Transcript 投影片 1

Distinguished Talk
On January 26, 2011
National Taiwan University of
Science and Technology
Prof. Wai Ho Mow,
Senior Member of
IEEE
Dept. of ECE,
HKUST
Professor Wai Ho MOW received his BSc (Electronics) in 1989, MPhil and PhD
(Information Engineering) in 1991 and 1993 respectively, all from the Chinese University
of Hong Kong (CUHK). In 1994, he was a visiting assistant professor at the Department of
Information Engineering, CUHK. He was a visiting scholar at the University of Waterloo,
the Munich University of Technology (TUM), the Kyoto University, and the University of
California at San Diego in 1995, 1996, 2000, and 2010, respectively. From 1997 to 1999,
he was an assistant professor at the Nanyang Technological University, Singapore. He
joined the Department of Electrical and Electronic Engineering, Hong Kong University of
Science and Technology, in March 2000. He has been an Adjunct Professor of the
Southwest Jiaotong University, Chengdu, China from 2003 to 2008. In spite of the
relatively short PhD study period, he received the Best PhD Thesis in Engineering Award
and the Young Scholar Dissertation Award. He was also the recipient of the Croucher
Research Fellowship (HK), the Humboldt Research Fellowship (Germany), the
Telecommunications Advancement Research Fellowship (Japan), the Tan Chin Tuan
Academic Exchange Fellowship (Singapore), the Wong Kuan Cheng Education
Foundation Academic Exchange Award (China), the Foreign Expert Bureau Fellowship
(China) and the Royal Academy of Engineering Award for Short Research Exchanges
with China and India (UK). His research includes: wireless communications, coding and
information theory. Dr. Mow was a past chair of the Hong Kong Chapter of the IEEE
Information Theory Society and has been a Senior Member of IEEE since 1999.
Speech: 11:00~12:00 (IB-201)
Robust Decoding for Unknown Impulsive
Noise Channels
In many real-world communication and storage systems, the extent of non-Gaussian impulsive
noise (IN) rather than Gaussian noise poses practical limits on the achievable system
performance. The decoding of IN-corrupted signals is complicated by the fact that accurate IN
statistics are typically unavailable at the receiver. Without exploiting the probability distribution
of the impulsive noise, the conventional method is to mark the IN corrupted symbol as erasures
before performing an error-and-erasure decoding. The main contribution of this work is to
propose a joint erasure marking and decoding approach to the design of a high performance
decoding algorithm for signals corrupted by impulsive noise as well as Gaussian noise. In this
presentation, a novel decoding metric for a joint erasure marker and decoder (JED) is
introduced. Some simple modulation schemes are investigated in detail to demonstrate the
sophistication in characterizing the decision regions and analyzing the performance of JED. Next,
the optimal joint erasure marking and Viterbi decoder (JEVA) is derived for a convolutionally
coded scheme. Our simulation results showed that JEVA performs only marginally worse than
the maximum likelihood decoder although unlike the latter, it does not exploit the knowledge of
the impulsive noise statistics at all. Finally, a sub-optimal variant of JEVA is devised to allow
different complexity-performance tradeoff.
Contact information : [email protected]
02-27376513 Prof. Der-Feng Tseng