Neural network-based radar signal classification system using probability moment and ApEn

Most of the existing electronic warfare systems use a threat library to identify radar signals. In this paper, new feature parameters for classifying various types of radar signals are introduced. The conventional method uses frequency, pulse repetition interval and pulse width sampled from the pulse description word column as characteristics of a signal. Such sampling technique cannot effectively model each radar signal when dealing with a complex signal array. This paper proposes probability moment and ApEn as an effective feature for the development of high-performance radar signal classifier. As shown in results, the proposed method can effectively classify ambiguous radar signals in the existing system because the signal values are similar but the order is different. In order to verify the performance of the proposed system, 100 types of radar signals in various bands were simulated, and the performance yielded 99% positive classification rate of the 100 radar signals.