Hierarchical Piecewise-Trajectory Planning Framework for Autonomous Ground Vehicles Considering Motion Limitation and Energy Consumption

Planning trajectories and trajectory tracking are significant and fundamental tasks for Lagrange-based autonomous ground vehicles. In this paper, a novel unified framework integrating path planning and trajectory tracking is proposed based on deep reinforcement learning for Lagrange-based autonomous ground vehicles considering motion limitation and energy consumption, namely, hierarchical piecewise-trajectory planning (HPP) framework. The framework consists of three layers, namely the path planning layer, the trajectory planning layer, and the local control layer. Firstly, the path planning layer enables the vehicle to find a discrete path from its initial position to its target position. Afterward, the trajectory planning layer ensures that discrete trajectory points are transformed into continuous trajectory functions based on the polynomial curve interpolation method. The adaptive asymptotic acceleration planning algorithm is proposed to satisfy the limitations of maximum velocity and acceleration for vehicles. Finally, the trajectory tracking control algorithm and poweroff trigger mechanism are developed to achieve the following two goals in the local control layer: 1) regulating the vehicle to follow its continuous trajectory curve, 2) switching off the power to save energy when its instantaneous kinetic energy is adequate to supply the energy consumption. Numerous simulation results show that our framework enables autonomous ground vehicles to accomplish integrated path planning and trajectory tracking tasks with the presence of motion limitation. Two extra examples are presented to demonstrate that our method is generalizable in terms of energy savings compared to existing optimization-based methods.