Research On Trajectory Planning Method For Lane Changing Of Intelligent Vehicle

With the increase of global vehicle ownership,traffic safety issues are becoming increasingly prominent.The causes of traffic accidents are mostly related to driver's Error operations.Therefore,intelligent vehicles have attracted more and more attention and become a research hotspot in the field of vehicle engineering all over the world.Lane change is one of the common behaviors of intelligent vehicles on structured roads,which has an important impact on traffic safety and efficiency.The formation of lane-changing trajectory is a prerequisite for the completion of lane-changing behavior.The performance of lane-changing trajectory determines whether the intelligent vehicle can run efficiently,comfortably and safely in the process of lane-changing.To aim at the problem of lane change for intelligent vehicles on structured roads,this paper studies the lane-change trajectory planning method.Firstly,a lane-changing decision-making strategy based on the behavior recognition of surrounding vehicles is proposed.A trajectory feature identification method based on adaptive hot zone is established.When the trajectory points of the surrounding vehicles get into the corresponding hot zone,the hot zone values are used to represent the observation of the surrounding vehicles by the main vehicle.When the main vehicle changes lanes,the hot zone can be switched.The hidden Markov model is used to model the trajectory,and the calculation is improved by incorporating higher-level conditional statements.The real-time performance of the method is to identify the behavior of surrounding vehicles,and to ensure that the behavior of surrounding vehicles does not threaten the main vehicle lane-changing,the reasons for lane-changing intention are deeply analyzed,and the concept of lane-changing tolerance is proposed to quantify lane-changing intention,and a safe and efficient lane-changing strategy is initially formulated.Secondly,to ensure the feasibility and safety of lane changing for intelligent vehicles on structured roads,the trajectory planning is investigated under multiple operating conditions.For simple barrier-free lane changing,a newly modified cosinetransformation model is proposed to makes the curvature of lane-changing trajectory continuous and solves the front wheel deflection angle constraint problem of intelligent vehicle at the beginning and end point of lane-changing.In dealing with the collision problems that may occur during lane changing,the concept of transit position is introduced,and the method of biquintic polynomial programming is adopted to ensure that the main vehicle can effectively avoid traffic when changing lane,which improves the security and real-time performance of the algorithm.Considering lane changing under curved roads,coordinate transformation is carried out,and the boundary conditions based on initial coordinate system are obtained by polynomial method.The results of lane change research on straight line section are extended to circular arc section.Then,considering the uncertainties caused by artificial boundary conditions in quintic polynomial algorithm,this paper creatively proposes a trajectory planning method based on the fusion of polynomial and particle swarm optimization.The longitudinal displacement of lane change and target speed of intelligent vehicles is taken as variables,based on the off-line collected data of lane change of real vehicles,and constrained by the trajectory performance evaluation index,the two variables are iteratively optimized by using the optimization ability of particle swarm optimization algorithm,to obtain the optimal lane change trajectory.Through the joint simulation of PreScan and MATLAB software,it is verified that the algorithm not only greatly improves the reliability of lane change,but also makes the lane change trajectory efficient and comfortable.Finally,the validity and reliability of the lane-changing trajectory planning method studied in this paper are verified by setting up different scenarios based on the "UJS intelligent driving vehicle" experimental platform.