Distributed Multi-robot Collision Avoidance under Localization Uncertainty

Multi-robot collision avoidance is an avoidable problem in the actual application of robots. Whether in multi-robot collaborative work or human-machine cooperation, the problem of multi-robot collision avoidance must be solved. In this paper, we put forward a distributed multi-robot collision avoidance algorithm, which ensures a crowd of robots to avoid colliding to each other while they are navigating to the goals cooperatively. To achieve this goal, a variant of the Voronoi diagram, the buffered Voronoi diagram, is utilized to incorporate robot path planning with their localization errors to guarantee safety. In order to verify our algorithm, a series of simulation experiments have been carried out. According to the analysis, the algorithm is effective in avoiding robot collision, which interlocks to each other when they are moving in space simultaneously. These results shed light for the robots can avoid collisions with each other under the control of the algorithm, and this research can be extended in the aspect of unmanned driving technology.