Risk Generation and Identification of Driver–Vehicle–Road Microtraffic System

Abstract

The highly nonlinear and uncertain driver-vehicle-road (DVR) traffic system will be unstable and cause potential risks under certain conditions. This paper analyzes the attributes and interactive characteristics of DVR, and constructs the physical and mathematical models of the DVR microtraffic system. We first consider the potential accident consequences caused by vehicle-road interaction, the behavior uncertainty of driver-vehicle interaction and the risk sensitivity of driver-road interaction, construct a DVR system model to characterize the influence of DVR interaction process on the system safety. Further, by revealing the risk generation mechanism of a DVR system, we can achieve system risk identification and early warning. Experimental results verified by naturalistic driving dataset show that the risk trend generated by the DVR system model is consistent with the output of traditional risk indicators (THW, TTC). Compared with TTC, the proposed model is more universal to different traffic participants and road topology scenarios and can assist DVR systems to make early warning and control strategy.