Voronoi-Based Collision Avoidance Using Localization Uncertainty Region

Multi-robot systems have been applied to numerous professional areas in industry, consisting of applications from warehousing logistics to drone swarm for military. Under realistic working conditions, errors of localizations may lead to collisions between robots although the multi-robot system is theoretically collision-free. This paper introduces a comprehensive control algorithm for collision-free multi-robot systems based on buffered Voronoi diagram enhanced by localization uncertainty region. Compared to previous related works, this paper combines the features of the Voronoi diagram, two-dimensional Gaussian distribution and the deadlock avoidance algorithm together to derive a control algorithm endowing the multi-robot system with high robustness and practical reliability.