Method for predicting depth-averaged current velocities of underwater gliders based on data feature analysis

Baoqiang Tian,Shijie Liu,Yonglai Zhang,Yaojian Zhou

doi:10.1063/5.0058318

Abstract

In this paper, the data feature of depth-averaged current velocities (DACVs) derived from underwater gliders is analyzed for the first time. Two features of DACVs have been proposed: one is the complex ingredients and small samples, and the other is the stationarity that occurs as the length of a DACV sequence increases. With these features in mind, a set of methods combining statistical analysis and machine learning are proposed to realize the prediction of DACVs. Four groups of DACV data of different gliders from sea trials in the South China Sea are used to verify the prediction method. Based on three general error criteria, the prediction performance of the proposed model is demonstrated. The persistence method is used as a comparison model. The results show that the prediction methods proposed in this paper are effective.

Highlights

Underwater gliders are a new type of autonomous underwater vehicle (AUV)
The reason for this is that RF is well suited for predicting deep averaged current velocities (DACVs) data with feature 1, while Autoregressive integrated moving average (ARIMA) is highly targeted for feature 2, but there is a certain coupling between feature 1 and feature 2, which cannot be clarified
The features of DACVs are analyzed, and methods based on data statistics and the method based on machine learning are proposed as the targeted forecasting model

Summary

INTRODUCTION

Underwater gliders are a new type of autonomous underwater vehicle (AUV). Compared with the traditional AUVs, underwater gliders are of low cost, are easy to maintain, and offer high endurance over long distances. Various sensors can be integrated into underwater gliders to measure different ocean indicators, such as salinity, depth, chlorophyll, and dissolved oxygen. In order to improve the universality and practicability of the DACV prediction method, this paper proposes a prediction method based on the analysis of the features of the DACV series, and the classical time series prediction method is used to forecast the DACVs. The forecast process does not require the inclusion of specific complex sea tide models, and there is no limit to the applicable sea area. The research results of this paper can provide critical support for the efficient planning and accurate navigation of the underwater gliders and provide a theoretical basis for the research on the prediction methods of ocean currents, sea winds, waves, and other ocean environmental elements matching with other types of ocean mobile observation platforms.

DACVs OF UNDERWATER GLIDERS AND THEIR FEATURES

Feature 1

Feature 2

Random forests

Proposed method

Evaluation criteria

RESULTS AND ANALYSIS

CONCLUSION AND FUTURE WORKS