In this paper, we investigate the leader–followers Time-Varying Formation Tracking (TVFT) problem of a networked Multi-Agent Systems (MASs) based output feedback. The agents’ behavior is featured by linear dynamics. The interaction topology among the agents is switching over time, directed and only assumed to contain a directed spanning tree. This latter is rooted by a leader-agent whose control input is unknown and bounded. A novel fully distributed TVFT controller is proposed that exhibits a reduced network information exchange property among the agents, thus less communicating-resources are utilized. The proposed controlled design is based on an adaptive observer and disturbance rejection technique, where the unknown leader-input is viewed as an external disturbance. The key feature lies in introducing a local observer in each agent-controller design to observe all the relative neighboring output-measurements and the relative neighboring distributed observer-outputs, all gathered in one signal that we denote the agent’s network information signal. The Lyapunov theory is used to prove that the closed-loop MASs tracking error is stable. An analysis of the effect of the interaction topology structure on the tracking-error convergence rate is further provided, showing the validity of the proposed formation control for switching interaction topologies. Finally, to verify the effectiveness of the obtained results, the proposed controller is extended to a cooperative guidance of a networked quadrotors to track and entrap an uncooperative aerial target that plays the role of a passive leader.

In this paper, a novel robust backstepping-based fault-tolerant control is designed, implemented and validated experimentally on a quadrotor unmanned aerial vehicle testbed under actuator fault conditions for tracking control. Backstepping is known as a robust method to maintain system performance and keep it insensitive to disturbances. The proposed nonlinear controller is mainly based on the advanced robust backstepping method to solve the system uncertainties caused by disturbances and actuator faults, which appear in the motor part of the system and are compensated without diagnostics or fault identification. The various parameters of the proposed approach are optimized using the particle swarm optimization (PSO) method. Owing to the minimized control effort to accommodate uncertainties compared to the conventional backstepping, the proposed approach can still maintain the system performance when severer faults occur. The proposed approach is validated, first, in a simulation using the nonlinear model of the quadrotor, then experimentally, using the Quanser 3DOF Hover quadrotor. Both theoretical and experimental analyses have demonstrated that the effectiveness of the proposed improved backstepping fault-tolerant control strategies faces significant loss of actuator efficiency.