User Tools

Site Tools

Writing /home/users/ashutosn/public_html/CommNetS2016/dokuwiki/data/cache/0/0b434ec221fdfebc6f6420f5c32eace8.i failed
Unable to save cache file. Hint: disk full; file permissions; safe_mode setting.
Writing /home/users/ashutosn/public_html/CommNetS2016/dokuwiki/data/cache/0/0b434ec221fdfebc6f6420f5c32eace8.metadata failed
Writing /home/users/ashutosn/public_html/CommNetS2016/dokuwiki/data/cache/0/0b434ec221fdfebc6f6420f5c32eace8.i failed
Unable to save cache file. Hint: disk full; file permissions; safe_mode setting.
Writing /home/users/ashutosn/public_html/CommNetS2016/dokuwiki/data/cache/0/0b434ec221fdfebc6f6420f5c32eace8.xhtml failed

Abstract: Control is an intellectual sibling to communication. Both are about removing uncertainty with limited resources — communication by sharing something about the world and control by shaping the world itself. While information theory has for decades been providing insights into problems of communication, traditional approaches to control did not use information-theoretic techniques or ideas. Recently, we have found some surprising connections between wireless information theory and some central problems in decentralized control. In addition, we have begun to understand how modern insights can be used to better make wireless protocols that support control problems for the Internet of Things.

On the theoretical side, it turns out that the machinery of linear deterministic models that has been so helpful in understanding problems of relaying and interference in communication can be brought to shed light on the fundamental limits of performance in control. Approximately-optimal strategies can be found and the control-theoretic counterparts to ideas like generalized degrees-of-freedom and cut-set bounds can be discovered. There are control/estimation counterparts to ideas like non-coherent communication channels.

All this suggests that there is an entire parallel realm of information theory that connects to control problems — just waiting to be explored. This talk will give some glimpses into this.

If I have the time, I might talk a bit on the practical side, where we can connect to ideas of diversity and cooperative communication. High-performance for the Interactive Internet of Things (I^2oT) involves a different corner of the performance space than traditional communication. Latency and Reliability take center-stage as opposed to pure spectral efficiency, and performance has to scale well as the number of nodes increases.

Biography. Anant Sahai received his B.S. in 1994, from the University of California, Berkeley, and his S.M. and Ph.D. from MIT in 1996 and 2001, respectively. He is an associate professor in the EECS Department at Berkeley, where he joined as an assistant professor in 2002. Prior to that, he spent a year as a research scientist at the wireless startup Enuvis in South San Francisco, developing software-radio signal-processing algorithms to enable very sensitive GPS receivers for indoor operation. From 2007 to 2009, he was the treasurer for the IEEE Information Theory Society. His current research interests are at the intersection of information theory and decentralized control, as well as in wireless communication, particularly dynamic spectrum sharing and its regulatory dimensions. He enjoys working very closely with his small group of graduate students on fun and deep problems. He usually teaches small intimate courses but this semester, is teaching a giant intro course with hundreds of students.

information_theory_meets_control.txt · Last modified: 2016/09/01 19:15 (external edit)