Automatic Lane-Changing Decision Based on Single-Step Dynamic Game with Incomplete Information and Collision-Free Path Planning

Traffic accidents are often caused by improper lane changes. Although the safety of lane-changing has attracted extensive attention in vehicle and traffic fields, there few studies considering lateral comfort users decision-making. Lane-changing decision-making single-step dynamic game with incomplete information path planning based on Bézier curve proposed this paper to coordinate performance from payoff, velocity payoff. First, distance which is improved collecting data through simulated driving, time obtained introduced into so that selection start considers safety, power passenger process. Second, without collision forward constrained curve, further obtain a smoother path. The tracking sliding mode controller front wheel angle compensation radical basis function neural network designed. Finally, model loop simulation hardware experiment carried out verify advantages method. results three conditions designed show performance, controlled method better than human drivers discretionary change mandatory scenarios.