Automatic Vertical Parking System Research Based On Parking Scene Intelligent Recognition Technology

In recent years, automatic parking technology as an important driving assistance technology has became a hot topic in the field of automotive engineering technology.The semi-automatic parking system is gradually popularized and applied, but which still need the driver to control the speed, gear and brake, the intelligent degree should be improved. In order to avoid the problem of driver was unable to open the door after vertical parking which cause by side storage space is too small. The existing parking system generally take very conservative strategy setting in the identification of available parking(i.e. set the width of the margin is too large parking). Lead to some relatively small spaces effective parking can not be identify, thereby reducing the effective utilization of parking spaces. In order to improve the effective utilization of parking space, the convenience of park and pick up the car, enhance the intelligence and automation level of the parking system, the study take electric vehicle as the carrier, based on parking scene intelligent recognition technology to research the automatic vertical parking system. This paper gives the four key aspects of system solutions-- vertical parking scene recognition, parking path planning, path tracking and control, parking control system design.Firstly, the vertical parking scene recognition methods were studied. Using distance sensors to measure the size of parking spaces, the use of visual sensor to identification the parked vehicle toward on both sides of the empty spaces. Focus on the parked vehicle toward detection. Training vehicle detector using Adaboost algorithm, and then using color model algorithm to detect red taillights. Detected parked vehicles parked toward, thus achieving the purpose of vertical parking scene recognition.Secondly, briefly analysis of the parking process of the vertical parking system, set a target size of parking spaces according to national regulations and the size of the test vehicle. Analysis the vehicle kinematics and motion constraints during the parking,set up the vehicle kinematics model and analysis parking space constraints. Modeling analysis the common vertical parking scene, draw the actual path of the vehicle motion control when the driver skilled vertical parking, planning the idealparking path of movement.Thirdly, take electric vehicle as the carrier to analysis the vehicles’ movement law during parking, proposed a non-smooth control theory based on the path tracking control method. In order to accurately reflect the true situation of vehicle motion,designed automatic parking path tracking controller based on the vehicle dynamics model, control vehicle tracking planning routes, improve parking accuracy, and the simulation results show the controller’s effect of path tracking.Finally, the automatic vertical parking controller is developed based on the carrier of electric vehicle. The parking system mainly includes parking control subsystem,power steering and speed control module. Wherein the parking control subsystem consists of four modules: environment perception, sensor signal processing, voice and power supplies. The hardware circuit and the system program of the control system are designed, the anti-interference performance of the software and hardware are considered, and the loading test of the parking controller is carried out. The results show the controller has better performance on the vertical parking scene recognition,parking path selection and parking precision.