Modelling overtaking manoeuvre of motorized two wheelers on urban undivided roads

Abstract

Motorized Two-wheelers (MTW) are popular in congested urban areas with heavy traffic since they offer a quick and adaptable means of transportation. Overtaking and lane changing manoeuvers happen when traffic does not flow at the intended speed. They cannot be avoided, especially in mixed traffic scenarios when there is a constant speed differential between fastmoving and slow-moving cars. Collisions during overtaking manoeuvers are one of the leading causes of two-wheeler fatalities among crashes involving two-wheelers. Considering these issues, there is a need to perform thorough analysis of the overtaking manoeuverability of MTW. In this study, the MTW overtaking manoeuverability is modelled using binary logit modelling (BLM) and support vector machine (SVM) and both were compared using performance parameters. The data was collected using video-graphic survey and was extracted using Kinovea. The speed of subject MTW, overtaken vehicle and the oncoming vehicle, the presence of pillion rider, lateral and longitudinal distances were the significant input variables for BLM. The performance parameters obtained from confusion matrix showed SVM being better than BLM. This study can be used to prepare an advanced driver assistance system (ADAS) which can increase the safety of MTW drivers while overtaking on undivided roads in urban areas.