Decentralized estimation of ocean current field using underwater acoustic sensor networks

Area of Science:

• Oceanography
• Acoustic Sensing
• Distributed Systems

Background:

• Accurate estimation of 2D horizontal ocean current fields is crucial for various marine applications.
• Underwater Acoustic Sensor Networks (UASNs) offer a promising platform for in-situ ocean monitoring.
• Existing methods often face challenges in decentralized processing and adapting to dynamic ocean conditions.

Purpose of the Study:

• To propose a novel decentralized method for estimating the 2D horizontal ocean current field using UASNs.
• To integrate advanced tomography and time-series modeling techniques for enhanced current field estimation.
• To develop a robust and computationally efficient distributed estimation framework.

Main Methods:

• The proposed "UASN-decentralized" method combines triangle-division-based travel time difference tomography with a spatiotemporal autoregressive model.
• A distributed information Kalman filter is employed for decentralized estimation and tracking.
• Sensor nodes are classified into types I and II to optimize computations, and a shortest-path-based consensus weight matrix adapts to ocean dynamics.

Main Results:

• Synthetic data validation demonstrates the feasibility and robustness of the UASN-decentralized method against measurement errors.
• Monte Carlo simulations confirm the method's effectiveness in estimating ocean current fields.
• The study shows that increased communication rounds enhance performance for fast-varying dynamics and lower sensor rates.

Conclusions:

• The UASN-decentralized method provides a feasible and robust solution for estimating 2D horizontal ocean currents.
• The integrated approach effectively handles complex ocean dynamics and sensor network constraints.
• The method's adaptability through communication rounds offers flexibility for diverse oceanographic monitoring scenarios.