Flocking for Multirobots Without Distinguishing Robots and Obstacles.

Most existing studies of multiple mobile robots assume that robots can distinguish between other robots and obstacles. However, if a flocking algorithm was available that did not require this ability, easier implementation could be expected. In this brief, we propose a flocking algorithm that does not distinguish between a robot and an obstacle. In other words, all detected objects are regarded as obstacles in the proposed algorithm. Thus, velocity information on neighboring robots is not required. We also show that the proposed algorithm maintains the desired properties of existing flocking algorithms, even though only limited information is used in the proposed algorithm. Furthermore, unlike many previous studies, the effectiveness of the algorithm is demonstrated not only by simulations, but also in real robot experiments. In the experiments, neighboring robots and obstacles are detected using only laser range finders, since the robots do not need to distinguish between another robot and an obstacle.