Robust Distributed Formation Control of Agents With Higher-Order Dynamics

Abstract

We present a distributed control strategy for agents with a variety of holonomic dynamics to autonomously achieve a desired formation. The proposed control is fully distributed, and can be implemented locally on agents using the relative position measurements. Furthermore, agents do not need to communicate or have a common sense of orientation. Our strategy allows control gains designed for agents with single-integrator dynamics to be used directly for agents with higher-order linear (or linearizable) holonomic dynamics. We provide rigorous mathematical analysis to prove convergence of the agents to the desired formation. The proposed method is applied to quadrotors with linearized dynamics, and simulations are provided to typify the theoretical results.