Pedestrian detection based on automotive radar

Automotive radar sensors are able to measure the target range, azimuth angle and radial velocity simultaneously with high accuracy and update rate even in multiple target situations. Target recognition of radar objects is still a technical challenge and one of the most important topics to support future advanced driver assistance systems (ADAS). There is especially a large interest to distinguish between radar echo signals from human beings and from other objects in the target recognition scheme. Therefore, this paper elucidates the motion sequence of a walking pedestrian and describes this complicated process in a suitable technical model with different reflection points and related Doppler frequencies. Based on some measurements with a 24 GHz radar sensor several signal features are generated, which are used for the target recognition procedure.