PARALLEL MARKOV CHAIN MONTE CARLO FOR SENSOR SCHEDULING

In this paper, we present a novel parallel Markov Chain Monte Carlo(MCMC) based solution to the sensor scheduling problem. In the context of space situational awareness, the objective of sensor scheduling is to maximize the information gain from observing a large number of space based targets using a limited number of sensors. The parallel MCMC approach is a sampling based optimization approach that can explore the space of configurations efficiently and quickly. We consider a scheduling scenario that involves a single sensor and multiple targets. The parallel MCMC method is used to obtain the look directions of a ground based sensor that maximizes the information gain over a receding horizon. The effectiveness of the new method is demonstrated through a simulation study.