Distributed Robust Fixed-Time Dynamic Average Consensus Protocol for Networked Agents

This paper seeks a robust dynamic average consensus protocol that makes the networked agents reach a common time-varying average state of multiple reference signals within an advanced preset time. It can be applied in fields such as multi-robot map merging and DC microgrid secondary control. Finite-time dynamic average consensus algorithms are studied in existing papers, while the convergence time is dependent on the initial states of the agents, so larger differences among their initial states may result in a slower convergence rate. Hence, a robust fixed-time dynamic average consensus scheme is presented in this paper. The settling time of the proposed scheme only relies on the parameters of the protocol and is independent of the initial values of agents. Specifically, the theoretical fixed-time convergence analysis is achieved by adopting the Lyapunov function, which indicates that dynamic average consensus can be reached within a fixed-time interval. Moreover, a damping term is added to the internal dynamics of agents to enhance the robustness of the protocol once the network interaction is changed. Finally, some numerical simulations are shown to verify the effectiveness of the proposed strategy.