Average consensus problems of neutral multi-agent systems with time-varying delay and switching topology

This paper introduces a new type of consensus protocol in undirected networks, which leads to a corresponding neutral multi-agent system. Such a protocol not only comprises the usual linear protocol as a particular case, but also allows a less conservative estimate on the upper bound of time delay. By decoupling the neutral multi-agent system in terms of the eigenvalues of the Laplacian matrix, necessary and sufficient conditions for average consensus have been established for the case of fixed topology. For the case of switching topology, a sufficient condition for average consensus is given by using a linear matrix inequality method. Finally, two examples are worked out to explain the effectiveness of the theoretical results.