Research On Personalized Lane-Change Trajectory Planning And Control Methods For Autonomous Vehicles

The development and popularization of autonomous vehicles provided new opportunities to improve road safety and enhance traffic efficiency.Vehicle lane-change is a crucial driving behavior in achieving the safety and intelligence of autonomous vehicles,and it has become a research focal point in the field of autonomous driving.Lane-change trajectory planning and trajectory tracking control were both essential components of a vehicle’s lane-change process.However,the existing lane-change trajectory planning methods failed to fully consider the driving preferences of both drivers,and the complexity of the lane change process required higher standards for trajectory tracking control.Therefore,this paper conducted relevant research on personalized lane-change trajectory planning and tracking control methods.The study aimed to address these challenges and improve the accuracy and efficiency of autonomous vehicle lane-change.The specific contents of this research included:(1)Safe lane-change trajectory cluster generation.In order to ensure the safety of the lanechange process,the safe lane-change area of vehicles was determined based on the potential field method.Next,the symmetric exponential moving average algorithm was used to filter and smooth the lane-change data in the NGSIM dataset.The vehicle state and environmental information at the beginning and end points of the vehicle lane-change were extracted as training samples.A prediction model for vehicle lane-change duration and lateral displacement was developed using particle swarm optimization and BP neural network.The prediction results of the model were then used as lane-change boundary conditions to generate a safe lane-change trajectory cluster.(2)Personalized lane-change trajectory selection for drivers.To account for the fuzziness of human thinking,an improved optimal and worst method based on fuzzy set theory was used to identify the driving preferences of both drivers.This involved obtaining the weights of indicators that reflected the driver’s preferences,such as driving safety,riding comfort,and lanechange efficiency.Based on the obtained indicator weights,the VIKOR method was used to rank vehicle lane-change tracks.This process resulted in identifying lane-change tracks that met the personalized needs of drivers.(3)Design of track tracking controller for autonomous vehicle lane-change.In order to simplify the calculation of the model and take into account the motion characteristics of the vehicle,a two-degree-of-freedom model of the vehicle was established.Based on this model,an LQR controller was designed in the lateral direction to control the front wheel angle of the vehicle,thereby achieving accurate tracking of the lane-change trajectory.Additionally,considering the driving state of the vehicle,such as speed,acceleration,and position,a speedposition dual PID controller was designed to achieve vehicle longitudinal speed tracking.(4)Verification of personalized lane-change trajectory and its tracking control effect.To validate the effectiveness of the proposed approach,a hardware-in-the-loop experimental platform was built using a Logitech G29 driving simulation simulator to collect lane-change data from drivers with different driving preferences.Based on the collected data and the planned trajectory,a similarity analysis was conducted to verify the rationality of the lane-change trajectory.Additionally,a software simulation platform for trajectory tracking control was built using Prescan,Matlab/Simulink,and Carsim software to evaluate the tracking control effect of the controller.The results showed that the generated trajectory was able to meet the user’s expectations in the driving environment while ensuring safety.Specifically,the similarity between the planned trajectory and the real driver reached 86%.The tracking error and obtained front wheel steering angle meet the experimental requirements,which verifies the effectiveness of the designed controller.In summary,this paper conducted a comprehensive research on the personalized lanechange trajectory planning and tracking control model of autonomous vehicles.A personalized lane-change trajectory for both drivers was generated in the safe lane-change area,and accurate tracking of the generated trajectory was achieved through LQR and PID controllers.An experimental platform was built using a driving simulator and simulation software to verify the effectiveness of the designed trajectory planning method and tracking controller,providing valuable technical support for the promotion and application of autonomous driving technology.