Sensor-Management for Multitarget Filters via Minimization of Posterior Dispersion

This paper presents a new sensor management method for multitarget filtering, that is designed based on maximizing a measure of confidence in accuracy of the multitarget state estimate. Confidence of estimation is quantified by optimal subpattern assignment-based dispersion of the multitarget posterior about its statistical mean. Implementation of the algorithm for generic multitarget filters is presented. Simulation studies with labeled multi-Bernoulli filter demonstrate excellent performance in challenging sensor control scenarios.