Adaptive predefined time neural filtered control design for an uncertain nonlinear system and application to flight control.

The paper studies the control problem of an uncertain nonlinear system with tracking error constraint, unknown nonlinear functions and unknown control input gains. An asymmetric barrier Lyapunov function composed with predefined time prescribed performance function is constructed to ensure tracking error enters into the predefined asymmetric constraint in a given time. Then radial basis function neural network is adopted to approximate unknown functions. To reduce computation load, minimal-learning parameter technique is applied. Meanwhile, adaptive method is used to solve actuator faults and the unknown control input gains. Moreover, an adaptive neural control strategy is designed in the framework of backstepping method. An adaptive fixed time filter is developed for avoiding the "explosion of complexity" problem, where the convergence speed of the filter error is improved compared with fixed time filter. It is proved that all signals of the closed-loop system are bounded and tracking error is kept in its constraint boundary. At the end, compared numerical simulations and application simulation of a hypersonic vehicle are demonstrated to verify the efficiency of the designed control scheme. • A composed performance function is designed to ensure the tracking error enters into the constraint in a given time. • An adaptive fixed time filter is developed to solve the 'explosion of complexity' problem. • Based on minimal-learning parameter method, neural networks estimate unknown functions, which reduce computation load. • Compared simulation and application simulation of a hypersonic vehicle are given to test the efficiency of the controller. [ABSTRACT FROM AUTHOR]