FMCW Radar-Based Vehicle Detection Techniques for Pedestrian Safety Systems

An FMCW radar-based detection system was proposed to identify approaching vehicles and enhance pedestrian safety. The proposed system combined Doppler-based velocity estimation, the power spectrum density analysis, and a Kalman filter algorithm to effectively distinguish between vehicles and pedestrians, thereby achieving a high detection reliability. In addition, by analyzing the movement direction and trajectory of the detected objects, the system classified vehicles approaching or departing from pedestrians in real time and provided the monitoring information only for vehicles that might pose potential risks to pedestrians. Implemented using a commercial 77 GHz FMCW radar module, the proposed system successfully distinguished up to four vehicles simultaneously, achieving a detection accuracy of 98.03% and false alarm rate of 5.88%.