Research On Decision-Making And Path Planning Of Automated Valet Parking System

With the increase of private car ownership year by year,the problem of parking difficulty has become increasingly prominent.The research on intelligent transportation systems such as autonomous vehicles has gradually attracted the attention of all parties.In view of this situation,this thesis carried out a research on the decision-making and path planning methods of automated valet parking(AVP)system.Automated valet parking system,as a milestone of L4 autonomous driving,aims to achieve completely autonomous driving in parking lot.Among them,the decisionmaking and path planning module,as an indispensable and important module in autonomous vehicles,is also a research hotspot in the field of autonomous driving.This thesis mainly studies the global path planning method and behavioral decision method in the AVP system.On this basis,the global path planning module based on the hybrid A* algorithm and the behavioral decision module based on the behavior tree method are established.The main contents are as follows:(1)Developed the AVP system global path planning overall plan.In this thesis,the principle flow of hybrid A* algorithm,node extension mode and heuristic method used in global path planning are analyzed in detail.In addition to expatiating the technical details of the algorithm,the cost estimation function including extended cost and heuristic function was determined.(2)The overall scheme of global path planning is used to realize the task of global path planning in the AVP system.The program flow of the planning method is given,and the simulation test is carried out based on the path planning program.The simulation compares the global path planning results using traditional A* algorithm and hybrid A* algorithm,and the planning results using different heuristic.(3)Established the behavior decision-making model of the AVP system.Based on the analysis of driving environment and driving behavior,an behavior decision-making model of AVP system under the constraint of basic traffic laws and driving experience is established for the whole parking lot scene.(4)The behavioral decision-making method based on the behavioral tree is used to realize the behavioral decision-making task in the AVP system.The program based on the behavior tree method was used to make a real car test in the real parking lot environment,and the output decision results of the behavior decision-making method in four typical scenarios were listed in detail,and the test results were analyzed to verify the rationality and enforceability of the behavior decision-making model.