A Kalman filter to estimate altitude change during a fall

Area of Science:

• Biomedical Engineering
• Sensor Technology
• Wearable Devices

Background:

• Barometers enhance fall detection accuracy in wearable devices.
• Traditional barometers are power-intensive, limiting device longevity.
• Integrating barometers into fall detectors requires power-efficient solutions.

Purpose of the Study:

• To evaluate a Kalman filter (KF) for low-power barometer operation in fall detection.
• To assess the KF's ability to estimate pressure changes during falls.
• To enable extended battery life for fall detection devices.

Main Methods:

• Offline evaluation of a Kalman filter (KF) algorithm.
• Utilizing data from simulated falls and activities of daily living.
• Leveraging pressure equalization delay caused by a semi-permeable membrane in the fall detector enclosure.

Main Results:

• The KF successfully differentiated between fall and non-fall activities.
• An average pressure change of 8 Pa during falls was best estimated with a 20-second delay.
• The KF detected altitude changes faster than a moving average filter (MAF).

Conclusions:

• The Kalman filter enables low-power barometer operation for fall detection.
• KF-based pressure change estimation improves fall detection accuracy and efficiency.
• This approach offers a viable solution for power-constrained wearable fall detection systems.