PCA-Based Vehicle Size Estimation in High-Resolution MIMO FMCW Radar System

In this paper, we introduce a new approach to simultaneously estimate the length and width of the vehicle ahead using a high-resolution automotive frequency-modulated continuous wave (FMCW) radar sensor. First, the point cloud data is generated by detecting a moving vehicle ahead using a high-resolution FMCW radar. The proposed vehicle size estimation method is based on the principal component analysis (PCA). The length and width of the target are estimated by orthogonally projecting each point onto the eigenvectors of the point cloud data obtained using the PCA method. The length and width of the target estimated using the proposed method results in average error rate of 2.18 % and 2.51 %, respectively. Furthermore, the efficiency of the proposed method has been validated under the presence of substantial road noise, demonstrating its robustness and reliability in traffic applications.