Algorithm for planning energy-efficient local trajectories of ground-based robotics, considering the global route

Abstract

In this paper, an algorithm is presented that provides dynamic planning of a local trajectory of movement based on the data from the 3D lidar, considering a predetermined global route on rough terrain. The proposed local planning algorithm considers the features of the LRLHD-A* global energy-efficient trajectory planning algorithm previously developed by the authors. The novelty of the proposed solution lies in the algorithm for generating a graph of the surrounding space from the data obtained from a three-dimensional lidar installed on a mobile robot with O(nlogn) complexity, which reduces the time to build a local trajectory. The developed solution makes it possible to combine the processes of constructing a local and global trajectory of movement by using the energy-efficient algorithm for constructing trajectories LRLHD-A* as the basis. During the experiments, it was revealed that when choosing a step of 0.5 m for a local grid of 10 × 10 m in size, the time for constructing the grid and the trajectory does not exceed 0.1 s. The developed algorithm can be used as a local scheduler in close to real time mode, providing prompt re-sponse to various obstacles on the global trajectory of the robot, considering the features of the terrain and the cur-rent global trajectory.