Fully Adaptive Radar for Target Classification

Abstract

In this paper, the fully adaptive radar (FAR) framework is applied to the problem of target classification. The FAR framework mimics the perception-action cycle of cognition and adapts its sensing based on its perception of the environment, system performance goals, and sensor costs. Performance is demonstrated on simulated data and on experimentally collected data from a pulse-Doppler radar testbed in which micro-Doppler features are used to classify the motion of a moving target. In both examples, the FAR classification algorithm is able to achieve the system performance goals while conserving system resources in a dynamically changing environment.