Multiplatform information-based sensor management: an inverted UAV demonstration
Chris Kreucher, John W. Wegrzyn, Michel Beauvais, R. S. Conti · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2007
This paper describes an experimental demonstration of a distributed, decentralized, low communication sensor management algorithm. We first review the mathematics surrounding the method, which includes a novel combination of particle filtering for predictive density estimation and information theory for maximizing information flow. Earlier work has shown the utility via Monte Carlo simulations. Here we present a laboratory demonstration to illustrate the utility and to provide a stepping stone toward full-up implementation. To that end, we describe an inverted Unmanned Aerial Vehicle (UAV) test-bed developed by The General Dynamics Advanced Information Systems (GDAIS) Michigan Research and Development Center (MRDC) to facilitate and promote the maturation of the research algorithm into an operational, field-able system. Using a modular design with wheeled robots as surrogates to UAVs, we illustrate how the method is able to detect and track moving targets over a large surveillance region by tasking a collection of limited field of view sensors.