A new recognition system for radar emitter signals

摘要

Purpose - The purpose of this paper is to develop a new recognition system for identifying advanced radar emitter signals (RES). Design/methodology/approach - Initially, the framework of the new recognition system is outlined. Then, feature extraction using resemblance coefficient and wavelet packet decomposition, and feature selection based on quantum-inspired genetic algorithm, and a classifier combining K-means clustering, support vector machines and Mahalanobis distance are applied to actualize the recognition system. Finally, experiments are carried out on RES. Findings - A valid recognition system with its framework and implementation is presented to solve the difficult problem of advanced RES recognition. Research limitations/implications - Initial investigation is made on modern RES recognition. Further work may be done on decreasing the error rates and enhancing recognition efficiency. Real signals instead of simulated signals can be applied. Practical implications - The system developed here can be applied to electronic reconnaissance systems such as electronic support measures, electronic intelligence and radar warning receiver. Originality/value - The paper presents a novel recognition system and its implementation for modern RES. Extensive experiments conducted on 155 RES with eight intra-pulse modulations show the feasibility and validity of the introduced system.