The Stochastic and Integrative Prediction Methodology and Modeling for Reliability of Pedestrian Crossing on Roads

A very large number of pedestrians die every year across the globe in developed and developing countries. A pedestrian is under constant influence by various factors including the approaching vehicles equipped with Advanced Driver Assist Systems (ADAS), road infrastructures, environment and pedestrian's state, abilities, and conduct. The authors’ view in the analysis of road accidents has been to identify and address the root causes of pedestrian-to-vehicle accidents. It is a simplistic approach to blame human error as the cause of accidents whether on part of drivers or pedestrians and not to address the causes of human errors logically, systematically, and dynamically. The interfaces and dependencies among identified causes of pedestrian-to-vehicle accidents as unwanted events have been illustrated using fault tree in this research. Furthermore, the time-dependent reliability and failure probability functions of each of the contributing causes of this unwanted event have been modeled distinctively. Subsequently, the failure probability and reliability functions of pedestrian crossing on roads are determined. The results of this research will assist researchers and practitioners to logically and stochastically predict and analyze risk of the pedestrians involved accidents and provide measures in elimination or reduction of this risk.