true benefit of these schemes to Network
Operators and Manufacturers. We conclude by discussing some of
the open issues from a practical and theoretical point of view.
Links :
presentation (pdf)
home_page
Presenter: Jian
(Andrew) Zhang
Date: Monday 28th April
Location : Seminar Room
Title: Performance of RAKE Reception for Ultra Wideband Signals in a
Lognormal-fading Channel
Abstract:
In this talk, some new problems emerging in the design
of RAKE receiver for TH-UWB will be highlighted and
the performance of RAKE will be analyzed focusing on
the following four questions:
1) When can we get CLEAN RAKE output with negligible ICI and ISI?
2) How to derive explicit expressions of RAKE performance
in a lognormal fading channel in the case of CLEAN output?
3) How to analyze the effect of ICI in a practical UWB system?
4) What's the speciality of the optimization of signal's parameters
in consideration of the constrained energy of UWB systems?
Liinks:
Presenter : Mr Kim Holburn
Date : April
Location : Seminar Room, Ground Floor, RSISE
Title : Introduction to Networking
Links :
presentation (pdf)
Presenter : Dr Dhammika Jayalath
Date : Thursday 17th April
Location : Seminar Room, Ground Floor, RSISE
Abstract :
Orthogonal frequency division multiplexing (OFDM) is successfully
used in many wireless digital communication systems. OFDM is also
a potential candidate for future broadband wireless communication
systems. A cyclic prefix (guard interval) is used in OFDM systems to
avoid the intersymbol interference. In this talk I will present a novel
cyclic-prefix based delay-spread estimation technique for OFDM
systems.
This technique uses the change of gradient of a correlation function
as the strategy to detect the delayed arrival paths. Estimation of the
symbol timing and frequency synchronization
information is also inherent in the technique. This technique can be
efficiently implemented using the Viterbi search algorithm.
The delay-spread information provided by this technique can be
adaptively used to improve the accuracy of the channel estimation
process in OFDM systems.
Links :
Presenter : Eric Lehmann
Date : 2pm Wednesday 16th April,
Location : Seminar Room, RSISE
Title: Sequential Estimation Methods for Acoustic Source
Localization
Abstract:
Traditional methods used in acoustic source localization
algorithms typically attempt to find the source location using data
collected from an array of sensors at the current time only. In the
presence of strong multipath, these traditional algorithms often
erroneously locate a multipath reflection rather than the true source
location. A recently proposed approach that appears promising in
overcoming this drawback of traditional algorithms, is to use a
state-space approach based
on particle filtering.
This seminar will present an overview of some methods used
traditionally for acoustic source localization, and will briefly
review the basic principles of Bayesian filtering and sequential
estimation. The methodology used to integrate the traditional source
localization techniques into a sequential Monte Carlo framework will
also be described. Finally, the tracking performance of four specific
particle filters
will be demonstrated by presenting experimental results obtained with
both simulated reverberant data and data recorded in a moderately
reverberant office room. These results indicate that a substantial
tracking performance improvement can be achieved when using a particle
filtering approach
for the problem of acoustic source localization.
Links :
Presentation
Home_Page
Author: Mark C. Reed
Last Update: Thu Apr 16 2003