Impartial Sensor Tasking via Forecasted Information Content Quantification

Abstract

The consistently increasing population of space debris in orbit about the Earth is becoming an increasingly unavoidable problem, necessitating a proper handling of the problem of sensor tasking to effectively and efficiently obtain observations of space objects to maintain catalogs of their positions and velocities. A variety of approaches have been explored to achieve this in an autonomous fashion, assisting with space traffic management. This work presents new statistical analysis of the Kullback–Leibler divergence justifying the use of its expected value in generating sensor schedules. This divergence’s use as an objective measure is extended to simultaneously consider, in an impartial manner, multiple observations of potentially varying quality and type into a common space, namely the information space, by mapping them to an arbitrary reference time. This time mapping is then shown to exhibit improvements in performance not only with respect to the estimate at the reference time, but over the entire tracking interval.