Decentralized adaptive tracking control for nonlinear large-scale systems with unknown control directions

This article investigates decentralized adaptive tracking problems for nonlinear large-scale systems with strong interconnections, unknown control directions and actuator faults. In the presence of measurable system states, a decentralized state feedback control algorithm with a certain inherent fault tolerance capabilities is developed using the dynamic surface control and adaptive techniques. Particularly, a special Nussbaum gain is introduced in the control laws to compensate unknown control directions and unknown bounds of actuator efficiency factor. Using the graph theory and Lyapunov analysis method, it is proved that the designed fault-tolerant state feedback controller can drive the tracking error for each subsystem to a small neighborhood of the origin while keeping the semiglobal uniform ultimate boundedness for all other closed-loop signals. When the system states are unmeasurable, an output feedback fault-tolerant control framework is also formed by introducing a new K-filters with adjustable parameters and constructing appropriate coordinate transformations. Finally, the effectiveness of the proposed fault-tolerant control algorithms is verified by simulating two system models.