Robust Adaptive Dynamic Programming for Dual-Redundant Actuation Systems in Aircrafts

Redundancy actuation systems are widely applied in the fields of aviation and aerospace. This paper establishes a nonlinear dynamic model of the dual-redundancy actuation system driven by electro-mechanical actuators, involving control constraints and unmatched bounded perturbation. We next focus on the reset control of the dual-redundancy actuation system and raise a robust control problem, which can be transformed into a specified optimal controller design problem theoretically. To acquire the optimal controller, we propose a modified robust adaptive dynamic programming approach. Using multiple critic neural networks, a simultaneous approximation of the value function and the controller are achieved, and all the weight parameters can be estimated through an optimization framework, which is feasible guaranteed via well-defined basis functions. Finally, a numerical example from the dual-redundancy actuation system is presented to illustrate the derived controller.