Research On Path Simulation And Motion Control For Automatic Parking

The parking space is gradually reduced because of the increase of city carownership and the environment around the parking space also become morecomplex. Parking is a problem for inexperienced drivers and the security of drivingcan be reduced. With the rapid development of automobile electronic technology,the study of automatic parking system has become more in-depth. The system canreduce the parking difficulty, improve security. So automatic parking system is animportant part of the automotive safety systems.Based on a SUV car model, a kinematics model of the vehicle at low speed isbuilt. The possible collision points during the parallel parking and vertical parkingprocess are analyzed by reverse path planning analysis. The minimum parkingspace and the initial point of the parking of one completed parking can be got bymodel calculation. The real car parking data are collected by test. Different datasamples are used in neural network which has been optimized by particle swarmoptimization in order to avoid the analysis of some constraint relations, such assafety distance. So the parking path generated can be better used in actual parkingprocess. The simulation results and the real car test show that a higher success rateby using the path generated by the above method for parking. Based on fuzzycontrol and neural network theory, parking fuzzy controller is built. Through theparking simulation in different size of parking space, the car can finish the parkingin smaller space and the path is more reasonable and safe.According to the function demand of automatic parking system, the hardwareof automatic parking system is designed. Freescale MC9S12XDT512chip isselected as the system main control chip. The environment around the parkingspace is detected by ultrasonic. The fuzzy controller output corner that the steeringwheel required. The rotation angle of the steering motor is controlled in the way ofPWM.