Hierarchical trajectory planning for multi-robots systems

Abstract

In this work, a hierarchical trajectory planning method to guide a multi-robot system from an initial point to a given end point is presented, considering the possibility of division and meeting of the training. A global planner based on the Dijkstra algorithm computes sequentially the optimal paths for each robot in the formation, while a local planner determines the set of movements required for each robot. This set of movements makes it possible to move a quadcoptertype flying robots along the path(s) obtained by the global planner, avoiding collisions with obstacles in the environment and between them. The motion vectors were obtained by simulation, using a non-linear position-tracking controller, and a repulsion model based on a potential function to avoid collisions between robots.