Key result
RCAKF reduces heart rate-derived core body temperature estimation error by ~21% vs standard Kalman filters.
Why the study?
Accurate real-time core body temperature estimation is critical to prevent heat-related illness, but existing Kalman filter-based methods use fixed observation models that fail to capture complex, non-linear, state-dependent physiological dynamics.
Does the Residual-Compensated Adaptive Kalman Filter (RCAKF) improve the accuracy of core body temperature estimation from heart rate compared to standard baseline models in healthy adults during exercise and heat stress?
Observational (n=22)
Does the Residual-Compensated Adaptive Kalman Filter (RCAKF) improve the accuracy of core body temperature estimation from heart rate compared to standard baseline models in healthy adults during exercise and heat stress?
Absolute Event Rate: 0.31% vs 0.39%
The RCAKF framework provides a highly accurate, single-sensor method for real-time core body temperature estimation from heart rate, enabling practical wearable safety monitoring for heat-related illness.
• A hybrid model is proposed that combines a Kalman filter with an LSTM to learn and correct observation residuals. • Using only heart rate, core temperature estimation (RMSE=0.31 °C) significantly outperforms five baseline models. • High sensitivity in detecting elevated temperatures supports its use in real-time, wearable safety monitoring systems. Accurate, real-time estimation of core body temperature is critical for preventing heat-related illness. While existing Kalman filter-based methods offer interpretable, single-input (heart rate) solutions, they are limited by fixed observation models that fail to capture the complex, non-linear, state-dependent dynamics of physiological signals. To address this, we propose the Residual-Compensated Adaptive Kalman Filter (RCAKF), a novel hybrid framework. The RCAKF integrates a long short-term memory (LSTM) network to learn and correct structured, state-dependent errors in the observation model, alongside an adaptive noise estimator that dynamically adjusts for measurement uncertainty. This architecture enhances the classic Kalman filter with data-driven flexibility while maintaining its recursive structure and interpretability. Evaluation was conducted on a controlled experimental dataset with 22 participants performing exercise and recovery under varied thermal conditions. Compared to five baseline models: extended Kalman filter (EKF: RMSE = 0.39 °C), the improved ECTemp model with a sigmoid observation function (ECTemp-S: RMSE = 0.40 °C), biphasic Kalman filter-based model (BKFB: RMSE = 0.48 °C), moving-average Kalman filter (MAKF: RMSE = 0.38 °C), and a standalone LSTM network (RMSE = 0.46 °C), RCAKF achieved the best accuracy with an RMSE of 0.31 °C. By augmenting the Kalman filter with a learned residual correction and adaptive uncertainty, the RCAKF framework significantly enhances core temperature tracking from a single heart rate signal. Its accuracy and reliance on a single, common sensor make it a practical and promising solution for real-time deployment on wearable devices for safety monitoring.
No takes yet. Share an insight, caveat, or question.
Zhao et al. (2025) conducted an observational in Thermal strain (n=22). Residual-Compensated Adaptive Kalman Filter (RCAKF) vs. Extended Kalman filter (EKF) was evaluated on Root mean squared error (RMSE) of core body temperature estimation. The Residual-Compensated Adaptive Kalman Filter (RCAKF) improved core body temperature estimation from heart rate, achieving an RMSE of 0.31 °C compared to 0.39 °C for the standard extended Kalman filter.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: