Evaluation of remote pedestrian sensor system based on the analysis of car-pedestrian accident scenarios

摘要

There is a need to evaluate the effectiveness of a remote sensor system for pedestrian detection in a vehicle traffic environment. For this purpose, the car-pedestrian accident scenarios were analyzed using selected data from the Swedish Traffic Accident Data Acquisition (STRADA), a database developed by collecting accident data from the police and hospitals. The two most common scenarios were identified as cars entering and leaving intersections colliding with pedestrians crossing the road. The knowledge of these two scenarios was then developed in terms of the factors such as the pedestrian's trajectory, the pedestrian's speed, the car's trajectory and the car's velocity. Based on the developed knowledge, a mathematical model was presented and the remote pedestrian sensor system was evaluated by using this model. It was found that the sensor system can detect almost all the pedestrians in these two scenarios in time, in the case where the detective angle was larger than 60 degrees.