Cognitive Radar

Systems & Technology Research (STR) is pleased to provide this Phase 1 proposal to develop neural network-based hierarchical cognitive radar architectures inspired by research in cortical memory and neuroplasticity. The U.S. Navy airborne radars operate in challenging environments against ever improving adversary systems and capabilities. While adaptive architectures that employ feedback between the receiver and the transmitter provide performance improvements, they will be challenged by emerging electronic attack and concealment techniques. The proposed effort will develop cognitive architectures with plasticity and retention enabling the radar to learn perception/action pairs, store them as memories and recall actions from memory as the perceived RF environment changes. We will implement these capabilities using cognits – neural networks that represent a unit of memory – and evaluate their efficacy in maritime and littoral surveillance missions. Throughout the effort, STR will work to establish a cognitive radar system development plan for Phase II using the fundamental approach developed during Phase I and the inclusion of feedback from the U.S. Government team to hone the research towards a direction that is amenable for a successful transition to the Navy.