Visualizing Effectiveness at Reducing Traffic Accidents—Enhancing Simulation Accuracy—

Abstract

Fostering social acceptance is necessary for the smooth introduction and use of automated vehicles and vehicles providing driver assistance. In this project, we reproduced the traffic environment of selected model cities using a multi-agent simulation of the perception, recognition, decision-making, and actions of various traffic participants (developed in a project⁽¹⁾ in the first phase of SIP-Automated Driving for Universal Services). We then integrated driver distraction and other factors that cause accidents in the simulation to recreate real-world accident conditions. We also used the automated driving (driver assistance) system deployment scenarios for every five years from 2015 to 2050 provided by the Socioeconomic Impacts of Reducing Traffic Accidents project to estimate effectiveness at reducing traffic accidents. Since those deployment scenarios differ for each vehicle classification, we further subdivided those classifications and defined rates of automated driving (driver assistance) system market propagation separately for categories such as passenger car, bus, and truck. The nationwide effectiveness at reducing traffic accidents was estimated from traffic accident statistics using the ratio of accident reduction calculated using the simulation applied to model cities.