基于AIT方法的雷达辐射源信号识别效能评估

摘要

为了有效对雷达辐射源信号识别效能进行评估，建立了基于识别测试结果的评估指标体系，提出了利用AHP法求解权值、采用区间TOPSIS法计算综合评分值的识别效能评估方法(AIT)。以复杂度特征作为实验所用的特征参数，采用概率神经网络(PNN)、支持向量机(SVM）、基于遗传算法的支持向量机(GA-SVM)和基于粒子群算法的支持向量机(PSO-SVM)四种识别算法进行了仿真实验。仿真结果验证了所得结果是合理的，该评估方法是可行的。%In order to assess the effectiveness of radar emitter signal recognition, this paper sets up an evaluation index system based on recognition test results, and puts forward a novel recognition effectiveness assessment method by using AHP method to solute the weight and utilizing the TOPSIS method to compute the synthesized assessment grades. By taking the complexity as experimental characteristic parameter, simulation experiment is performed on the four types of recognition algorithms in terms of using the probability neural network (PNN), support vector machine (SVM), SVM based on genetic algorithm (GA-SVM) and SVM based on particle swarm optimization (PSO) algorithm (PSO-SVM). Simulation proves the achieved results are reasonable, and this proposed method is available.