Research on Lane-Changing Decision-Making for Autonomous Driving Based on Multi-player Dynamic Game Theory

In a hybrid traffic environment with autonomous vehicles (AVs) and human-driven vehicles (HVs), lane change decisions are critical for ensuring both safety and efficiency. This research introduces a novel lane change decision framework for AVs, grounded in multi-player dynamic game theory. It comprehensively addresses complex interactions among lane-changing vehicles, trailing vehicles in the target lane, and those in the original lane. The framework employs advanced computational techniques to determine optimal decision points, enabling safe, reliable, and efficient lane changes. Simulations using the NGSIM dataset demonstrate that this approach significantly outperforms traditional rule-based methods in terms of success rate and safety across diverse traffic conditions. This work substantially enhances traffic flow efficiency, reduces collision risks, and provides a robust and flexible strategy for deploying AVs in real-world mixed traffic scenarios.