Energy-efficient multimodal motion control of hybrid underwater gliders based on an optimized model predictive active disturbance rejection framework

Abstract

Hybrid Underwater Glider (HUG) serve as critical platforms for extended and large-scale ocean observations. However, achieving precise trajectory tracking in complex and dynamic marine environment, remains a significant challenge. This paper addresses the multimodal trajectory tracking problem for HUG by proposing a composite control framework integrating Active Disturbance Rejection Control (ADRC), Model Predictive Control (MPC), and Adaptive Line-of-Sight (ALOS). First, a six-degree-of-freedom dynamic model is established, ALOS is designed to achieve decoupling and smooth guidance of horizontal and vertical motions by dynamically adjusting the look-ahead distance and depth-pitch coupling mechanism. Second, an Extended State Observer (ESO) was validated for real-time estimation of total system disturbances. MPC replaced the conventional Nonlinear Law State Error Feedback (NLSEF), utilizing rolling optimization strategies to address actuator physical constraints and state constraints, thereby enhancing the predictive capability and robustness of the controller. Finally, the framework’s performance was validated through multimodal task simulations. Simulation results demonstrate that this framework significantly outperforms traditional methods. During trajectory tracking, horizontal tracking error is maintained within 3 meters, depth deviation is reduced to 0.52 meters. In the constant-depth navigation phase, depth fluctuations are less than 0.15 meters. This approach effectively addresses the challenge of synergistic control under model uncertainty and multi-mode switching.