Intelligent fault diagnosis for seeker based on least squares support vector machine with genetic algorithm
基于遗传优化最小二乘支持向量机的导引头故障诊断

In order to improve the correct rate of seeker fault diagnosis, this paper proposed a multi-class classification model for seeker based on LSSVM (least squares support vector machine) optimized by the improved genetic algorithm. The model constructed the multi-class LSSVM classifiers of two layers according to the one-against-one strategy and the improved vote method. And the genetic algorithm which was improved by an adaptive search strategy with a variable step length was employed to select the kernel function parameter and the regularization parameter of LSSVM. Then the data of diagnosing missile seeker were brought into the model to test its functions. Result is shown that the method has higher diagnosis accuracy and computational efficiency by comparing with the standard SVM and BP neural network diagnosing method.