Prediction of pedestrian risk at signalized intersections

Abstract

Signalized intersections in an urban area are expecting the high volume of mixed traffic with various conflicting movements. This situation would greatly risk the safety of pedestrian as vulnerable road users, especially in a busy capital city like Kuala Lumpur. However, predicting risk based on accident data would neglect many risk factors associated. Thus, this paper presented an application of the pedestrian crossing risk assessment (PedCRA) model using Petri Nets approach to be compared with the risk calculated using accident data for thirty signalized intersections in Kuala Lumpur. The prediction of risk using this model is able to consider various risk factors, including illegal behavior, signal setting, traffic volume, road geometric layout and environment. Prediction of risk for selected sites were compared with the pedestrian accident data for 5 year period as the actual risk value. The Chi-Square goodness of fit was performed and the result showed that the predicted risk value and actual risk values follow the same distribution pattern.