Radar targets identification based on perfect parametric modeling using artificial neural networks

In this paper, we have developed a new target recognition procedure. In order to efficiently obtain feature vectors for target discrimination, the closed-form expression of geometrical wave front have been derived to provide efficient and accurate computation. The resulting small dimensional feature vectors based on the developed extractor are identified using the well-known ANN classifier. The results show that the presented scheme gives successful correct automatic target recognition (ATR) in the low SNR with low computational costs. Although this work concentrates mainly on parametric modeling of electromagnetic target, further work should study more efficient ANN architecture when the number of candidates is dramatically increased in the practice.