Local Path Planning And Tracking Methods Of Intelligent Vehicle In Complex Environment

Local path planning and path tracking are two important technologies to autonomous vehicle.The existing methods can not properly solve the local path planning and tracking problem of intelligent vehicle in complex environments with many irregular and random obstacles.Complex environment will increase the complexity of path planning,reduce the response speed ofpath planning system,and then threaten driving safety.Meanwhile,path tracking will often fail in the complex environment because of the combined action of the path tracking error and obstacles.Therefore,the research on local path planning and path tracking of intelligent vehiclein complex environment has theoretical significance and practical value.The path planning method proposed in this thesis consists of two steps: preliminarypath searching and post-processing.CE-RRT algorithm based on RRT algorithm framework is used for preliminary path searching.Combined with reasonable distancefunction,heuristic extension strategy and dynamic neighbor region,CE-RRT algorithmcan get a more stable preliminarypath with shorter time.During post-processing step,The reference path is obtained by removing redundant sections from the preliminarypath.Then,the predicted trajectory is obtainedthroughsimulatingwith proposed vehicle model and path tracking model.The simulation results verify the effectiveness of the path planning method.Aiming at the problem that the path tracking failure caused by the combined action of the path tracking error and the obstacle,a two-state path tracking model is propose in this thesis.The model consists of two parts,namely longitudinal part and lateral part.Each part have two states: forward tracking state and reverseadjusting state.When the longitudinalpart works under the forward tracking state,the command speed is decided according to the predicted trajectory.Otherwise,a constant command speed is used for reverseadjusting.In the lateral part of path tracking model,the classical Pure-Pursuit model is used for forward tracking state and a novel control law proposed in this thesis is used for reverseadjusting state.The simulation results show that the proposed path tracking method can overcome path tracking failure with a series of maneuvers and make the vehicle track the reference path continuously.According to the methods mentioned in this thesis,a path planning and tracking system is implemented with the intelligent vehicle experimental platform of our school.The effectiveness of the proposed method and the developed system is verified by real vehicle experiments.