Research on Remote Anomaly Detection Method for Underwater Gliders Based on Multi-Feature Parameters

Underwater glider is a new type of underwater unmanned vehicle, which realizes the submersible movement by adjusting the net buoyancy and adjusts its attitude by relying on the movable and rotating internal mass block. It has the characteristics of low power consumption, long endurance and low cost. The existing airborne anomaly detection mechanism can respond to most anomalies by setting a fixed threshold for a single detection parameter, but it is difficult to detect some anomalies with gradualism in time. At the same time, due to the limited data transmission and low power consumption of the platform, it is difficult to run the complex anomaly detection mechanism based on big data online. To solve the above problems, this paper takes underwater glider as the research object, and proposes a remote anomaly detection method based on multi-feature parameters under the condition of limited data, and verifies its feasibility through experiments.