Adaptive Neural Control of Flight-Style AUV for Subsea Cable Tracking Under Electromagnetic Localization Guidance

Abstract

This article focuses on an industrial system for the intelligent inspection of subsea cables based on a flight-style autonomous underwater vehicle (AUV) that tracks the route of subsea cable guided by electromagnetic localization signals. The intelligent inspection system consists of electromagnetic localization and autonomous tracking control modules. The electromagnetic localization algorithm for subsea cable is geometrically derived based on the electromagnetic sequences measured with two triaxial sensors that are symmetrically mounted on the forewings of a flight-style AUV. The argument results on the singularity, sensitivity, and consistency of the localization algorithm provide guidelines for designing the cable-tracking AUV prototypes and planning the near-bottom cable-tracking tasks. Subsequently, the cable-tracking control schemes for decoupled AUV heading and diving subsystems are developed by using the localization results. Particularly, the vector field guidance based on the localization results is designed to overcome the challenges, including underactuation and the lack of cable-tracking error dynamics. To minimize the electromagnetic noise interference and to relax the dependence on the accurate AUV hydrodynamics, the kinematic virtual control signals are filtered with low-pass filters, and neural networks are applied to approximate dynamics. Finally, an integrated simulation study is presented to demonstrate the effectiveness and robustness of the proposed electromagnetic localization-guided autonomous cable-tracking system.