Distributed neural-based fault-tolerant control of multiple flexible manipulators with input saturations

This paper focuses on a novel fault-tolerant control design for networked flexible manipulators with actuator failures, input saturations, and external disturbances. Unlike the existing methods, the new control strategy has three characteristics. First, the second-order auxiliary system is used to eliminate the influence of input saturation. Second, by designing some proportional–integral (PI) auxiliary variables, the boundary dynamics of manipulators are reconstructed as some new second-order systems, where the nonlinear uncertainties are estimated by the adaptive radial basis neural networks (RBFNNs) and the compound disturbances are compensated by the adaptive method. Third, a boundary local angle interaction protocol is given under a digraph. Then combining the virtual control and Lyapunov design, some novel distributed fault-tolerant control algorithms are obtained to ensure that the vibration states and the angle consensus error are uniformly ultimately bounded. Finally, numerical simulations are provided to show the effectiveness of these control algorithms.