基于QPSO的雷达辐射源信号识别方法研究

Study of signal recognition of radar emitter based on QPSO

鉴于支持向量机(SVM)方法对雷达辐射源信号具有较理想的识别结果,但对模型参数没有具体选择方法的问题,设计了一种以具有量子行为的粒子群优化(QPSO)算法为参数优化方法的SVM分类器,并提出了基于QPSO-SVM的雷达辐射源信号识别方法.QPSO-SVM分类器在采用QPSO算法对SVM进行优化改进的同时,继承了SVM分类器泛化能力强的特点,对雷达辐射源信号识别问题具有良好的适应性.实验结果表明,与其他方法相比,本文方法在保证识别准确率的同时,降低了参数选择时间.

Considering that the SVM algorithm has an ideal recognition result for radar emitter signals, but no specific selection method for the model parameters, this paper designs an SVM classifier that uses QPSO algorithm as the optimization algorithm of parameters, and proposes a scheme of signal recognition of radar emitter based on QPSO-SVM. While the QPSO-SVM classifier optimizes and improves SVM using the QPSO algorithm, the proposed classifier keeps the strong generalization capability of SVM classifier, thus having the better adaptability for the signal recognition of radar emitter. Experimental results show that this proposed method can shorten the time for parameter selection while guaranteeing the accuracy rate of recognition, compared with other algorithms.