A new method based on MTL and LS-SVM for crosstalk predicting in electric vehicle

In electric vehicle crosstalk plays an important role in electromagnetic compatibility. However, the influence factors of crosstalk of lines in electric vehicle are very complex and variable. To get higher precision prediction, as many of the factors as possible are input in the forecast model in predicting computing. Multi-conductor transmission line (MTL) is one of multivariate electromagnetic analysis, which achieves parsimony and reduces dimensionality to simplify computation by extracting the smallest number of irrelevant components with little loss of information. In this paper, a new method for crosstalk predicting based on MTL and least squares support vector machine (LS-SVM) is presented. Firstly components are extracted from various factors of crosstalk by MTL and to be inputs of LS-SVM. Then LS-SVM is applied to train and forecasting. The model is characterized by simple computing. Analysis of the experimental results proved that the method proposed achieved greater accuracy and efficiency than conventional LS-SVM.