RBFNN-Based Adaptive Fixed-Time Sliding Mode Tracking Control for Coaxial Hybrid Aerial–Underwater Vehicles Under Multivariant Ocean Disturbances

In this study, the design of an adaptive neural network-based fixed-time control system for a novel coaxial trans-domain hybrid aerial–underwater vehicle (HAUV) is investigated. A radial basis function neural network (RBFNN) approximation strategy-based adaptive fixed-time terminal sliding mode control (AFTSMC) scheme is proposed to solve the problems of the dynamic nonlinearity, model parameter perturbation, and multiple external disturbances of coaxial HAUV trans-media motion. A complete six-degrees-of-freedom model for a continuous water–air cross-domain model is first established based on the hyperbolic tangent transition function, and, subsequently, based on a basic framework of FTSMC, a fixed-time and fast-convergence controller is designed to track the target position and attitude signals. To reduce the dependence of the control scheme on precise model parameters, an RBFNN approximator is integrated into the sliding mode controller for the online model identification of the aggregate uncertainties of the coaxial HAUV, such as nonlinear unmodeled dynamics and external disturbances. At the same time, an adaptive technique is used to approximate the upper bound of the robust switching term gain in the controller, which further offsets the estimation error of the RBFNN and effectively attenuates the chattering effect. Based on Lyapunov stability theory, it is proven that the tracking error can converge in a fixed time. The effectiveness and superiority of the proposed control strategy are verified by several sets of simulation results obtained under typical working conditions.