| In recent years,with the rapid development of the national economy and the automobile industry,the continuous increase in the number of domestic automobiles and insufficient construction of urban parking lots have led to serious parking difficulties.Insufficient parking space,narrow parking spaces,and difficult parking operations in urban areas have become a common problem in people’s daily travel.Autonomous valet parking technology is a low-speed unmanned driving technology designed for parking lot scenarios,which can effectively help people solve parking difficulties and improve vehicle travel efficiency.Autonomous valet parking technology mainly includes perception,positioning,planning,and control.This study focuses on the planning and control aspects of autonomous valet parking technology,and constructs the motion planning and motion control system in the autonomous valet parking system,including global path planning,local obstacle avoidance path planning,parking path planning,and vehicle path tracking motion control.The main contents of this study are as follows:(1)This article proposes a novel autonomous parking driving path planning method that combines the directed Hybrid A* global path planning algorithm with the path discretization and optimization-based local obstacle avoidance method to generate a highquality collision-free path for vehicles from the parking lot entrance to the vicinity of the parking space.In terms of global path planning,a directed Hybrid A* global path planning algorithm based on the generalized Voronoi diagram is proposed,which can accurately and effectively generate a collision-free route from the parking lot entrance to the parking starting point while avoiding redundant searches.Compared with the traditional Hybrid A* algorithm,this algorithm improves efficiency by 30%.In terms of local obstacle avoidance path planning,a directed quasi-uniform B-spline path sequence discretization optimization path planning method is proposed,which selects the optimal path by constructing an evaluation function based on the safety,smoothness,path length,and end point position during vehicle driving to achieve high-quality local obstacle avoidance path planning.(2)In this study,we investigated single-step or multi-step parking for three types of parking spaces: parallel,perpendicular,and diagonal.We established a complete parking motion planning framework and generated corresponding parking paths using collision constraints obtained from inverse kinematics and a method that combines straight lines and circular arcs.We established collision constraint equations for the parking process using inverse kinematics and used numerical computations to generate suitable parking paths.We used B-spline curves to smooth and interpolate the parts of the path with curvature discontinuities to ensure that the path is easy to follow.Simulation results show that our method can generate smooth parking paths for the three types of parking spaces.(3)For autonomous parking control,this study derives the model predictive control theory based on the vehicle kinematic model and designs a corresponding autonomous parking motion controller.Using the joint simulation method of Carsim and Simulink,this method was compared with other control methods in terms of tracking the parking path,and the results showed that the autonomous parking motion controller performs better in tracking the parking path.(4)A Matlab/Simulink co-simulation platform was constructed to verify the autonomous parking motion planning and control algorithms using a 3D virtual environment of the parking lot created by Matlab 3D engine combined with Carsim vehicle model.Finally,the algorithm was validated through real-world experiments in the vertical parking scenario.During the experiment,the vehicle completed the autonomous parking task in 79 seconds,with a lateral error of less than 0.25 m,and a lateral error of0.035 m upon completion of the parking.The results demonstrate the effectiveness of the algorithm in practical scenarios. |
PDF Full Text Request