基于最小二乘支持向量机预测器的传感器故障检测与数据恢复

本文介绍了最小二乘支持向量机（LS-SVM）回归的基本原理，提出了一种基于LS-SVM回归的时间序列预测器，并将其用于传感器的故障检测和数据恢复。论述了LS-SVM预测器的实现方法和步骤，并且将其应用于压力传感器的故障检测和数据恢复，同线性神经网络预测器、RBF神经网络预测器和BP神经网络预测器的比较结果表明，LS-SVM预测器具有更高的预测精度，更好的外推能力，计算效率最高，因此，LS-SVM预测器是传感器故障检测和短期数据恢复的一种有效方法。

Sensor fault detection and data recovery based on LS-SVM predictor

In this paper, we introduce the fundamental theory of least squares support vector machine (LS-SVM) for regression, propose a predictor based on LS-SVM regression for sensor fault detection and data recovery, and present the principle of the predictor and its on line algorithm. This method was applied to a pressure sensor, and compared with linear neural network predictor, RBF neural network predictor and BP neural network predictor. The experimental results show that LS-SVM predictor has higher prediction and recovery accuracy, consumes least time than neural network predictors. It can detect sensor fault and recover sensor signal successfully.