A Novel Method for Real-Time Human Core Temperature Estimation Based on Extended Kalman Filter.

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Signal Processing

Background:

• Current core temperature (CT) monitoring methods are invasive and costly.
• Kalman filter applications for indirect CT estimation have been studied for over a decade.
• Existing Extended Kalman Filter (EKF) approaches require comparative analysis.

Purpose of the Study:

• To compare state-of-the-art Extended Kalman Filter (EKF) methods for core temperature estimation.
• To introduce and evaluate a novel pre-emptive mapping layer using heart rate (HR) signals for EKF input.
• To enhance the accuracy and efficiency of non-invasive CT monitoring.

Main Methods:

• Comparative analysis of different Extended Kalman Filter (EKF) algorithms.
• Development of a pre-emptive mapping concept utilizing heart rate (HR) signal characteristics.
• Algorithm training and testing on two distinct datasets comprising 18 subjects.

Main Results:

• The best-performing EKF approach incorporated the novel pre-emptive mapping layer.
• This enhanced method achieved an average Root Mean Squared Error (RMSE) of 0.34°C.
• The pre-emptive mapping resulted in a 17% performance improvement compared to EKF without mapping (RMSE of 0.41°C).

Conclusions:

• The proposed pre-emptive mapping significantly enhances EKF-based core temperature estimation accuracy.
• This non-invasive method offers a more efficient and potentially cost-effective alternative to traditional CT monitoring.
• Future research should validate this method on larger datasets to confirm its clinical applicability.