A neural network ATR for high range resolution radar signature recognition of moving ground targets

High range resolution (HRR) radar produces high resolution target signatures of moving ground targets that can be separated from ground clutter using Doppler processing. As a result, HRR sensors are a leading candidate to provide all weather, day/night, long-standoff identification of moving ground targets. Previous automatic target recognition (ATR) studies have demonstrated the usefulness of HRR ATR with traditional statistical techniques such as mean-squared error template-based classifiers. This paper presents the results on an initial investigation of the use of artificial neural network (ANN) classifiers for HRR ATR in a continuous tracking (CT) scenario.