Dynamic Average Consensus Over Strongly Connected Digraphs Based on Integral Surplus

This paper addresses the design of a dynamic average consensus (DAC) algorithm over strongly connected but may not necessarily balanced digraphs, which seems to be the first time in the literature. Specifically, a new concept of integral surplus is proposed for DAC problems. On this basis, an integral surplus DAC algorithm is developed for agents to track the average of their multiple dynamic input signals with a bounded steady-state error. Such error is tunable by some algorithm parameters and even vanishes for special classes of input signals. Simulation examples are presented to verify the theoretical results.