Research On Path Planning And Tracking Control Algorithm Of Automatic Parking System

With the increase of car ownership,urban road traffic pressure is increasing day by day and the driving and parking environment of drivers has become increasingly complex^[1],in recent years,Autonomous driving technology has become a research hotspot of domestic and foreign scholars in the automotive industry,With the rise of cutting-edge technologies such as artificial intelligence and big data,autonomous driving has also ushered in a phase of rapid development,as an important part of automatic parking technology,it has been produced in various types of vehicles and has been favored by many users.Automatic parking technology can effectively solve a series of problems such as complex parking environment,small parking space,and difficult parking and it can also improve the safety,convenience and comfort of parking process.Parking space detection and recognition,path planning and path tracking technology are three important components of automatic parking system,in this paper,it takes a certain vehicle type as the research object and studies and analyzes the three important components of automatic parking system.Firstly,in terms of the detection and recognition of parking Spaces,aiming at the shortcomings of the detection and recognition of non-standard parking spaces without parking lines（mainly inclined parking Spaces）at the present stage,vehicle sensors are used to scan and detect the parking environment,and a plane parking model which integrates ultrasonic sensor,in this paper,wheel speed sensor and vision sensor information to identify parking parameters and vehicle attitude on both sides is established,the intelligent recognition system for parallel parking space,vertical parking space and non-standard parking space is constructed based on the fuzzy reasoning method,it realizes the effective identification of multiple parking Spaces in complex parking environment.Secondly,the vehicle kinematics model was established,and the parking motion equation at low speed is calculated.in this paper,the two most commonly used parking methods such as parallel parking and vertical parking are conducted deeply research in terms of path planning,for parallel parking,under the premise of satisfying the constraints of physical conditions and obstacle avoidance,a multi-section parking path with the combination of circular curve and tangent line is preliminarily planned based on the driver’s habits.considering the comfort and comfort of the parking process,theβ-spline curve was used to smooth the parking path,which obtains curvature continuous and gently varying parking paths.For vertical parking,it is mainly divided into two parking conditions,When the initial position of the vehicle is relatively far from the parking space and the lateral distance is large,a circular arc parking path is adopted;When the initial position of the vehicle is close to the parking space and the lateral distance is small,the combination of two arcs and a straight line is adopted to preliminarily plan the parking path,in order to avoid the abrupt change of curvature at the junction point of arc and straight path,the clothoid curve is used to smooth the path,and finally the parking path with continuous curvature change and meeting the obstacle avoidance constraints is obtained.Then,in the aspect of path tracking control,the layered control strategy is adopted in order to improve the tracking accuracy,the upper controller uses the MPC control algorithm to control the lateral deviation between the vehicle and the reference trajectory,and the lower controller uses the fuzzy PID control algorithm to control the longitudinal velocity.In this paper,it uses the common preview control algorithm to make a comparison test.Finally,it uses Carsim and Matlab/Simulink to carry out the co-simulation test,which verifies the effectiveness and superiority of the path tracking controller based on MPC control algorithm.