Study On Optimal Trajectory Decision And Control Algorithm Of Automatic Parking System

The auto parking system is an important part of the advanced driver assistance systems.In recent years,it has attracted great attention of domestic and foreign scholars and car manufacturers.The automatic parking system acquires the environmental information through the on-board sensors,after finding a suitable parking space,it assists the driver in controlling the vehicle to complete the parking.The automatic parking system greatly reduces the driver's stress and tension in the parking process,meanwhile avoids the occurrence of traffic accidents.Discovered from the research status of domestic and foreign,existing auto parking solutions are mostly for single parking condition and have poor application.Therefore,this paper proposes an optimal trajectory decision and control algorithm for both parallel parking and vertical parking conditions.The main research content of this paper is as follows.First of all,this paper uses a vehicle-mounted sensor system for the detection of parking spaces and obstacles,which is based on twelve ultrasonic radars.Based on the traditional singleprobe parking space detection algorithm,a novel parking space detection algorithm based on dual-sensor data fusion is proposed.The two ultrasonic radars on the same side of the vehicle detect the parking spaces at the same time,and the accuracy and stability of the parking spaces detection are improved through the fusion of sensor data.The parking space detection test verifies the effectiveness of the dual-sensor data fusion parking space detection algorithm.After detecting the effective parking space,this paper proposes a segmented optimal trajectory decision algorithm by analyzing the safety,feasibility,smoothness and efficiency of the parking trajectory.First of all,for the safety analysis,the parking scene is digitally modeled and the vehicle's travelable state area is obtained.Then,a kinematics model that is independent of speed is established so that the generated trajectory satisfies the constraints of the vehicle kinematics.And the optimal parking reference trajectory is determined based on the optimization of the evaluation function.In order to improve the efficiency of the generated trajectory,this paper formulates some rules for segment planning by imitating the habits of skilled drivers in the process of parallel parking and vertical parking,finally,an optimal parking reference trajectory that satisfies safety,feasibility,smoothness and efficiency at the same time is obtained.The results of Matlab simulation experiments show that the optimal trajectory decision algorithm is feasible and can be applied to various parking conditions at the same time.In order to follow the generated reference trajectory,this paper builds the motion control strategy of the automatic parking system.An EKF-based dead reckoning is used to obtain accurate vehicle location information by filtering the noise signals of the sensors during the process of path following.Then,an optimal tracking controller is used to implement the trajectory tracking by eliminating the current deviation between the vehicle pose and the reference trajectory.In the tracking controller,steering LQR controller and speed fuzzy controller are used to control the steering and speed independently in the parking process to improve the applicability of the controller.Finally,a software-in-the-loop test platform is built on Matlab/Simulink and CarSim,the parking motion control strategy is validated through the joint simulation.Then the real vehicle test platform is built,and a rapid control prototype dSPACE is used as a controller to integrate the algorithm of this paper into a real vehicle for verification.The experimental results of the simulation and real vehicle test show the effectiveness and feasibility of the proposed algorithm.