Multi-Sensor Data Fusion in A Real-Time Support System for On-Duty Firefighters

Area of Science:

• Wearable technology
• Biomedical engineering
• Safety systems

Background:

• Firefighters face significant risks including falls, incapacitation, and toxic gas exposure.
• Existing safety systems lack comprehensive monitoring for falls, performance loss, and environmental hazards like carbon monoxide (CO).

Purpose of the Study:

• To develop a portable system for detecting firefighter falls, loss of physical performance, and high carbon monoxide (CO) levels.
• To enhance firefighter safety through real-time monitoring and alerts.

Main Methods:

• Utilized a microcontroller with data fusion from a 3-DOF accelerometer, gyroscope, magnetometer, barometer, and MQ7 CO sensor.
• Developed novel algorithms for fall detection, loss of physical performance assessment, and CO monitoring.
• Implemented wireless alerting to the incident commander (IC) via an nRF224L01 module.

Main Results:

• Achieved 100% accuracy, specificity, and sensitivity in distinguishing falls from on-duty activities (ODAs) in experimental datasets.
• Demonstrated high accuracy (97.96%, 100%, 95.89%) on public datasets for fall vs. ODA detection.
• Successfully differentiated loss of physical performance from elevator movements using barometric data fusion.

Conclusions:

• The developed system effectively detects firefighter falls, performance degradation, and hazardous CO levels.
• Real-time alerts significantly improve emergency response and firefighter safety.
• Data fusion from multiple sensors enhances the reliability and accuracy of the safety system.