Application of POMDPs to Cognitive Radar

In recent years, hardware advances have resulted in software configurable radar systems that lend themselves well to decision-making systems. Partially observable Markov decision processes (POMDPs) are evaluated herein as a framework for decision-making in radar scenarios, and value iteration is examined as a method for computing an optimal decision policy with a POMDP. A scenario is investigated wherein a radar is competing with a greedy agent for spectrum. Results demonstrate improvement over a heuristic decision-making agent that seeks to maximize immediate reward.