Noncooperative-Game-Based Intelligent Vehicle Decision Method for Lane-Changing Interactive Behavior

In the future transportation, it is a foreseeable trend that human-driven vehicles and autonomous vehicles will coexist and interact on roads for a long time. How to integrate autonomous vehicles into human drivers’ traffic ecology and make behavioral decisions and predictions in the complex environment considering vehicle-vehicle interaction is worthy of deep consideration. Therefore, the proposed game theoretic method is applied to automatic lane-changing scenarios. Then, the safety risk, ride comfort and traffic efficiency of vehicles in game are quantitatively evaluated as cost functions. Specifically, the driving maneuvers of interactive agents could be predicted by reasoning, and the optimal strategy is obtained by calculating Nash equilibrium. Finally, simulations are conducted using Commonroad under typical driving conditions. The results illustrate that our model could help autonomous vehicles make reasonable and explainable lane-changing decision when interacting with heterogeneous agents, demonstrating the feasibility and reliability of our proposed approach.