A distributed model predictive control strategy for finite-time synchronization problems in multi-agent double-integrator systems

摘要

In this paper, a synchronization problem for a group of agents described by double-integrator dynamics and time-varying communication network is formally addressed. Specifically, the main aim consists in ensuring that the agents, organized as a swarm, are capable to synchronize in finite-time the swarm centroid to a given reference trajectory and to keep this property for all future time instants. To this end, an ad-hoc distributed control architecture is developed by taking advantage from two key ingredients: a novel swarm formation modelling that guarantees finite-time convergence properties of the agent kinematics and a hybrid model predictive control scheme in charge to adequately exploit such a kinematics description. One of the relevant features of the resulting distributed algorithm relies on the fact that both parallel and sequential paradigms are used with the aim to ensure collision avoidance requirements amongst the involved agents. (C) 2019 European Control Association. Published by Elsevier Ltd. All rights reserved.