Data-Driven Distributed Coordinated Control for Cloud-Based Model-Free Multiagent Systems With Communication Constraints

Abstract

This paper aims to solve the problem of the coordination of cloud-based model-free multiagent systems (CBMF-MASs) with communication constraints. To improve the data processing ability of multiagent systems for massive real-time data and decrease the communication burden of individual agents, cloud computing systems are utilized to establish the multiagent system. To actively compensate for network delays and data losses in all communication channels and coordinate the output of the CBMF-MAS, a novel data-driven networked distributed predictive control method (DDNDPC) is presented, which is independent of system’s structure model and only relies on system’s input and output data. Furthermore, the stability and consensus criterion of the CBMF-MAS are established by proposing a simultaneous analysis approach for the stability and consensus. Finally, the effectiveness and practicality of the DDNDPC method are verified through numerical simulations and cloud-based experimental tests. The achievements promote the application of large-scale multiagent systems in practice.