Research On Automatic Parking Path Planning And Parking Pose Estimation Algorithm

As an important part of advanced driver assistance systems,the automatic parking system can help the driver quickly and safely complete the parking operation,both to reduce the level of tension in the process of driver parking,improve the driving comfort,but also to a certain extent,reduce the risk of accidents in the process of parking.Starting from the kinematics model of automatic parking,this paper discusses parking path planning,parking position estimation,parking system construction and real vehicle test.Firstly,the development status of automatic parking system at home and abroad is briefly introduced,and the working principle of automatic parking system is analyzed.The kinematics model is built,and the model is simulated and analyzed.Then,based on the principle of ultrasonic detection and the trajectory of key points,the different path planning and collision constraints under different types of parking spaces are analyzed.Choose "arc-line-arc" parallel parking,backward vertical parking,"arc-line" diagonal parking as the path planning control strategy of the automatic parking system.Take the diagonal parking path planning as an example,the fuzzy controller and path optimization are designed.Secondly,based on the improved unscented kalman filter algorithm,the parking position and attitude estimator is designed.Combined with the state equation and observation equation,the optimal value is predicted,and the position and posture of vehicle in the parking process are estimated,and the accuracy of the algorithm is verified by simulation.Finally,the automatic parking program,build the test platform,the ultrasonic ranging test,recognition test of parallel parking spaces,parking ramps real car test and real vehicle test.The results show that the automatic parking system has a high success rate,which validates the practicability of path planning and path tracking.