Optimal Automated Valet Parking Trajectory Planning Based On Safe Tunnels

The main function of autonomous parking trajectory planning is to plan the parking trajectory according to the current parking environment,so that the vehicle can complete the parking task autonomously.In addition to the proper environmental simplification for vehicle trajectory planning,accurate kinematic model and efficient solution methods are also key factors in determining trajectory quality.The paper takes intelligent networked vehicles as the research object and proposed the optimal autonomous parking trajectory planning method based on safety tunnel.The main contents are as follows:(1)A high-precision parking lot topology map containing road topology network and parking space parameters is designed,which simplifies the complex parking lot environment into multiple topology nodes.The topological node is used as the input of the RANSAC algorithm.The topological nodes are clustered,and the road model between the nodes after clustering is obtained.The iterative expansion method is used to expand the road model into a safe tunnel.Finally,according to the specific parking task,the Dijkstra algorithm is used to select the appropriate safety tunnel for parking.Simulation results show that this method can simplify the complex parking lot environment into multiple rectangular tunnels.(2)A trajectory planning modeling method based on optimal control is proposed.Aiming at the simplified parking environment,an optimal control model for segmented obstacle avoidance is proposed.Taking the coordinates,swing angle,steering angle and speed of the vehicle as the state variables.Taking acceleration and steering angular velocity as the control variables.Taking parking time as the target functions,the parking trajectory planning problem is transformed into an optimization solution problem by establishing the vehicle kinematics equation constraint,the state/control variable interval constraint,the edge value constraint and the path constraint.(3)Optimization solution strategy based on path coordination is proposed.In order to improve the stability and efficiency of the optimization solution process,the initial parking trajectory is generated by using the path coordination strategy.The forward and reverse segments of the global path and the parking path are planned according to the map information.The global path is smoothed based on the Bézier curve.Then,through the path coordination strategy,the smoothed global path and parking path are fused.Finally,according to the trapezoidal speed planning,the speed is planned for the initial path.The initial parking trajectory is obtained.The optimal solution is obtained by substituting it into the optimization model.Simulation shows that the initial trajectory generated by the proposed method can effectively speed up the solution time of the optimization model.4)Experimental research based on intelligent networked vehicle platform is conducted.The environment simplification method and the trajectory generation method are simulated and experimentally verified.Experimental results show that the designed environmental simplification method can effectively reduce the complexity of the parking environment.The proposed trajectory planning modeling method can calculate a reasonable parking trajectory and make the vehicle successfully parked in the parking space.