Adaptive backstepping sliding mode heading control for wave glider using neural network–based extended state observer

• A speed-dependent dual-body coupled heading coordination model is proposed, accurately reflecting the UWG’s real-world dynamics under time-varying conditions. • An RBF-IESO-based adaptive backstepping sliding mode controller is developed, with parameters self-tuned via an improved particle swarm optimization algorithm. • The controller demonstrates superior coordinated heading tracking and disturbance rejection in simulations and sea trials, outperforming conventional PID control. This study develops a speed-dependent dual-body coupled dynamic model and a coordinated heading control strategy for an unmanned wave glider (UWG). Initially, a coupled float-glider heading control model is developed, considering the time-varying speed of the UWG under the influence of external disturbances. To estimate unmeasurable system states, an improved extended state observer (IESO) is designed, enhanced by a radial basis function (RBF) neural network to approximate unknown system dynamics. Building upon the IESO, a backstepping sliding mode controller (BSMC) is designed, combining the recursive structure of backstepping with the robustness of sliding mode control. Additionally, an improved particle swarm optimization (IPSO) algorithm is employed to optimize controller parameters. Simulation results and sea trials demonstrate that the proposed RBF-IESO-IPSO-BSMC control system achieves superior heading coordination and disturbance rejection, outperforming conventional control methods.