Cluster space collision avoidance for mobile two-robot systems

The cluster space state representation for multi-robot systems provides a simple means of specifying and monitoring the geometry and motion characteristics of a cluster of mobile robots. In previous work, this approach has been experimentally verified and validated for controlling the motion of mobile multi-robot systems ranging from land rovers to autonomous boats. In this paper we introduce a compact collision avoidance algorithm that operates at the level of the cluster, leading to coordinated translational and rotational motions that allow obstacles to be avoided while maintaining the relative geometry of the cluster. This paper formulates the potential-field based obstacle avoidance algorithm, describes its integration within the cluster space control architecture, and presents successful experimental results of its application to two simple, diverse multi-robot testbeds.