Analysis of impact of elderly drivers on traffic safety using ANN based car-following model

Abstract

The purpose of this study is to quantitatively investigate the impact of elderly drivers on traffic safety. Based on the data collected from a driving simulator, this study utilizes artificial neural network (ANN) to establish a car-following model with consideration of elderly drivers. The model validation demonstrates that the ANN based car-following model has a high fitting degree and strong predictive capacity. Furthermore, two surrogate safety indicators, number of potential collisions (NPC) and Time Integrated Time-to-collision (TTT) which characterize the number of potential collisions and the degree of potential crash risk respectively, are used to quantify the road traffic safety with elderly drivers. The simulation results show that car-following behavior of elderly drivers made a strong impact on the stability of the following platoon. The increase of elderly drivers could make the variability in the speed of all the followers greater, and lead to drastic increases in both NPC and TTT per vehicle per second. Besides, the two indicators increase firstly and then decrease as the stationary time of the first vehicle increases. And they decrease as the absolute value of deceleration for the first vehicle increases when the proportion of elderly drivers is higher. We hope the results in this study may be helpful to the development of road safety interventions under the aging population background.