Value of Information Based Data Retrieval in UWSNs

Area of Science:

• * Underwater Sensor Networks
• * Robotics and Autonomous Systems
• * Data Communication and Networking

Background:

• * Underwater sensor networks face challenges with high data acquisition rates exceeding communication capabilities via acoustic channels.
• * Autonomous underwater vehicles (AUVs) are used for data offloading, but this can introduce significant latency.
• * Timely reporting of high-priority events (e.g., catastrophes, intrusions) is crucial, demanding minimal end-to-end delay.

Purpose of the Study:

• * To enhance the reporting of high-priority events in underwater sensor networks.
• * To investigate the effectiveness of the Value of Information (VoI) metric for optimizing AUV data collection and delivery.
• * To compare different path planning strategies for AUVs based on data value and priority.

Main Methods:

• * Development and application of a Value of Information metric to classify sensor data by value and priority.
• * Utilizing a hybrid approach for data transmission, combining acoustic and optical communication channels.
• * Experimental evaluation using a binary event model (high-priority vs. low-priority events) and distinct AUV path planning strategies.

Main Results:

• * AUV path planning algorithms incorporating the Value of Information metric significantly improve the timely reporting of high-priority data.
• * The proposed approach leads to the accumulation of greater overall information value within the network.
• * Effective classification and prioritization of data are key to reducing latency for critical event dissemination.

Conclusions:

• * The Value of Information metric offers a viable strategy for optimizing data reporting in underwater sensor networks with AUVs.
• * Intelligent scheduling of AUV visits based on data value and priority is essential for efficient and timely event reporting.
• * This research demonstrates a method to balance data offloading needs with the critical requirement for rapid high-priority event detection.