Human-Like Decision Making for Autonomous Vehicles with Noncooperative Game Theoretic Method
Peng Hang, Chen Lv, Xinbo Chen · 2022
The lane-change decision making is a common issue for AVs, which has been widely studied. The non-cooperative game theoretic approach is easy to describe the lane-change interaction and decision making. Due to the few players, the interaction process of the lane-change scenario is relatively simple. Two non-cooperative game approaches, i.e., the Nash equilibrium and Stackelberg games, are adopted to address decision making and interactions for AVs. Finally, the developed human-like decision-making algorithms for AVs are tested and validated via simulation in various scenarios. Besides the Nash equilibrium game theoretic approach, Stackelberg game theoretic approach is another noncooperative game approach. To verify the two kinds of game theoretic human-like decision-making algorithm, two driving scenarios are designed and carried out based on the MATLAB-Simulink simulation platform. This chapter presents a human-like decision-making framework for AVs. In the decision-making issue of AVs at unsignalized roundabouts, AVs are limited within the control boundaries.