MEMS Seminar: Distributed Data Fusion - Neighbors, Rumors and the Art of Collective Knowledge

Distributed data fusion is the process whereby a group of agents sense their local environment, communicate with other agents, and collectively try to infer knowledge about a particular process. The applications are many: cooperative robots mapping a room; cooperative UAVs tracking a moving object on the ground or searching for survivors; a distributed formation of space telescopes; a game of hide and seek with kids; a group of people discussing an interesting issue, either in person or on-line. This talk will use a paradigm called Bayesian Distributed Data Fusion (DDF) to explore the challenges, solutions and applications related to a group of agents working together sensing, communicating, and inferring processes in their environment. These challenges include: how to build a scalable solution; how to maintain consistent estimates and avoid rumor propagation; how to handle complex information types (such as semantic human inputs); and how to handle varying network topologies. The theoretical underpinnings of these solutions will be studied, along with a series of applications focused cooperating UAVs, robots, and people.

Lunch will be served from 1-1:30 pm.