An Improved Representation of Measurement Information Content Via the Distribution of the Kullback-Leibler Divergence

Proper utilization of sensor networks is key in target-dense or measurement-scarce environments, such as in the creation and maintenance of reliable records for space objects in Earth orbit. In recent years, there have been many investigations of utilizing different information-theoretic measures as performance measures in allocating sensor tasks to maximize the information gained. More specifically, information divergences have been considered in sensor tasking schemes to effectively and efficiently utilize the available sensor resources. However, it is typical that only the expected information gain with respect to the measurement likelihood is considered, while the rest of the distribution of the divergence in question is disregarded. This work studies the full distribution of the Kullback-Leibler divergence and if the utilization of this knowledge when committing to an action regarding the acquisition of measurement information is beneficial