Research On Route Planning Method Of Driverless Vehicle In Structured Road Environment

Nowadays,as the world’s growing population,urban traffic congestion and traffic accidents are becoming more and more serious,it is urgent to solve the traffic problems.Autonomous driving vehicle technology is regarded as the fundamental way to solve traffic problems and one of the leading topics of academic research in the world.Its technologies include vehicle engineering,cognitive science,artificial intelligence,robotics and other fields,involving a large number of theoretical methods and breakthroughs in key technologies.And route planning is the precondition to ensure the safe and accurate completion of driving tasks of driverless cars,which is of great significance to the development of modern urban traffic.Path planning can be divided into global path planning and local path planning.Firstly,this paper introduces and analyzes the research status of driverless vehicles at home and abroad and relevant algorithms of driverless vehicle path planning,so as to understand the basic methods of driverless vehicle path planning;Then the paper introduces the characteristics of the structured road and the architecture of today’s driverless cars and the functions of each sub-module and the correlation between them,and the design criteria of global path planning and local path planning algorithms are defined by combining the road traffic rules and the constraints to be considered in motion planning;Then aim at the problems of real-time and traffic rule constraints encountered by today’s self-driving cars in static global planning and sports obstacle avoidance planning under structured road environment and the system’s design requirements for the safety,stability,smoothness of the planning results.This paper has conducted in-depth research on global path planning and local path planning.The specific research content includes the following aspects:1)Aiming at the problem that the traditional global path planning algorithm handles a large number of road nodes with poor real-time performance,the global path planning method of driverless cars is studied,and an improved A~* algorithm based on binary tree and two-way search,namely BB-A~* algorithm,was proposed,and the effectiveness and better real-time performance of the algorithm were verified by simulation experiments and global planning experiments based on GIS data of real map of Beijing.2)In order to solve the problem that the existing safety distance model is inaccurate in calculating the braking distance of the vehicle and cannot adapt to various road conditions such as rain and snow,the driving safety distance model of driverless cars is studied.The pressure change process of the vehicle brake mechanical structure during braking is analyzed,and a more accurate mathematical model of the driving safety distance is established.The simulation results verify the accuracy of the braking distance model and the feasibility of the driving safety distance model.3)Local path planning method for driverless cars is studied.An improved RT-RRT* algorithm(SP-RT-RRT* algorithm)based on the urban lane scatter points strategy is proposed,which improves the problem of the random sampling point location of the traditional algorithm resulting in the unstable real-time search and the obtained path not meeting the traffic rule constraints.Then the initial path obtained by the algorithm is smoothed by the five-time Bessel curve,so that the vehicle can drive safely and autonomously in the dynamic environment.The feasibility and superiority of the local path planning method in real time and satisfying traffic rules are verified by simulation experiments.Finally,the research results of this paper are summarized,and the prospect of further research is made.