Based on principal component analysis and D-S evidence theory and application of sensor fault diagnosis
基于主元分析和D-S证据理论的传感器故障诊断与应用*

Because the type sensor for underground complex, the measured parameters of the data are huge, use principal component analysis to reduce the dimensions of the data. Use RBF neural network to carry out the feature level data fusion, and establish distribution function of basic trust, further use the evidence theory advantage of representation and reasoning to inaccurate information, realize the fault detection and isolation capabilities effectively. The simulation show that the use of principal component analysis and DS theory can be correctly located and accurately isolate the failure of sensors.