Distributed Adaptive Observer-Based Control for Output Consensus of Heterogeneous MASs with Input Saturation Constraint

Without assuming that the global connectivity information of the interaction graph is available to any agent, fully distributed saturated controller design for output consensus of heterogeneous multi-agent systems (MASs) under directed leader-follower communication topologies is addressed in this paper. The main difficulty of the problem lies in the coupling of the nonlinearities caused by the input saturation and the adaptive feedback control. To decouple these two types of nonlinearities, a distributed adaptive observer is first designed for each follower to realize asymptotical estimation of leader's state, where the adaptive coupling gain is embedded to estimate the graph connectivity, and then an extra tracking error observer is introduced to tackle input saturation constraint. The controller is then constructed based upon these two observers, which is fully distributed with only the information of the distributed adaptive observer transmitted between neighboring agents and the local output information of each agent involved. To reduce the order of the controller, an in-depth analysis of the proposed controller is conducted, and the novel controller is designed by removing the tracking error observer and modifying the multi-level saturation control parameters for each follower. Furthermore, to remove the limitation of using absolute output information of each agent, novel controller is proposed with tracking error estimation designed in a distributed manner, which can achieve output consensus tracking for heterogeneous non-introspective MASs with undirected subgraph among followers.