Pedestrian accident prediction modelling—A case study in Thiruvananthapuram city

Abstract

Pedestrians are the most neglected road users and witness an alarming accident rate each year. In India, annually, around one-third of road accidents involve pedestrians as victims. The pedestrian accident-prone urban road stretch along National Highway 66 in Thiruvananthapuram City was selected as the study stretch. In this paper, a statistical model for predicting the number of pedestrian accidents and their severity levels such as grievous, minor and fatal injury was developed for urban midblock sections. Poisson regression models were found to be the most suitable model to fit the data. Several factors such as number of bus stops, speed of vehicles, traffic volume, pedestrian volume along footpath and crosswalk, side roads, signalized crossings and the quality of crossings were found to considerably affect the pedestrian accidents. This paper helps to provide an insight into the hazardous conditions to pedestrians in urban areas and also proposes recommendations for enhancing the safety of the pedestrians.