Optimal Design And Research On Predictive Control Of Automatic Parking System

In response to the current poor user experience and low penetration rate of fully automatic parking systems on the market,in response to the development direction of the "new four modernizations" of automobiles proposed by the country,the current research status of key technologies for automatic parking systems is analyzed and studied.This article uses a certain type of car as a carrier to design a user-oriented fully automatic parking system,including parking space detection,parking trajectory planning,and motion control of the automatic parking system,and focuses on the body positioning technology and trajectory during parking.The tracking control algorithm has been researched to ensure the stability of automatic parking,improve the success rate of parking,and improve the user experience.First of all,after conducting a full investigation and analysis of the background and key technology research status of the automatic parking system,this thesis conducts the hardware solution and system software architecture of the entire automatic parking system according to the specific functions and actual requirements of the automatic parking system.Design development.According to the functions of different modules,suitable equipment is selected and loaded on the actual vehicle,which is used as the experimental platform of the actual vehicle.Secondly,in the parking space detection module,it is the first to propose a data fusion scheme using laser rangefinder and ultrasonic radar for effective parking space recognition,which can adapt to more complex parking environments and improve the effective parking space recognition rate.The designed parking space recognition algorithm makes the system suitable for more parking scenarios.Real vehicle experiments verify the system's stable parking space recognition effect.For the trajectory planning of parking,this thesis chooses the more common parking trajectory that conforms to the driver's parking habits.The three-segment parking trajectory is used for trajectory planning for parallel and vertical parking scenarios.Matlab simulation experiments show that the designed parking trajectory has high feasibility,strong applicability and obvious advantages.Thirdly,for the body positioning technology which is one of the two core technologies of motion control for autonomous driving,this thesis considers the characteristics of low speed,small range and low equipment cost of the automatic parking scene,and uses the track push algorithm for real-time body positioning.According to the problems of inaccurate vehicle speed and steering feedback in real vehicle experiments,on the basis of traditional path estimation,a wheel speed type path estimation algorithm is proposed and improved to a fusion type path estimation algorithm.Carsim and Matlab joint simulation experiments and Real vehicle experiments have verified the practicability and superiority of the algorithm.Finally,a layered control strategy is used to design the motion controller of the parking system.After obtaining valid real-time body position information,model predictive control is selected as the algorithm of the upper controller,and the lower controller uses PID control.In this thesis,the currently used preview control algorithm is used as a comparative experimental algorithm,and joint simulation experiments show that the model predictive control algorithm is significantly better than the preview control.The actual vehicle experiments also prove that the control effect based on rolling optimization is more stable,more robust,and the parking success rate of the automatic parking system is higher.