基于决策树SVM的某型无人机发射机故障诊断

Fault diagnosis of transmitter for UAV based on decision tree SVM

为能够准确地对无人机数据链发射机进行故障诊断,通过对某型无人机数据链发射机工作原理和故障模型分析,结合长期的维护、保障经验,提出了一种基于故障优先级的决策树支持向量机故障诊断方法,避免了现有的多类分类支持向量机在多类故障诊断方面存在不足,优化了多类分类支持向量机组合策略。经试验,与几种常用的多类分类支持向量机方法对比,该诊断策略有效得提高了故障诊断正确率,能够准确地定位发射机内部故障功能模块,具有一定的实际意义。

To accurately troubleshoot the data-link transmitter of UAV,by analyzing the working principle of the data-link transmitter of UAV and the fault model,combined with long-term maintenance experience,a decision tree SVM of fault diagnosis method based on the fault priority is proposed.The method can avoid the disadvantage of the existing multi-class classification SVM in the fault diagnosis and optimize the combination of the strategies of the multi-class SVM.After testing,compared with commonly used multi-class classification SVM,the diagnostic strategies can effectively improve the correct rate of the diagnosis,accurately locate fault within the transmitter modules and has some practical significance.