Parking Slots Intelligent Detection And Parking Path Tracking Control Of Automatic Parking System

As one of the most important components of intelligent vehicle and driverless vehicle,automatic parking system has broad application prospects,and it has been a research hot-spot in the field of intelligent vehicle technology.At present,most of the parking assist systems are semi-automatic parking assistant systems.The semi-automatic parking controller only controls the steering wheel angle and it requires the driver to control the speed.It is difficult to control the speed and steering angle coordinately due to the intervention of the driver,therefore,the result of parking stability and controllability is poor.In addition,the existing parking slot detection methods mainly relies on a single kind of sensor,which have limited detection accuracy and can only recognize particular types of parking slots.To solve these problems,this paper focus on the study of slot intelligent recognition and parking path tracking control algorithm of automatic parking system based on data fusion of image and ultrasonic sensors.Particularly,the environment sensing sensors,parking slot intelligent detection,parking path planning and tracking control algorithm,and automatic parking system architecture are studied.Firstly,the function principle and calibration method of environmental sensing sensors are analyzed.The functional characteristics of various environment perceiving sensors are analyzed.In addition,the working principles of the ultrasonic sensors and fish-eye lens used in the automatic parking system are introduced in detail,and the installation and calibration methods of the surround view camera are studied.Secondly,a high-precision parking slots detection method based on multi-sensor data fusion is proposed.Various types of slots in China are comprehensively analyzed,on this basis,the horizontal and vertical parking slots which are the most common in cities are selected for modeling.Accurate width and depth information of parking slots are obtained by ultrasonic and image sensors.Vehicle position and head angle are estimated via speed and front wheel angle information.The world coordinate system of parking slots is established to lay the foundation for parking path planning.A parking slot detection method based on fuzzy logic is proposed.The environmental information is obtained for feature level fusion by ultrasonic sensors and cameras.The width,depth of the slot and the obstacle information in the slot are extracted as inputs of the fuzzy logic,and the type of parking slot is judged by the fuzzy logic.A fuzzy classification model of the type of parking slots is established and verified via Simulink.The simulation results prove the effectiveness of the proposed fuzzy classification method.Then,parking path planning and path tracking control algorithm are studied.The parking path is designed for parallel parking as an example after the geometric relationship between vehicle and parking slots is determined.Besides,the vehicle kinematics model is established,and on this basis,the parking path tracking controller based on model predictive control with soft constraints is proposed.The simulation of parking path tracking control under low speed condition based on Simulink/Carsim is carried out.The simulation results show that the proposed model predictive parking controller achieves great coincidence between the actual path and the reference path in the whole process of path tracking and the final parking effect of the vehicle is good.Compared with the PID controller and open-loop controller,the MPC controller has obvious advantages in all aspects.Finally,the architecture of automatic parking system is designed and developed,and the experiments of real vehicle are carried out.The advantages and disadvantages of the two system architectures with APA(Automatic parking assistance)and SVM(Surround view management)as the control core are analyzed,and the system architecture with APA as the vehicle control core is determined.On this basis,the data interaction method of the APA and the SVM is designed.The parking slots detection experiments under various experimental conditions are carried out.The results show that the parking slots intelligent detection method proposed in this paper can effectively detect various types of parking slots and effectively improve the intelligent level of the automatic parking system.The parking path tracking control experiment based on MPC algorithm is carried out.The results show that the parking path tracking effect based on MPC controller is better than that based on PID controllerand open-loop controller.