LS-SVM Based Human Recognition on Automotive Active Safety System

In this article, a human recognition method, using wavelet transform and least square support vector machine (LS-SVM), is introduced. This had been performed in automotive active safety system. This method takes two-dimension discrete wavelet transform (2D-DWT) to do pretreatment, and extracts coefficients of correlation as eigen-coefficients, and then uses LS-SVM to search humanoid in certain area. The searching result will be presented to controller of a vehicle for making decision. A real vehicle experiment proves the validity of proposed method.