Game-Theoretic Lane Change Decision-Making Method Considering Traffic Trend

In order to improve the safety and adaptability of the lane-changing decision process of autonomous vehicle, a lane-changing decision-making method considering the traffic trend is proposed in this article. In the proposed method, the longitudinal and lateral driving intentions of the vehicle are predicted by nonlinear autoregressive with external input neural network and Gaussian mixture models and hidden Markov model, respectively, which is trained using the next generation simulation traffic database. Besides, the payoff matrix is constructed based on the game theory with the interaction of other vehicle and future traffic trend both considered. The advantage of proposed method is that it not only takes into account the complex interaction of other vehicle in the lane-changing decision-making process, but also highlights the flexible adjustment to different traffic trends. Several typical lane-changing scenarios tests and analysis are given under a hardware-in-the-loop testing platform to verify the effectiveness and real-time implementation performance of the proposed method. Notably, related results are compared with the lane-changing method based on traditional game theory. The results show that the proposed method is able to make feasible and safe decision, which is in line with the actual driver and correct corresponding unreasonable decisions by traditional game theory, which verifies the effectiveness of the proposed method.