CoFAR: Cognitive fully adaptive radar

A new and fully adaptive environmentally aware (cognitive) radar and signal processing architecture is introduced to meet the challenges of increasingly complex operating environments. The system features fully adaptive transmit, receive, and controller/scheduler functions. “Cognition”, i.e., learning/understanding the complete multidimensional radar channel (targets, clutter, interference, etc.) and operating environment is achieved via a sense-learn-adapt (SLA) approach, which is a radar centric application of the OOPDA (observe, orient, predict, decide, act) loop concept. Learning in turn is achieved via expert system, knowledge-aided (KA) supervised training. Lastly, a MIMO probing approach is introduced as a learning aid for signal dependent channel effects and illustrated with a MTI radar example where it is shown that a full rank estimate of the clutter covariance matrix is possible from the returns in a single range bin, thereby alleviating the so-called “sample starved” covariance estimation problem that arises in highly nonstationary environments.