Adaptive State Consensus of Uncertain MIMO Dynamical Multi-Agent Systems via Output Feedback over a Directed Graph

This paper investigates the state consensus problem of uncertain multi-input multi-output (MIMO) multi-agent systems (MASs) over a directed graph, where only the outputs of the agents are available. We consider two types of agent dynamics, namely, square systems with equal input-output dimensions, and non-square systems where outputs exceed inputs. For square systems, an adaptive consensus algorithm is developed that interacts solely through relative output information. Luenberger-like observers and local reference models are designed for each agent to handle both parametric uncertainties and inaccessible states. For non-square systems, an observer-based adaptive output feedback algorithm is proposed to achieve full-state consensus despite input-output mismatch. Simulation results demonstrate the effectiveness of the proposed approach.