Doctor of Philosophy, The Ohio State University, 2005, Electrical Engineering

Sensor management (SM) is one of the key factors that determine the performance of a multi-sensor system, especially when the system consists of mobile sensor agents (MSA). Due to the strong interconnection between information processing and MSA motion control, a new set of theoretical foundations, design principles and performance metrics are needed. This dissertation presents a few attempts toward this goal by jointly studying the target-track-maintenance problem and the MSA motion-planning problem through the MSA-target scenario. First, a generic method for target track maintenance, the BF-HMap approach, is proposed based on the Bayesian filtering method and the hospitability map . An advanced version of this approach with much less computational and memory load using a particle filter, the PF-HMap algorithm, is also introduced in this work. Both BF-HMap and PF-HMap are capable of exploiting non-analytic prior environmental knowledge as well as handling intermittent and regional measurements caused by the coverage and motion constraints on MSAs. Meanwhile, a generalized particle filter for both in-sequence and out-of-sequence measurements is developed for possible extensions of the PF-HMap algorithm to distributed MSA networks. Secondly, the MSA motion-control problem is studied in such an information-theoretic way that the conditional entropy (i.e. given the measurements) of the target state is chosen as a generic performance metric. Several key properties of the evolution of the entropy are identified, which are further exploited to model the sensor management problem in two cases of studies: target search and target surveillance, which are modeled as a stabilization problem of the entropy and an optimization problem to minimize the average revisit time on each target, respectively. Based on that, a necessary condition and a sufficient condition by means of the number of MSAs to perform a non-escape search for a moving target are derived. In the meantime, a co (open full item for complete abstract)

Committee: Umit Ozguner (Advisor) Subjects: