15th European Signal Processing Conference EUSIPCO 2007

EURASIP Fellow Inaugural Lectures

The EURASIP Administrative Committee (AdCom) has recently decided to initiate a 'Fellowship Programme', to recognize outstanding achievements of its members and volunteers. Each year, a select group of signal processing researchers will be elevated to 'EURASIP Fellow', the Association's now most prestigious honor.

It is my pleasure to announce the result of the 2007 'EURASIP Fellows' selection process. In recognition of their many important contributions to the field of signal processing, the EURASIP AdCom elevates the following individuals to 'EURASIP Fellow 2007':

     Prof. Peter Grant, University of Edinburgh, Edinburgh, UK

     Prof. Wolfgang Mecklenbräuker, Vienna University of Technology, Vienna, Austria

     Prof. Peter Stoica, Uppsala University, Uppsala Sweden

     Prof. Martin Vetterli, Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland

I wish to congratulate our new EURASIP Fellows on this most important achievement. The Fellowship Awards will be presented at the conference. In addition, to highlight the new Fellowship Programme, Prof. Peter Grant and Prof. Martin Vetterli will present a 'Fellow Inaugural Lecture'.

Marc Moonen
President EURASIP

Review of real world MIMO system performance

Wednesday, September 5th, 9:00 - Sala Kongresowa (Congress Hall)

Peter Grant, University of Edinburgh, United Kingdom

This review lecture will introduce the basic techniques which are adopted in multiple input multiple output (MIMO) systems and show the potential theoretical increase in data throughput or transmission efficiency in bit/s/Hz. This is dependent on a rich multipath environment supporting many unique propagation modes. Modelling and actual results will be presented from the universities of Bristol and Illmeneau on typical channel characteristics and provide practically realisable channel capacity in urban LOS and non-LOS channels.

The presentation will also include real time hardware implementations of the

receiver architectures from TU Vienna which use FPGA solutions. These provide particularly rapid design time. It will also describe simplified receiver designs using sphere decoders which achieve close to maximum likelihood performance combined with rapid FPGA implementation plus a high data rate capability.

[Peter Grant - photo]

Peter Grant, born in St. Andrews, received the B.Sc. degree in electronic engineering from the Heriot-Watt University, in 1966, the Ph.D. degree from the University of Edinburgh, in 1975, and an honorary DEng from the Heriot-Watt in 2006.

He worked intially in radiocommunications for the Plessey Company, before he was appointed to a research fellowship at the University of Edinburgh. He was subsequently promoted to a Professor of Electronic Signal Processing in 1987 and in 2002 appointed head of the School of Engineering and Electronics.

During academic year 1977-78, he was a visiting professor at the Ginzton Laboratory, Stanford University, and in 1985-86 he was a visiting staff member at the MIT Lincoln Laboratory. He was awarded the 82nd (2004) Faraday Medal by the Institution of Electrical Engineers (IEE) for his work on CDMA receiver designs and adaptive filters. In 1974 and again in 1977 he was awarded the Bulgin premium from the then Institution of Electronic and Radio Engineers and in 1982 their Lord Mountbatten premium. In 1994 he was awarded the IEE Marconi and Langham Thompson premia.

Professor Grant was president of EURASIP, the European Association for Signal Processing, in 2000-2002, chairman of EUSIPCO-94 and technical programme chairman for ICASSP-89 international conferences. In 1998 he was appointed by the the US IEEE Signal Processing Society as a distinguished lecturer on DSP for Mobile Communications, presenting at 25 locations over five continents.

He served from 1980-1996 as an honorary editor of IEE Proceedings title "Vision Image and Signal Processing". He was chair of the 2001 Universities Funding Council research assessment panel for the UK Electrical Engineering Departments and has served as research assessor at: Queensland University of Technology, University of West Australia, City University Hong Kong and ETHz in Zurich. He is a member of the Scottish Science Advisory Committee.

He holds fellowships of the IEEE, IEE, Royal Academy of Engineering and the Royal Society of Edinburgh.

Sampling Sparse Signals at Occam's Rate

Thursday, September 6th, 9:00 - Sala Kongresowa (Congress Hall)

Martin Vetterli, Ecole Polytechnique Fédérale de Lausanne, Switzerland and University of California, Berkeley, USA

joint work with T. Blu, L. Coulot, A. Hormati and P.L.Dragotti

Shannon's sampling theorem gives a sufficient condition for reconstructing the projection of a signal onto the subspace of bandlimited functions, and this by taking inner products with a sinc function and its shifts. Recently, this framework has been extended to classes of non-bandlimited signals and their perfect reconstruction from a suitable projection. This gives a sharp result on the sampling and reconstruction of sparse continuous- time signals, namely that 2K measurements are necessary and sufficient to perfectly reconstruct a K-sparse continuous-time signal.

We first review this result and show that it relies on structured Vandermonde measurement matrices, of which the Fourier matrix is a particular case. Because of this structure, fast, O(K^3) methods exist. When then generalize these results to a number of cases where sparsity is present, including piecewise polynomials as well as local measurement kernels like splines.

Of course, real cases always involve noise, and thus, retrieval of sparse signals in noise is considered. Lower bounds by Cramer-Rao are given, and an iterative algorithm due to Cadzow is shown to perform close to optimal over a wide range of signal to noise ratios. This indicates the robustness of such methods.

Next, we consider the connection to compressive sampling, a recent approach involving random measurement matrices, a discrete set up, and retrieval based on convex optimization. We compared the two approaches, highlighting differences, similarities, and respective advantages.

[Martin Vetterli]

Martin Vetterli received his Engineering degree from ETH in Zurich, his MS from Stanford and his Ph.D. from EPFL in Lausanne.
In 1986, he joined Columbia University in New York, first with the Center for Telecommunications Research and then with the Department of Electrical Engineering where he was an Associate Professor of Electrical Engineering. In 1993, he joined the University of California at Berkeley, were he was Full Professor until 1997. Since 1995, he has been a Professor at EPFL, where he headed the Communication Systems Division (1996/1997) and heads the Audiovisual Communications Laboratory. From 2001 to 2004 he directed the National Competence Center in Research on mobile information and communication systems. He has also been Vice-President for International Affairs at EPFL since October 2004. His research interests are in the areas of applied mathematics, signal processing, and communications.