A human-like response model for following vehicles in lane-changing scenario

Abstract

A human-like following vehicle (FV) model for lane-changing scenario was developed. First, the non-linear relationship between FV responses and relative speed was integrated. Normalized transformation was performed to better investigate the correlation characteristics between variables. Second, environmental factors and kinematic parameters were characterized using a hierarchical linear model. Next, vehicle control modes were incorporated. Finally, a human-like response model (HRM) for the FV was proposed and tested. It is found that the brake intensity of the FV is affected by lane-changing direction, but it is not affected by turn-signal usage or traffic density. Turn-signal usage and traffic density are related to kinematic parameters. Lane-changers use their turn signal more when relative velocity is negative or when traffic density is high. Therefore, turn-signal usage and traffic density will become redundant variables when kinematic parameters are considered. Driving data verification shows that the HRM can match both non-critical and critical lane-changing cases and can simulate the no reaction behavior of the FV in lane-changing, which can help to improve the human-like driving ability of intelligent vehicles.