Multi-robot Safe Navigation Under Localization Uncertainty
Zhiqing Yu · 2022 International Conference on Machine Learning and Intelligent Systems Engineering (MLISE) · 2022
This paper introduces a distributed avoidance collision algorithm for multi-robot that can account for the uncertain localization of each robot. The algorithm in this paper can safely drive the robot to its goal position by using local information while ensuring that it does not collide with other robots in the same environment during the movement. The algorithm achieves the above target by adding a buffer for each robot based on the Voronoi diagram called buffer Voronoi cell (BVC). The operational range of each robot retreat a safe distance based on the VD-based boundary to ensure avoidance of collision of robots. This paper demonstrates the efficacy of these algorithms via a series of simulations with different uncertain localization ranges of robots while exploring these effects on robot search time and path length. The results show that robots can safely move to their goal position in uncertain localization ranges. The algorithm also solves the deadlock problem that often occurs in multi-robot by obeying the right-hand rule for each robot in the same environment.