Control Information

Anant Sahai
Associate Professor, UC Berkeley
Given on: April 3rd, 2015


Control provides a conceptually interesting area for understanding the nature of information beyond traditional communication problems. After all, both control and communication are about the reduction of uncertainty ---- in communication it is about informing the beholder so that its idea of the world is closer to reality while in control, it is about changing the world so that it more closely conforms to the idea of a beholder. Building by analogy with communication's standard picture of a source, channel, and destination/sink, this talk will discuss the source-nature of control systems, the channel-nature of control systems, and also the sink-nature of control systems. This sink-nature in particular, is something that we don't normally think about in communication systems. For control, however, it is natural to consider it and I will talk about a concept that we have developed recently that we call "control capacity."

Joint work with Gireeja Ranade and Se Yong Park.


Anant Sahai got his BS degree in EECS from Berkeley in '94 and then got his MS and PhD degrees from MIT in '96 and '01. Before joining Berkeley as a faculty member in '02, he spent a year at the wireless startup Enuvis working on ultrasensitive software-radio algorithms for GPS. His current research interests are in the foundations of information theory for control, low-latency wireless communication protocols to support the high performance Active IoT applications of the future, and wireless spectrum sharing (where his interests span system architectures, game theory, law, and policy).