Estimating Pedestrian and Characteristic Extraction for ADAS

Kikuta, R., & Kageyama, I. (2020). Estimating Pedestrian and Characteristic Extraction for ADAS. JASE Kanto Branch. Online: Society of Automotive Engineers of Japan.

This research discusses estimating pedestrian dynamics who wants to cross a road and characteristic extraction when pedestrians across a road. To estimate their behavior, we paid attention to the pedestrian joint posture angle. To analyze that, we used a motion capture system that is made by the VICON company. We obtain a pedestrian joint Euler angle and analyzed that data, a pedestrian will change their lower leg to thigh Euler angle when they across. According to that, Characteristic is existed on lower leg to thigh's posture. As a result, we understood that we had better use lower leg to thigh angle to estimate pedestrian dynamics.