Investigation Of Lane-change Trajectory Planning And Replanning Methods For Intelligent Vehicles In Multi-vehicletraffic Environments

With the development of advanced sensing technologies and control techniques,the related research about intelligent vehicles has attracted more and more attention from researchers in industry and academia,and the full-autonomous driving has become the goal of many scholars.On the ordinary highway,normal driving behaviors mainly include carfollowing and lane-changing.The lane change maneuvers of human-driven vehicles are closely related to many serious accidents on roads,while the active lane change system of intelligent vehicles which can well prevent the improper operations of human drivers in the lane change process could greatly reduce the probability of road crashes and the driving burdens;and at the same time,improve the efficiency of traffic flow.A complete active lane change system of intelligent vehicles consists of four modules,i.e.,the detection module,the lane change decision-making module,the trajectory planning module,and the trajectory tracking control module.The task of detection module is to sense and detect the status of surrounding environments and the motions of traffic vehicles.The decision-making module uses the information obtained by the detection module to make the specific lane change decision.Then the planning module plans the safe and stable lane change trajectories after receiving the instruction from the decision-making module.When the reference trajectory is determined,the controller is employed to control the vehicle to track the desired trajectory.This project relies on the National Natural Science Foundation Project "Research on control mechanism and evaluation method of new steering-by-wire system based on driver characteristics"(No.: 51575223).When taking the multi-vehicle environment as the research background and the off-line planning as the basic scheme,this paper studies the lane change trajectory planning methods and the tracking control strategy in the active lane change system.To be specific,the trajectory planning strategy includes the normal planning strategy,which enables the vehicle to complete lane change in a safe and stable way,and the trajectory re-planning strategy,which ensures the vehicle to avoid possible crashes in the normal lane change process.The trajectory tracking control strategy includes the lateral displacement tracking control method and the longitudinal speed tracking control method,which can well realized the longitudinal and lateral tracking of the reference trajectory.The key innovation of this paper is that the optimal lane change trajectory planning method is designed considering the multi-vehicle traffic environment,and the lane change trajectory replanning method is designed considering the disturbance of traffic environment.This dissertation mainly includes the following three parts.(1)Design of the normal lane-change strategy for the intelligent vehicle in the multivehicle driving environments.This paper utilizes the quintic polynomial as the basic normal lane change trajectory model when planning trajectories in multi-vehicle environments.The longitudinal planning and the lateral planning are coupled through assuming the longitudinal velocity is invariant in the ground coordinate system.Then,a lane change trajectory cluster is determined by taking the lane change duration as the variable.Next,a stable handling envelope that is derived from the vehicle dynamics model is applied to extract the stable trajectories from the overall trajectory cluster.In a specific traffic scenario,the heuristics collision avoidance algorithm used online could select the collision-free trajectories from the stable trajectory cluster.Finally,the TOPSIS-based multi-objective optimization algorithm is employed to obtain the optimal trajectory from the collision-free trajectory cluster as the reference.(2)The lane change trajectory re-planning strategy in complex driving environmentsConsidering that there may have disturbance during the normal lane change process,and such disturbance may impose the collision risks on the intelligent vehicle which travels along the pre-planned trajectory.In such circumstance,the lane change trajectory replanning strategy,which consists of the sinusoidal attenuation acceleration model-based longitudinal motion planning model and the quintic polynomial-based lateral motion planning model,is designed to eliminate the collision.The longitudinal planning model plans a collision free longitudinal destination and the lane change duration,while the lateral planning model designs a stable trajectory from the vehicle's current position to the destination.According to the lane change trajectory re-planning model,the intelligent vehicle could complete lane change before/after the pre-planned terminal time,or return back to the original lane to well avoid the possible collisions with surrounding objects.(3)The trajectory tracking control strategy and the validation of the proposed schemeTrajectory tracking control consists of two parts: the lateral displacement tracking and longitudinal speed tracking.In this paper,on the basis of the incremental vehicle dynamics model,a lateral displacement tracking controller is designed using the model predictive control(MPC)algorithm.The quadratic programming problem in MPC can be solved by the dual algorithm,which guarantees the feasibility of optimization.In addition,a speed tracking controller using vehicle longitudinal dynamics model is designed to track the reference longitudinal velocity.Finally,simulations and hardware-in-loop experiments are conducted to verify the efficiency and accuracy of the planning strategy and control method.