Select a calendar:
Filter May Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Events for May 04, 2005
-
The Impact of Feedback on Communication Systems: Capacity and Complexity
Wed, May 04, 2005 @ 11:15 AM - 12:15 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
SPEAKER: Prof. Youjian (Eugene) LiuABSTRACT: Channel state feedback from receivers to transmitters is not difficult to implement in modern communication systems because most of them are full duplex systems. The third Generation (3G) cellular phone systems such as CDMA 2000 and WCDMA are typical examples of such systems where a few bits of channel state information can be sent to the transmitter every few milliseconds.Perfect or imperfect channel state information at transmitter (CSIT) has a profound impact on communication system design. We examine two aspects, capacity and complexity, and focus on capacity due to time constraint of the talk. For capacity, the channel state information can be exploited to allocate transmission power and adjust "transmission direction", also called eigen-channels, in case of Multiple-Input Multiple-Output (MIMO) systems. Shannon has proved that a memoryless device, which selects a portion of a codeword to be transmitted according to feedback, can optimally utilize the given memoryless feedback in a memoryless channel. However, if one has the freedom to design channel state feedback, it not clear whether the channel state feedback should be memoryless. If it is not memoryless, then the memoryless device at the transmitter is not optimal either. We prove that a memoryless feedback generator is indeed optimal, give the capacity formula, and therefore, obtain a coding theorem for memoryless channel with designable feedback. In addition, the proof shed light on how the feedback error may affect the performance. For application in MIMO systems in independent fading channel, we solve the open problem of performance analysis when the systems cannot be made equivalent to a Single-Input Single-Output systems. Using random matrix theory and extreme order statistics, we accurately characterize the performance gain as a function of the feedback rate (bits/channel use) for a power on/off strategy and show this strategy is near optimal. A new method is given to translate the effect of finite-feedback-rate constraint to a power loss factor compared with the case of infinite feedback rate. The analytical results match with simulation results almost perfectly. An application of the research would be the standard evolution of 3G cellular systems.For complexity, even in situations that Channel State Information at Transmitter (CSIT) does not improve capacity, we demonstrate that it can be exploited to allocate computational resource in code design, resulting in reduced encoding or decoding complexity. Consequently, computational resource allocation according to CSIT is a design problem that is as fundamental as power water filling. Advances in the past decade have resulted in error control codes with performance very close to capacity. It is expected that further advance in the field would be centered on lowering the coding complexity to support increasingly prevailing low power and portable applications.BIO: Youjian (Eugene) Liu received the Ph.D. and M.S. degree in Electrical Engineering from The Ohio State University in 2001 and 1998 respectively, the M.S. degree in Electronics from Beijing University, China, in 1996, and the B.E. degree in Electrical Engineering from Beijing University of Aeronautics and Astronautics, China, in 1993. Since August 2002, he has been an Assistant Professor of Department of Electrical and Computer Engineering, University of Colorado at Boulder, Colorado. From January 2001 to August 2002, he worked on space-time communications for 3G mobile communication systems as a Member of the Technical Staff in CDMA System Analysis and Algorithms Group, Wireless Advanced Technology Laboratory, Lucent Technologies, Bell Labs Innovations, Whippany, New Jersey. His current research interests include communications, coding theory, and information theory.Host: Professor Keith M. Chugg, chugg@usc.edu
Location: Frank R. Seaver Science Center (SSC) - -319
Audiences: Everyone Is Invited
Contact: Mayumi Thrasher