Output-Feedback Consensus-Formation Control of Nonholonomic Vehicles With Input Constraints and Time-Varying Delays

We provide a solution to the consensus-based formation control problem of multiagent nonholonomic vehicles, under the assumption that the magnitudes of control inputs must satisfy certain preimposed bounds, to avoid saturation of the actuators. We consider that the communications are affected by time-varying delays and the vehicles lack velocity measurements. The desired control objective is that all the vehicles converge to a given desired formation and that they reach a consensus on the localization of the center of such formation and on their orientations. The control architecture exploits several features. From a control viewpoint, the systems’ dynamics are split into their angular-motion and linear-motion parts, so each of these is driven by a separate distributed dynamic output-feedback controller, passively interconnected to the plant via a virtual spring. From a network-topology viewpoint, the overall networked system contains two layers of interconnections with undirected graphs. Finally, from an analytical viewpoint, each vehicle-controller closed-loop system may be regarded as a cascade with a $\boldsymbol{\delta }$-persistently exciting interconnection, which ensures the achievement of the control goal, in spite of the nonholonomic constraints.