Research On Intelligent Parking Path Planning And Tracking Method Based On Multi-sensor Information Fusion

As an important part of the advanced driving assistance system,the automatic parking system can help the driver to complete the parking operation quickly and safely.Aiming at the existing smart parking technology,due to the problem of parking space perception that irregular parking spaces cannot be identified and the path planning does not meet the actual needs,and then affects the actual vehicle path tracking reference effect,this paper designs a smart parking path planning based on multi-sensor information fusion And tracking methods.Based on the previous work,this paper discusses the parking environment perception,parking space identification,and path planning and tracking methods based on the kinematics model of automatic parking.First,the environmental awareness technology of intelligent parking is studied.The principle and functions of ultrasonic radar and lidar are introduced respectively.In view of the shortcomings of traditional environmental perception that the perception information is not rich enough,the detection distance of ultrasonic radar and the detection space of lidar and the surrounding vehicle space are proposed.The distributed environment perception method extracts the relevant parameter information of the parking space and the posture information of the obstacle vehicle.And carry on the input analysis to the fusion sensor data needed for parking.Then,for the problem that traditional parking space recognition technology cannot recognize irregular parking spaces,a fuzzy inference parking space type detection and recognition method based on multi-sensor information fusion is proposed.By establishing a parking space model,a multi-input multi-rule fuzzy inference parking space detection and type recognition is designed based on multi-sensor information fusion.Based on the joint simulation of Prescan and MATLAB software,the results show that the method can effectively and accurately detect regular parking spaces and irregular parking spaces,and can adapt to the needs of multiple parking space type recognition in actual parking scenarios.Then,the vehicle kinematics model and Ackerman steering are used as the theoretical basis to model the vehicle motion.Firstly,the motion constraints and obstacle avoidance constraints during parking are analyzed.We studied and analyzed the corresponding path planning methods from parallel parking spaces,vertical parking spaces and oblique parking spaces respectively,and realized parallel parking path planning based on arc-arc fitting and vertical parking path planning based on back-forward-backward and oblique parking path planning based on arc-straight line.The path planning simulation experiment of various parking spaces was completed according to the parking space identification,which verified the effectiveness of the designed path planning method.Finally,the vehicle path tracking control is designed.We improve and design the sliding mode control based on the shortcomings of large errors and instability,so as to realize the sliding mode tracking control based on fuzzy control gain.For this method,the path tracking simulation is performed based on the simulation test platform to verify the effectiveness of the sliding mode path tracking strategy.