Neuroadaptive Fully Actuated Attitude Tracking for Robot Manipulators with Asymmetric Full-State Constraints

This paper studies the neuroadaptive attitude tracking problem for robot manipulators systems with asymmetric time-varying full-state constraints, model uncertainties, and external disturbances. Most existing robot manipulator attitude control works are constructed through the state-space approach. In this paper, a neuroadaptive attitude control strategy is developed based on the fully actuated system (FAS) approach, which has shown its simplicity and flexibility for controller design of nonlinear systems. We adopt adaptive multilayer neural network (NN) to tackle lumped uncertainties due to the model uncertainties and external disturbances. On this basis, a nonlinear transformation is further integrated into the FAS approach-based control design to achieve exact asymptotic convergence while satisfying full-state constraints. The attitude tracking error is rigorously proven to asymptotically converge to zero, which is a notable improvement over existing NN control works. Moreover, using the FAS approach, stability analysis and controller design are much simpler compared to the conventional backstepping or sliding mode design. Stability analysis and numerical simulations validate the control performance of the proposed control strategy.