Intent Inference of Deceleration Maneuvers and Its Safety Impact Evaluation by Driving Simulator Experiments

Abstract

Though the autonomous driving is becoming a key solution to mitigate rear-end collisions, the driving assistance such as Advanced Driving Assistance System (ADAS) is still required to support drivers to be aware of the risk of rear-end collisions. The authors have been developing the intent inference system that advises the appropriate deceleration timing to mitigate the collision risk with the preceding vehicle. The purpose of the system is to infer the intent of a vehicle ahead 1.5 s in advance of its driver’s most likely action. The proposed system has been well evaluated through a driving simulator experiment and a car-following simulation in previous studies by the authors. However, the car-following simulation itself is not enough to identify the safety effect of the assistance system on a vehicle platoon of a convoy. In this paper, the safety impact of the proposed assistance system on the platooned vehicles was assessed using additional driving simulator experiments. Numerical analyses showed that the proposed assistance was significantly effective for the safety of multiple vehicles in the platoon including the assisted car itself.