Lane-change trajectory planning method for driverless vehicles based on trajectory prediction

Lane-change is a common and dangerous condition in the process of vehicle driving. Especially in the multi-vehicle environment, the relative safety between lane-change vehicles and surrounding vehicles is very important. In this paper, a lane-change trajectory planning method based on quintic polynomial is proposed by considering the trajectory prediction results of surrounding vehicles. Firstly, by fusing the driving state of the vehicle, the lane-change trajectory cluster is obtained by using the quintic polynomial programming method, and the stability boundary of the trajectory cluster is obtained according to the three-degree-of-freedom dynamics model of the vehicle. Then, in order to get a relative in trajectory cluster environment vehicle safety of lane-change trajectory, we assume that the information data of surrounding vehicles is available, according to different location of the vehicle, to establish a different objective function, to evaluate different lane-change trajectory of the vehicle relative security, according to the objective function to solve trajectory optimal solution in the cluster. Finally, the co-simulation based on MatLab and CarSim verifies the reliability of the proposed lane-change trajectory planning method.