A Hierarchical Optimization Framework for Parking Maneuver of Automated Vehicle

In this article, the problem of optimize autonomous ground vehicle (AGV) parking trajectory is studied. To speed up the convergence of the optimization algorithm when the problem is established in a constrained environment, a hierarchical optimization framework is proposed. In the first layer, an enhanced A-star algorithm is designed to produce an reference collision-free trajectory. Then, the obtained trajectory is used as the initial guess of the gradient-based optimization algorithm in the second layer. The overall algorithm is tested to plan the optimal parking trajectory, taking account of the AGV mechanical constraints and task-related constraints. Simulation studies were conducted to verify the effectiveness of the proposed scheme. The results show that such a design is able to produce collision-free trajectories for the considered problem.