Place: Duthie Center, Room 117
Date: Friday, February 12 , 2010
Reception: 3:00 – 3:30pm, Duthie Center 2nd Floor
Seminar 3:30

The Ambient Virtual Assistant Testbed and Cocktail Party Listening

Speaker: Kevin D. Donohue
Databeam Professor of Electrical and Computer Engineering
Center for Visualization
University of Kentucky

The "cocktail party effect" is the ability humans have to focus on a single conversation in a noisy and multi-talker environment, which is typically done at a "cocktail party." This term was coined by E.C. Cherry in 1953, and research continues today to understand the mechanisms by which human isolate conversations of single speakers. In recent years computer scientists and engineering have also investigated this effect in hopes of using functional descriptions for computer implementations. This presentation describes how understanding and implementing the cocktail party effect is important for applications in smart rooms and surveillance. Recent approaches developed at the Center for Visualization for isolating voices in noisy environments will be described, and early results using a combination of multiple microphones and processing inspired by the human auditory system will be presented. This talk also presents an overview of the Ambient Virtual Assistant (AVA) project at Center for Visualization at the University of Kentucky, along with a description of resources available through for research activities associated with smart spaces and virtual environments.

Kevin D. Donohue received the B.A. degree in mathematics from Northeastern Illinois University, Chicago, IL, in 1981, and the B.S., M.S., and Ph.D. degrees in electrical engineering from the Illinois Institute of Technology, Chicago, IL in 1983, 1984, and 1987, respectively. He is currently with the Electrical Engineering Department at the University of Kentucky, Lexington, KY where he is the Databeam Professor of Electrical and Computer Engineering and directs the Audio Sensing and Rendering Laboratory at the Visualization Center. His current research interests include novel signal characterizations, audio and ultrasonic signal processing, and distributed sensor networks and systems. Dr. Donohue is a senior member of the IEEE, a member of Sigma Xi and the Audio Engineering Society.