Research On Path Planning And Tracking Control Of Automatic Parking

The rapid economic development of modern society improves people’s living standards,which has aroused people’s demand of the convenient transportation,resulting in the continuing rise of existing vehicle ownership in the domestic market.Accordingly,traffic conditions got more congested.Moreover,concerns with vehicle management and traffic safety progressively grew in importance as obstacles to urban development.Domestic and international automakers performed extensive research on vehicle assisted driving technologies to fulfill market demands.The issue was considerably being resolved thanks to the development of cutting-edge aided driving technologies with automatic parking systems gained popularity as a solution.Parking space detection,path planning,and tracking control were used to successfully meet the objective of accurately parking automobiles in available parking spaces.The market demand for this technology was both universal and practical.Therefore,it emerged as one of the primary research topics for a number of academic institutions and automakers.The three crucial links in the autonomous parking process were systematically examined in this paper.The following are the research and innovation points:(1)In order to detect the actual state of the parking space,it was necessary to use remote ultrasonic sensors on both sides of the vehicle body.After analyzing the motion parameters under low-speed parking conditions,a corresponding vehicle kinematics model was created based on pertinent factors such as the vehicle’s length,width,front overhang,rear overhang,and wheelbase.Simulation tests demonstrated that the accuracy and dependability of the model satisfied the experimental requirements when the vehicle’s initial speed and equivalent rotation angle were fixed.(2)A thorough analysis of the parallel parking scenario was conducted with the aforementioned model as a reference model,and the reverse derivation was then carried out to finish the computation of the minimum parking space.A path planning method based on sine function and constant velocity offset superimposed curve was chosen.Besides,a comparison simulation experiment was conducted with the conventional double arc tangent path planning scheme to address the restrictions of how to avoid obstacles and vehicle kinematics.The experimental findings showed that the path planning strategy proposed in this study may meet a variety of vehicle obstacle avoidance and steering limitations during ensuring smooth paths.In addition,the strategy made the planning easier,drastically lowering the amount of calculations required,and enabling the vehicle tracking management much more feasible.(3)A better model predictive control was applied to create a controller.The objective function was identified with vehicle speed and front wheel angle as control variables.A feasibility experiment for trajectory tracking was conducted during parking,combined with the steering angle,steering angle increment,vehicle acceleration,collision limits,and other factors.For comparative tests,a controller built on a linear quadratic regulator was also adopted.The experimental findings were justified to the fact that the MPC control-based algorithm outperforms the LQR method in three tracking-related domains: tracking trajectory effect,heading angle effect and longitudinal position errors.(4)More evidence was provided to support both the effectiveness of the built controller module and the proposed trajectory algorithm.With Simulink/CarSim software,a collaborative simulation platform for autonomous parking was built.It was possible to collect the relevant data and curves.The reasonableness of the proposed path was confirmed and examined with these simulation data,and the practicality,stability,and error values of putting the controller module into use were all carefully examined.Joint simulation studies indicated that the MPC method could follow the parking path in this research more correctly and fairly.