A Multi-Node Energy Prediction Approach Combined with Optimum Prediction Interval for RF Powered WSNs

Area of Science:

• Wireless Sensor Networks
• Energy Harvesting
• Power Management

Background:

• Efficient power management is crucial for environmentally powered Wireless Sensor Networks (WSNs).
• Traditional Moving Average (MA)-based energy prediction relies solely on a node's past energy readings.
• Radio Frequency Energy Harvesting (RF-EH) WSNs present unique challenges and opportunities for energy prediction.

Purpose of the Study:

• To propose a novel Multi-Node energy prediction method for RF-EH WSNs.
• To enhance future energy availability predictions by incorporating data from neighboring nodes.
• To analyze prediction effectiveness concerning distance and optimize prediction intervals for energy neutrality.

Main Methods:

• Developed a Multi-Node energy prediction approach utilizing harvesting history from surrounding nodes.
• Analyzed the impact of effective prediction distance.
• Created a mathematical model to determine the optimal prediction interval for energy neutrality.

Main Results:

• The Multi-Node prediction method demonstrates reduced sensitivity to the prediction interval compared to traditional MA techniques.
• Distant nodes contributed less to the prediction, with their utilization decreasing as the prediction interval increased.
• A linear relationship was established between the prediction interval and the energy threshold limit.

Conclusions:

• Multi-Node energy prediction offers a more robust approach for RF-EH WSNs.
• Optimizing the prediction interval is key for balancing prediction accuracy and system design.
• The findings provide insights for designing more efficient and energy-neutral WSNs.