An Information-Theoretic Approach to Sensor Resource Management for Situational Assessment

Toyon Research Corporation proposes to develop algorithms for the active management of multiple sensors necessary for a situational assessment (SA) system. In our approach dynamic sensor resource management is cast as an optimization problem where the goal is to minimize the information-theoretic entropy of a target database subject to sensor usage constraints. Choosing an appropriate partitioning of database entities into classes is the key to using existing sensor resource manager (SRMgr) technology to provide data of importance to a SA system. Our system representation is based on probability mass and density functions. As a result, a Bayesian algorithm will be used for transforming the system state from individual targets to entities of interest to SA. Toyon will demonstrate the effectiveness of this SRMgr using our software testbed SLAMEMT on an example problem requiring tracking and identification of enemy units. If we are selected to continue development of the concept in Phase II, we would work to extend our algorithm and test it on a variety of situations and perform system demonstrations to evaluate the performance of the resulting SA system.