Research On Local Path Planning Methods Of Self-driving Vehicle On Typical Roads

At present,the problems of driving safety and traffic jam have increased rapidly,with the fast growth of the traffic accident rates and the rising ownership of vehicles.Therefore,the study on self-driving vehicle related technologies has attracted the attention of experts and scholars in various fields all over the world recently.The path planning system is an important subsystem of self-driving vehicles,the target of which is to complete the modeling and path searching in the driving environment.This thesis presented some new improved planning methods to solve the local path planning problems on the basis of related researches.The specific research contents are as follows:Firstly,the development process of international and domestic about self-driving vehicles was presented in detail.The problems encountered in the field of self-driving vehicles were analyzed,with summarized the commonly used path planning algorithms,the advantages and disadvantages of which were also compared.Secondly,the tasks and coordination of relationships in subsystems that make up the local path planning system were analyzed and discussed to clarify the design requirements of the subsystems,aiming at the structure of the local path planning system for self-driving vehicles.The factors that influence the result of local path planning were analyzed and discussed,and the environment model based on probabilistic maps was established according to the characteristics of structured roads.The no-nholonomic constraints of vehicles are studied,and the mathematical model of no-nholonomic properties of self-driving vehicles was also built.Thirdly,the Rapidly-Exploring Random Tree algorithm was studied,based on which,the new improved path planning methods and optimal schemes were proposed,which contains the RRT path planning method based on vehicle kinematics,the path node pruning method based on the distribution of obstacles,the path smooth and optimization method based on B-spline curve,and the path optimization method based on the steering controller model of self-driving vehicle.The simulation results showed that these methods can effectively tackle the local path planning tasks.Fourthly,pre-generated paths of straight road,cross road and U-shaped road were generated by using Dubins curves,based on which,the path planning system and pre-generated paths planning algorithm were also designed.Then,an RRT planning system based on pre-generated path was designed,combined with improved RRT algorithm and path planning method of pre-generated path.Finally,simulation tests were designed to verify the performance of the local path planning system,the results of which showed that the local path planning system can accomplish the planning task of the local path of the self-driving vehicle.