Study On Trajectory Planning Algorithm Of Automated Valet Parking System Under Urban Road

The emergence of automobiles has changed the way of travel of mankind and greatly improved the efficiency of travel,which plays a very important role in modern society.But with the increasing number of cars in cities,it has also brought a series of problems such as road congestion,parking difficulties,and frequent traffic accidents.Relying on advanced environmental perception systems,decision-making planning units and executive agencies,smart cars can effectively alleviate the above problems.As the level 4 technology of automatic driving,the automated valet parking system is the key technology to solve the parking problem.The vehicle can find and drive to the nearby parking space to park or park out of the parking space and drive to the target position.This paper researches the trajectory planning algorithm of automated valet parking system,and proposes a trajectory planning algorithm for long-range automated valet parking system that can adapt to the complex environment of urban roads.The main research contents are as follows:(1)Establishment of automated valet parking system architecture and vehicle model.This paper proposes a long-distance automated valet parking system that is not limited to parking lots,and divides it into local trajectory planning under urban roads and parallel parking planning for parking spaces.The coordinate system used in the trajectory planning problem and the conversion relationship between the coordinate systems are studied.In order to reduce the complexity of the planning problem,the Frenet coordinate system is introduced into the local trajectory planning algorithm for urban roads.The Cartesian coordinate system is used in the parallel parking planning algorithm.The forward and backward parking modes are compared and analyzed to determine the parking mode used in this paper.Finally,the vehicle kinematics model used in the algorithm is established,which lays the foundation for the research of the trajectory planning algorithm in this paper.(2)Research on local trajectory planning algorithms under urban roads.For urban complex road environment,the parking trajectory planning algorithm with a hierarchical framework is used to divide it into two sub-problems of path planning and speed planning,thereby reducing the computational dimension of the research problem.In order to ensure the trajectory quality and real-time planning at the same time,a hybrid search method is proposed by combining the method of graph search and numerical optimization.For path planning,a state lattice method is used to generate the original path,which is then optimized using a non-linear planning method.For speed planning,an improved A * search method is proposed in combination with the mechanical constraints of the vehicle.Based on this,an optimized speed curve is obtained through quadratic planning.(3)Research on parallel parking trajectory planning algorithm.Firstly,the collision points existing in the parking environment are analyzed and researched.The minimum parking space size that satisfies the safe parking of the vehicle in the parking space is determined by the collision points and the extreme steering conditions of the vehicle in the parking space.Secondly,the parking planning methods of two tangent arcs and arc-line-arc are studied in depth,and the problems existing in the two parking planning methods are analyzed.Finally,in order to solve the problem of turning in place during parking,an improved scheme combining an arc curve and a clothoid curve was proposed.The parking path was divided into six segments and the expression equation of each path was solved.(4)Simulation experiment verification of trajectory planning algorithm for automated valet parking system.Set up simulation environments for local mixed trajectory planning algorithms for urban roads and parallel parking trajectory planning algorithms,add moving obstacle vehicles to the urban road trajectory planning simulation experiments,and design three driving scenarios for lane keeping,lane change,and overtaking for algorithm performance testing.In the parallel parking trajectory planning simulation experiment,the effects of different types of parking planning algorithms are compared.Finally,the clothoid curve algorithm is selected and the influence of parking space size,starting position,and parking speed on the algorithm is analyzed.The simulation results verify the feasibility and effectiveness of the trajectory planning algorithm proposed in this paper.