A novel system identification technique for improved wearable hemodynamics assessment

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Wearable Technology

Background:

• Ballistocardiography (BCG) noninvasively measures cardiovascular events through body movements.
• Existing BCG platforms (beds, chairs, scales) require direct body coupling.
• Wearable BCG sensors offer continuous monitoring but capture local, not whole-body, signals.

Purpose of the Study:

• To propose a novel method for reconstructing weighing scale BCG (WS BCG) from wearable sensor data.
• To remove local signal effects and enable accurate whole-body BCG measurement.
• To validate the method for potential use in remote patient monitoring, such as for heart failure.

Main Methods:

• A novel reconstruction method using a training step to remove local effects from wearable BCG signals.
• Preliminary validation with 15 subjects, simultaneously recording wearable BCG and WS BCG.
• Application of a regularized system identification approach to reconstruct WS BCG from wearable BCG.

Main Results:

• Preliminary results indicate individual and sensor location significantly influence the local-to-central disturbance relationship.
• The proposed calibration method can resolve these differences.
• Accurate measurement of cardiac output and contractility from wearable sensors is feasible.

Conclusions:

• Reconstructing WS BCG from wearable sensors is possible through a calibration process.
• This approach allows for accurate assessment of cardiac function using easily accessible wearable devices.
• Potential for improved remote monitoring of cardiovascular conditions, including heart failure.