Research On Path Planning And Tracking Control Of Parallel Parking System

The increasing number of vehicles has led to an increasing shortage of parking Spaces.Automatic parking system can ease traffic difficulties,improve traffic efficiency,and reduce parking accidents.At present,the algorithm and model of automatic parking system in path planning and tracking control are too complicated.In this paper,the automatic parking system is improved from two aspects: path planning and tracking control.The main contents are as follows :1.Kinematic model of parallel parking vehicle: Based on Ackermann steering principle,the vehicle kinematic model is established in Cartesian coordinate system and Frenet coordinate system respectively,and the mutual conversion relationship of related parameters of vehicles in the two coordinate systems is deduced.2.Parallel parking path planning: in Cartesian coordinate system,the initial path is planned with double circular parking trajectory,and the path is smoothed with a quintuplical polynomial,and the parking path is planned according to parking parameters,body size,curvature constraints and obstacle avoidance constraints in the parking process;In order to reduce the complexity of the planning problem,Frenet coordinate system is introduced into the parallel parking planning algorithm of vehicles.Under Frenet coordinate system,parking motion is decoupled into horizontal and longitudinal motion for trajectory planning.After selecting the optimal trajectory cluster with relevant constraints,the horizontal and longitudinal motion planning trajectory is coupled according to the corresponding time relationship.The final parking trajectory cluster is obtained,parking is performed according to the planned trajectory cluster,and the optimal trajectory is screened with relevant constraints.The planned path is simulated and verified in Matlab.3.Parallel parking motion control and path tracking: In order to reduce the difficulty of parking motion control analysis,parallel parking motion is decoupage into horizontal and longitudinal motion for tracking control respectively.Based on model predictive control algorithm,the longitudinal control is studied theoretically.For lateral control,pure tracking control,linear quadratic regulation control and model predictive control are compared.The model predictive control has the best tracking effect,which can control and adjust the lateral parking movement.A vehicle motion controller was designed based on the Model Predictive Control/MPC,and the MPC motion controller model is built in Simulink.4.Parallel parking co-simulation experiment: In order to verify the effect of the parking path planning algorithm and the parking tracking controller proposed in this paper,the Carsim and Simulink co-simulation models were built to simulate and verify the functions of the designed path planning algorithm and model prediction controller.The results show that the proposed planning method can plan a reasonable parking path.And the designed motion controller can accurately track the planned parking path to ensure that automatic parking can be completed safely.