Anomaly Detection of Underwater Gliders Verified by Deployment Data

TL;DR

This paper presents an anomaly detection algorithm for underwater gliders, enabling real-time operational status monitoring and damage prevention, validated with real deployment data from two oceanographic gliders.

Contribution

It introduces a novel anomaly detection method specifically designed for underwater gliders and validates it with real-world deployment data.

Findings

Effective real-time anomaly detection in underwater gliders

Validation with USF Stella and SkIO Angus deployment data

Potential to improve glider safety and operational efficiency

Abstract

This paper utilizes an anomaly detection algorithm to check if underwater gliders are operating normally in the unknown ocean environment. Glider pilots can be warned of the detected glider anomaly in real time, thus taking over the glider appropriately and avoiding further damage to the glider. The adopted algorithm is validated by two valuable sets of data in real glider deployments, the University of South Florida (USF) glider Stella and the Skidaway Institute of Oceanography (SkIO) glider Angus.