Robust Finite-Time Dynamic Average Consensus With Exponential Convergence Rates
Kedong Xu, Lan Gao, Fei Chen, Chaojie Li, Qi Xuan
Abstract
This brief focuses on the dynamic weighted average consensus problem and aims to achieve accurate tracking of the weighted average of all the time-varying reference signals in a network. We first propose a robust dynamic weighted average consensus (RDWAC) algorithm that employs a simple fixed control gain and introduces an individual weight for each agent compared with recent works. Furthermore, a theoretical finite-time convergence analysis instead of an asymptotic one is provided by constructing a novel Lyapunov function, which shows that the accurate weighted average consensus can be reached exponentially within a finite time interval. In addition, the lower bound of the required convergence time is given and the relationship between the lower bound and the initial steady-state error and control parameters is established explicitly. Finally, some numerical examples are given to illustrate the effectiveness of the proposed algorithm.