Intelligent routing for human activity recognition in wireless body area networks

Area of Science:

• Biomedical Engineering
• Computer Science
• Machine Learning

Background:

• Wireless Body Area Networks (WBANs) face challenges with limited energy, mobility, and dynamic environments.
• Human Activity Recognition (HAR) using machine learning can detect user movements, offering potential for WBAN optimization.

Purpose of the Study:

• To develop a novel adaptive routing protocol for WBANs by integrating HAR capabilities.
• To enhance WBAN performance, focusing on optimizing power consumption and routing decisions.

Main Methods:

• Employed machine learning, specifically the Random Forest classification algorithm, for HAR.
• Developed a novel WBAN routing protocol that adaptively adjusts decisions based on recognized human activities.

Main Results:

• The proposed HAR-based routing protocol demonstrates significant promise for optimizing routing decisions and improving energy efficiency in WBANs.
• Performance evaluation showed improvements in energy consumption, throughput, and a reduction in the number of dead nodes compared to the mobTHE protocol.

Conclusions:

• Integrating HAR with machine learning offers a promising approach to enhance WBAN performance, particularly in healthcare.
• The developed adaptive routing protocol effectively optimizes energy consumption and network longevity for WBANs.