报告题目：Acoustic Beamforming and Localization on Wireless Sensor Network Systems
报 告 人：美国加州大学洛杉矶分校杰出终身教授、IEEE终生会士—Kung Yao教授
邀 请 人：何明一教授、陈景东教授
In recent years, there have been many publications on the theoretical and practical aspectsof sensor network systems. In this seminar, we will first review the diverse activities of the NSF funded UCLA Center for Embedded Networking Systems (CENS) for over ten years by faculty members from UCLA, UCR, USC, and Cal Tech. Then we will introduce our own research works in acoustic beamforming and localization using different sensor network systems. Our earliest beamforming algorithm was based on a maximum-energy criterion for a front-end hearing-aid array and also for adaptive acoustic/seismic source localization. Then we introduce the theory and practice of the wideband Approximate Maximum-Likelihood (AML) algorithm for beamforming and localization applications funded by DARPA, MURI, and NSF agencies. In particular, our NSF-based efforts are in collaboration with UCLA researchers who are interested in the vocalization, classification, and linguistic properties of different species of birds. Our AML-based array utilizing the Alternative Projection Algorithm can effectively steer adaptively two beams, one beam having a high-gain toward a desired source A and nulling toward a unwanted source B, while another beam has a high-gain toward a desired source B and nulling toward a unwanted source A. Our initial AML algorithm on an array was capable of dealing with beamforming and localization of the azimuth of a source (in a 2D setting). With two or more such arrays, the crossings of the estimated azimuth beams then yield the location of a source. Later, we generalized the AML-based array to deal with beamforming and localization of the azimuth and elevation of a source (in a 3D setting). Recently, we also tackled the fast AML-based algorithm for the 3D problem. Wireless hardware platforms will also be presented. Finally, we will briefly present the use of the random set theory (RST) methodology for source number estimation and localization when the number of sources may be unknown and/or time varying.
Kung Yao received the B.S.E. (Summa Cum Laude), M.A. and Ph.D. degrees in EE all from Princeton University, Princeton, N.J. He was a NAS-NRC Post-Doctoral Research Fellow at the University of California, Berkeley. Presently, he is a Distinguished Professor in the EE Department and the Co-Director of the Public Safety Network Systems Laboratory at UCLA. In 1985-1988, he served as an Assistant Dean of the School of Engineering and Applied Science at UCLA. In 2007, he was a Royal Society Kan Tong Po Visiting Professor at HK Polytechnic University. His research and professional interests include digital and array signal processing in sensor networks, wireless communication theory and system, statistical modelling of wireless fading channels, and systolic and VLSI algorithms, architectures, and systems. He was the co-editor of a two volume series of an IEEE Reprint Book on "High Performance VLSI Signal Processing," 1997. Dr. Yao received the IEEE Signal Processing Society's 1993 Senior Award in VLSI Signal Processing and the 2008 IEEE Communications Society/Information Theory Society Joint Paper Award. He is a Life Fellow of IEEE.