An Optimal, Power Efficient, Internet of Medical Things Framework for Monitoring of Physiological Data Using...

Area of Science:

• Biomedical Engineering
• Computer Science
• Wireless Communication

Background:

• Internet of Medical Things (IoMT) networks rely on battery-powered sensors that require frequent replacement or energy harvesting.
• Embedded sensors in IoMT pose challenges for power replenishment, necessitating innovative solutions for network longevity.
• Reducing data transmission and reconstructing signals at the receiver end is a potential strategy to conserve sensor battery power.

Purpose of the Study:

• To evaluate the accuracy of reproducing physiological signals using decision tree-based regression models.
• To assess the impact of varying the number of decision trees on model efficiency and accuracy.
• To demonstrate significant battery power savings and improved network longevity in IoMT applications.

Main Methods:

• Development and application of decision tree-based regression models for physiological signal reproduction.
• Transmission of two sets of physiological signals over cellular networks.
• Execution of regression analyses across three iteration varieties to test model performance with different numbers of decision trees.

Main Results:

• The proposed decision tree regression models achieved significantly lower error rates in data reproduction compared to existing methods.
• The study quantified the relationship between the number of decision trees and the efficiency of the regression model.
• Results indicate substantial potential for battery power savings and enhanced network longevity in IoMT.

Conclusions:

• Decision tree-based regression models offer a promising approach to reduce data transmission in IoMT, thereby conserving sensor battery power.
• Accurate signal reconstruction at the receiving end can significantly extend the operational lifespan of IoMT networks.
• This method provides a viable solution for sustainable and long-lasting IoMT deployments, especially for embedded sensors.