Distributed algorithm for dissensus of a class of networked multiagent systems using output information

In this paper, a distributed algorithm is developed to solve the dissensus of a class of networked multiagent systems only using output information. By introducing a gauge transformation, the dissensus problem is transformed to the problem of demonstrating that (A, B, C) is stabilizable and detectable. If the networked multiagent systems can reach dissensus, the signed digraph is structurally balanced containing a spanning tree. Furthermore, by solving a Riccati equation, the necessary condition becomes a necessary and sufficient condition. Finally, two examples are provided to illustrate our results. There are three main contributions in this paper: (1) a distributed algorithm with output information is introduced to deal with the difficulty of obtaining relative full-state observations; (2) the undirected communication graph is extended to the signed digraph which is more practical in physical implementations; (3) the method established in this paper is also applicable to discrete-time networked multiagent systems by using a gauge transformation, which further demonstrates the generality of our results.