Swarm Trajectories Generation for Target Capturing With Uncertain Information

Abstract

This work describes a methodology to generate trajectories for a swarm of agents, modeled as single integrators, for achieving target capturing/enclosing and containment. The discussed method proposes a robust solution that relaxes the strong assumptions about the exact knowledge of the target information. In particular, each agent requires a rough estimation of the reference, with a priori known bounded error both on the position and the velocity. It is shown that the swarm can be properly guided to enter a given capturing region while remaining outside a defined distancing region around the target. A steady-state analysis for velocity consensus is also provided for the case in which the reference information is exactly known by the agents.