Reactive Coordination and Adaptive Lattice Formation in Mobile Robotic Surveillance Swarms.

We present here a set of decentralised control laws to facilitate lattice formation and reactive coordination and control of a swarm of mobile ground based robots. The control laws rely on local, indirect communication, which we implement in the form of virtual forces governed by physics based laws, computed from range and bearing measurements relative to the individual robots. Furthermore, we introduce the Virtual Robot Node (VRN) architecture to extend the capabilities of the cooperative formation control in terms of lattice cohesion and reactive dynamic abilities. The characteristics of the control laws are analysed through a number of 3D physics-based simulation experiments. We show that the basic proposed methods exhibit robustness to simulated sensor noise. We further show a number of improvements made by employing the VRN architecture, in terms of reducing errors in specified formation constraints, and additional dynamic capabilities.