Clutter reduction technique based on clutter model for automatic target classification in forward scatter radar

Abstract

Classification becomes one of the important elements in the forward scatter radar (FSR) micro-sensors network. This classification performance is dependent on the target’s profile behaviour and the network’s surrounding; and one of the factors that cause the reduction of classification probability is the presence of ground clutter. As the volume of clutter increases, their masking effect becomes greater and may result in more significant errors in target classification. Hence, to reduce misclassification in the FSR sensor network, a new clutter reduction technique based on the ground clutter model is proposed. Simulated ground clutter is modeled based on the estimated signal to clutter ratio (SCR) of the received signal. The clutter effect is diminished by eliminating simulated like-clutter from the receiving signals. The result shows improvement in the classification accuracy, especially for the minimum value of the SCR and this new technique uses only one database which will shorten the processing time and reduce the overall database’s size.