FUZZY ESTIMATION OF VEHICLE SPEED IN PEDESTRIAN COLLISION ACCIDENTS

摘要

This paper proposes a fuzzy tool capable of determining a pedestrian's trajectory and vehicle collision speed based on relevant data readily obtainable from an accident site or reasonably approximated through a simulation analysis. The input parameters for the model can also take into account the longitudinal gradient of the road. In developing the proposed fuzzy tool, valid membership functions were determined using references from not only actual collision accidents and experimental test data but also 390 cases supplemented through computer simulation. The fuzzy tool developed deduces the category of the pedestrian's trajectory and the collision speed of the vehicle using 5 membership functions and 2 logic rules. The logic rules are organized into 2 stages, the first being designed to identify the pedestrian's trajectory and the second being designed to determine the vehicle collision speed. The throw distance membership function was set according to the deduced trajectory category. The fuzzy tool was implemented in a computer program and validated by comparison with the empirical and computer simulation data. This fuzzy tool, which uses information easily collectable at an accident site, may be more precise and of more practical value than other popular methods of analysis of pedestrian collisions.