雷达抗干扰性能评估是雷达系统研制、引进、装备过程中必要的环节. 如何综合评估复杂电磁环境下的雷达抗干扰性能评估已成为研究的重点. 针对现有雷达抗干扰性能评估方法的特点和局限性, 提出了一种基于粗糙集-自适应神经网络模糊推理系统(RS-ANFIS)的性能评估方法. 首先, 针对原始样本数据的不完备性和不确定性, 采用粗糙集理论对原始样本数据进行数据归一化、 离散化、属性约简处理, 并得到覆盖原始样本特征的最小规则集. 其次, 建立了基于ANFIS的Sugeno型性能评估模型, 设计了评估变量的隶属度函数和推理规则, 确定了评估网络各层输入输出关系以及网络学习算法. 最后, 以12组雷达抗干扰性能评估指标为例进行算法模型验证, 表明了方法的可行性和模型的有效性. 实验结果表明, 该方法能够有效改进网络结构, 提高雷达抗干扰性能评估结果的可信度.

The radar electronic counter-countermeasures (ECCM) capability evaluation (CE) method is the essential step in the radar development, introduction, and service process. Consequentially, the CE for radar ECCM in the complex electromagnetic environment becomes a research focus point. Considered with the present research situation, a technique for CE based on rough sets and adaptive neuro fuzzy inference system (RS-ANFIS) is proposed to solve the properties and vulnerabilities existed in the CE methods. According to the uncertainty and imperfection of the original sample data, the rough set is used to preprocess for the normalization of data, the discretization of continuous data and the attribute reduction in order to obtain the minimum feature subset. Then a model for CE based on ANFIS with Sugeno type is established. The membership functions and the inference rules of the system variables are devised with computational relations between layers of the input and output and the learning algorithm of neural network. Subsequently, 12 typical sample sets are used to check the validity and rationality of constructed model. The experiment results show that the proposed method can effectively optimize the structure of neural networks and make the radar ECCM capability evaluation more feasibly and practically.