A cognitive radar network: Architecture and application to multiplatform radar management

Abstract

The objective of a cognitive radar network is to optimise radar performance in the highly variable mission environments that current operational systems encounter, while minimising its interference with other systems and its vulnerability to countermeasures such as jamming and anti-radiation missiles. A cognitive radar network may achieve these challenges by fully exploiting the available radar resources, sharing data among network components and taking into account prior environmental and situational knowledge as well as experience accumulated during operations. This knowledge can vary from high level information such as intelligence about the threat to low level information such as clutter maps. This paper presents a cognitive radar network architecture that supports this functionality and the application of (self)-learning methods. In this paper reinforcement learning is used to maximise the survivability of naval surface ships in a littoral scenario by managing the modes of an air surveillance radar.