Vehicle Classification via Multi-dimension Feature Extraction with Millimeter Wave Radar

Accurate vehicle classification is an important issue in autonomous driving and traffic monitoring technology. In order to improve the classification accuracy of road vehicles in passive traffic monitoring scenario with millimeter wave (MMW) radar, a vehicle classification algorithm based on multi-dimension feature extraction and support vector machine (SVM) is proposed in this paper. According to the scattering points distribution of the vehicle in range-doppler-angle domain, 9 statistical features are mined and extracted to obtain the feature vector, and then an SVM classifier is used to realize vehicle classification. The real data experimental results based on a 77GHz MMW radar verify the effectiveness and superiority of the proposed method in road vehicle classification over the classification method based on point cloud feature and SVM.