AI‐ML-Enabled Electromagnetic Homing Guidance System for Scientific Autonomous Underwater Vehicles

Abstract

Abstract Subsea homing and docking systems are used for increasing the spatiotemporal capabilities of autonomous underwater vehicles (AUVs) involved in long endurance scientific, survey, and surveillance missions. They offer necessary guidance for the AUV, and maneuver into the dock, considering the vehicle attitude and the dynamic response capabilities. Short-range homing is critical for successful docking as carrying outgross course and heading corrections is difficult when the AUV is closer to the dock. The article describes the development of an artificial intelligence (AI)-enabled short-range AUV electro-magnetic homing guidance system (EMHGS) based on differential magnetometry principle. System engineering is carried out with the aid of electromagnetic Finite Element Analysis software; supervised and unsupervised machine learning algorithms are implemented for determining the range and heading correction in real time. The proposed algorithms aid the reliable and accurate homing operation in an unknown dynamic environment, and the algorithm is very easily implementable and independent of AUV configuration and payloads. The EMHGS with an AI-enabled MagHomer AUV is demonstrated for autonomous intelligent homing over a range of 7 m.