EECS 500 Homepage -- Winter 2003
Special Lectures in
COMMUNICATIONS AND NETWORKS

Lecture schedule as of 1/7/03

• Tuesday, January 7
  
  FIRST CLASS MEETING
  Intro to EECS 500
  
  

  ---

• Thursday, February 6
  
  Multi-Antenna Communications: From Information Theory to Algorithms
  Professor Babak Hassibi
  Department of Electrical Engineering
  California Institute of Technology
  
  Abstract -- The use of multiple antennas can, in theory, significantly boost channel capacity as well as lower the probability of error of a wireless communications link. The main signal processing challenge in multi-antenna communications is to devise practical space-time transmission schemes that are simple, yet efficient: simple so that all the processing can be done in real-time, and efficient so that the high rates promised by theory can be attained. In this talk I will describe several recent advances in the design of space-time codes for both quasi-static (as in fixed wireless) and rapidly fading (as in mobile wireless) channels, as well as some new work in the development of efficient (polynomial-time) maximum-likelihood decoding algorithms.

  ---

• Thursday, February 13
  
  On the Algebraic Design of Space-Time Codes
  Dr. A. Roger Hammons
  Chief Scientist
  Johns Hopkins University Applied Physics Laboratory
  
  Abstract -- Recently, it has become known that the outage capacity of the Rayleigh fading channel increases when multiple transmit and receive antennas are used. This new understanding of the wireless communication channel has resulted in vigorous research activity into signal design and signal processing techniques to realize the potential benefits of the multiple antenna system. Initial space-time code designs were either crafted by hand or found through exhaustive search based on ad hoc design rules. The space-time code designs of Hammons and El Gamal were the first to propose general algebraic design rules. This talk will provide an overview of the algebraic design of space-time codes for BPSK and QPSK modulation based on this seminal work. It will also survey extensions to other modulation schemes and other channels.
  
  Biosketch -- A. Roger Hammons Jr. received the B.S. degree with highest honors and distinction in mathematics and the humanities from Harvey Mudd College, where he was an S&H Foundation Scholar; the M.A. degree in mathematics from the University of California, Los Angeles, where he was a Teaching Assistant; and the Ph.D. degree in electrical engineering from the University of Southern California, where he was a Hughes Aircraft Company Ph.D. fellow. His doctoral work investigated the connections between quaternary sequence designs as CDMA signature sequences and as error control codes over the ring of integers modulo 4. This work resolved the longstanding problem regarding the algebraic relationship between the nonlinear Kerdock and Preparata codes, established the study of Z4-linear codes, and led to the discovery of best known sequence designs of large size.
  
  Dr. Hammons was with Hughes Electronics from 1981 through 2000, first with the Missile Systems Group of Hughes Aircraft Company and then with Hughes Network Systems. At Hughes Networks, he and his research group led the introduction of turbo codes into the emerging standards for third-generation CDMA cellular radio systems, where their designs have been adopted worldwide. In November 2000, he joined Corvis Corporation.
  Dr. Hammons was recipient of the IEEE Information Theory Society 1995 Best Paper Award for his doctoral research regarding the Kerdock and Preparata codes. At Hughes, he received numerous awards including the 1999 Hughes Patent Award and the 1999 Chairman¹s Honors Award for technical innovation. He holds 14 U.S. patents.

  ---

• Thursday, March 13
  
  Bit Stuffing and CONSTRAINED Sequences
  Professor Jack Keil Wolf
  Department of Electrical and Computer Engineering
  University of California, San Diego
  
  Abstract -- Bit-Stuffing has been used in various communication and storage applications to prevent Control Signals from appearing in the Information sequence. In this talk, we will consider how the rate loss for bit stuffing compares to the smallest possible rate loss predicted by Shannon theory. Surprisingly, we will show that with an appropriate modification, bit stuffing can be 100% efficient. Applictions to two dimensional arrays will also be considered.
  
  Biosketch -- Jack Keil Wolf is a Professor of ECE at UCSD. He was educated at the University of Pennsylvania and Princeton University. His research is in coding and information theory.
  
  

  ---

• Thursday, March 20
  
  Systematic Code Construction of Space Time Codes
  (Joint Work with Hamid Jafarkhani, Univ of California, Irvine)
  Dr. Nambi Seshadri
  Broadcom Corp.
  
  Absract -- Space-Time codes exploit the extra dimension of space, in addition to time and frequency, to create powerful channel codes for communicating over fading wireless channels. An area of weakness in space-time coding research has been the lack of elegant approaches to the design of such codes, not withstanding the "design rules" in the work of Tarokh, Seshadri and Calderbank. I will talk about a systematic approach to the construction of space-time codes by using "set-partitioning" principles. The "set-partition" is accomplished by considering a "super-orthogonal" set of space-time matrices. The design rules lead to simple and powerful code constructions for N transmit, 1 Receive communications system. I will further talk about some significant open issues that remain in the area of space-time coding.
  
  Biosketch -- Nambi Seshadri is senior director of wireless technology strategy at Broadcom Corporation, Irvine, CA. In this role he is responsible for corporate technology and business strategy in the area of wireless. His current research interest is in the area of space-time codes and coding for MIMO wireless channels. He received the 1999 IEEE Information Theory society paper award and is a Fellow of the IEEE.
  
  

  ---

• Thursday, March 27 -- SEMINAR CANCELLED --
  
  Fundamental Limitations of MIMO Wireless Links that use Polarimetric Antenna Arrays
  Dr. Thomas Marzetta
  Bell Laboratories, Lucent Technologies
  
  Abstract -- It was recently speculated that six-component polarimetric antennas - comprising six collocated electric and magnetic dipoles that generate the six components of electric and magnetic current or that measure the six components of the electric and magnetic field -could give six-fold improvements in capacity over single-component antennas in Rayleigh flat fading. Using an exact wave propagation formulation involving the plane-wave spectral decompositions of the transmitted and received fields, we show that the behavior of wireless links based on ARRAYS of polarimetric antennas is qualitatively different than that of links based on SINGLE polarimetric antennas. The MIMO channel can be viewed rigorously in the wavenumber domain. For dense planar arrays, MANY WAVELENGTHS IN EXTENT, the use of polarimetric antennas yields at most a four-fold improvement in capacity, which can be realized by particular combinations of four of the six dipoles. Six-component elements, the expansion of the planar arrays into volume arrays, or spacing elements at intervals less than a half-wavelength only give additional logarithmic improvements in capacity. Wave propagation physics imposes very powerful constraints and one cannot make arbitrary assumptions about the independence of the entries of the propagation matrix, even under the most optimistic scattering conditions.
  
  Biosketch -- Thomas L. Marzetta was born in Washington, D.C. He received the Ph.D. in electrical engineering from the Massachusetts Institute of Technology in 1978. His dissertation extended the three-way equivalence of autocorrelation sequences, minimum-phase prediction error filters, and reflection coefficient sequences to the two-dimensional case. He worked for Schlumberger-Doll Research from 1978 to 1987 to modernize geophysical signal processing for petroleum exploration. From 1987 to 1995 he performed research and development at Nichols Research Corporation under contracts from the U.S. Department of Defense, NASA, and Schlumberger; he headed a group that improved automatic target recognition, radar signal processing, and video motion detection. Since 1995 he has been with Bell Laboratories (formerly AT&T, now Lucent Technologies), currently in the Mathematical Sciences Research Center where he is the Head of the Mathematics of Communications Research Department. He specializes in multiple antenna wireless with particular emphasis on techniques for realizing extremely high throughputs with large numbers of antennas.
  
  Dr. Marzetta is active in the IEEE: at present he is a member of the Sensor Array and Multichannel Technical Committee of the Signal Processing Society, and a guest editor for the Special issue of the IEEE Transactions on Information Theory on Space-Time Transmission, Reception, Coding and Signal Design.

  ---

• Thursday, April 3
  
  NOTE: STARTING TIME WILL BE 3:00 PM
  (this seminar only)
  
  Electrical and Magnetic Imaging of Human Brain Function
  Prof. Richard Leahy
  Signal and Imaging Processing Institute
  University of Southern California
  
  Abstract -- Measurements of external magnetic fields and scalp potentials produced by neuronal activity in response to sensory, motor and cognitive tasks provide unique insights into functional organization and communication networks in the human cerebral cortex. These fields are referred to respectively as the magneto- and electroencephalogram, or MEG and EEG.. While other imaging modalities provide high resolution localization of brain activation, most notably, functional MRI, these modalities are limited in temporal resolution by the several second time-constants of the hemodynamic effects on which they are based. Conversely, MEG and EEG provide high temporal resolution, on the order of milliseconds, but present difficult challenges in the production of meaningful spatial maps of brain activation. It is this that makes MEG and EEG data analysis such an interesting problem from a signal processing perspective. I will briefly describe the forward problem and its solution using closed form or numerical boundary and finite element methods. Inverse methods typically fit either a low dimensional model using dipole or quadrupolar models of focal activity, or high dimensional images constrained to a tessellated representation of the cortical surface. I will describe methods used for estimating both high and low dimensional sources, including regularized and Bayesian approaches to cortical imaging, and the use of signal subspace methods for parametric source estimation. I will also discuss the relative merits and limitations of these methods. I will conclude by discussing some of our recent work on evaluating the stability and significance of estimated activation maps using permutation tests and bootstrapping methods.
  
  Biosketch -- Richard M. Leahy received the B.Sc. and Ph.D. degrees in Electrical Engineering from the University of Newcastle upon Tyne, England, in 1981 and 1985 respectively. In 1985 he joined the University of Southern California where he is currently a Professor in the Department of Electrical Engineering and Director of the Signal and Image Processing Institute. He holds joint appointments with the Departments of Radiology and Biomedical Engineering at USC. He was general co-chair of the International Conference on Information Processing in Medical Imaging in 2001 and will serve as General co-chair of the IEEE International Symposium in Biomedical Imaging in Washington, D.C. in April 2004. He is also an associate editor of the IEEE Transactions on Medical Imaging. Dr. Leahy's research interests lie in the application of signal and image processing theory to biomedical inverse problems. His current research involves the development of methods for anatomical and functional imaging with applications in neuroimaging, oncology, and gene expression using magnetoencephalography, positron emission tomography, fluorescence imaging, and magnetic resonance imaging.

  ---

• Tuesday, April 15
  
  Aspects of Secure Multiparty Computation
  Prof. Ian Blake
  Dept. of Electrical and Computer Engineering
  University of Toronto
  
  Abstract -- The notion of secure multi-party computation involves several parties wishing to compute some function of their inputs without revealing any information about their inputs beyond what is implied by the value of the function. A good example of this, is the Yao millionaire problem where two parties wish to determine who is the wealthier without revealing their net worth. The area has developed and certain interesting primitives such as bit commitment, zero knowledge proofs, verifiable secret sharing and oblivious transfer, have been widely studied. A formal framework for the area has been established but this talk informally develops a few of the ideas central to the area by considering some examples.
  
  Reference: The central reference for the topic is a book being developed by Oded Goldreich: Secure multiparty computation (http://www.wisdom.weizmann.ac.il/~oded/pp.html)
  
  Biosketch -- Ian F. Blake received his undergraduate education at Queen's University in Kingston, Ontario and his Ph.D. at Princeton University in New Jersey. From 1967 to 1969 he was a Research Associate with the Jet Propulsion Laboratories in Pasadena, California. From 1969 to 1996 he was with the Department of Electrical and Computer Engineering at the University of Waterloo, in Waterloo, Ontario where he was Chairman from 1978 to 1984 and Director of the Institute of Computer Research from 1990 to 1994. He is currently with the Department of Electrical and Computer Engineering at the University of Toronto where he is Director of the Bell University Labs program
  

  ---

• SUPPLEMENTAL LECTURE
  Friday, September 12
  Capacity-approaching codes
  G. David Forney, Jr. (LIDS-MIT)
  Details to follow

  ---

  Sponsors:

  Lucent Technologies and the University of Michigan Dept. of Electrical and Computer Engineering and College of Engineering.