Interactive on-ramp merging decision-making with motion prediction considering merging motivation

Abstract

On-ramp is one of the scenarios most likely to cause traffic accidents and congestion. In order to mitigate the conflict between mainline vehicles and ramp vehicles, and improve vehicle safety and travel efficiency, we propose an interactive on-ramp merging decision-making (IORMDM) method based on motion prediction. We first propose a motion prediction method for different actions of vehicles based on model predictive control. Meanwhile, the game theory is employed to construct the interaction behavior between the ramp vehicle and the mainline vehicle, and the merging motivation of the ramp vehicle is modeled to describe the collision risk between it with the end of the acceleration lane. Afterward, according to the predicted trajectories of vehicles, the comprehensive payoff of vehicle collision risk, efficiency, passenger comfort, and merging motivation of each action is solved. The Nash equilibrium solution is solved to determine the optimal action of each vehicle according to the payoff table. Simulation experiments demonstrate that our proposed method could reduce collision risk and improve traffic efficiency.