Distributed Dynamic Average Consensus in Asymmetric Networks

The distributed dynamic average consensus problem in an asymmetric network, represented by a weighted directed graph, is investigated in this paper. A novel distributed control scheme is proposed in this work in discrete time such that the state vector of all nodes converge to a neighborhood of the average of multiple time-varying smooth reference vectors. The convergence analysis of the proposed control procedure is performed by establishing an upper bound on the maximum tracking error from the viewpoint of all nodes provided the design parameters along with the sampling period are properly chosen. The efficiency of the results is subsequently demonstrated by simulations.