Improving heart variability and electrodermal activity measurements from wearable devices for fatigue monitoring

可穿戴技术在健康和健身跟踪方面越来越受到青睐，但确保它们提供的数据的可靠性，特别是在医疗背景下，对于正确的解释至关重要。该项目专注于两种电生理测量，即来自收集光体积描记(PPG)信号的光学传感器的心率变异性(HRV)和皮肤电活动(EDA)，这两种测量都可以深入了解自主神经系统的活动，包括交感和迷走神经反应。然而，这些信号容易因环境因素(如环境光线和温度)、生理因素(如年龄、性别和肤色)以及记录条件(包括身体位置和传感器放置)而发生变化。因此，PPG信号经常误判心跳，使得它们对于HRV分析是不可靠的，而HRV分析需要准确和连续的心跳检测。同样，EDA非常容易受到运动伪影的影响，需要进行细致的分析才能得出可靠的结论。该项目致力于采用稳健的信号处理和估计技术，从HRV和EDA测量中获得准确的生命和心血管数据，同时量化它们的不确定性。目前，可穿戴设备数据中缺乏伴随的不确定性估计，阻碍了对其的解读，突显了这一研究努力的重要性。这种全面的生理监测方法有望推进个性化医疗保健，促进健康状况的早期检测。

Wearable technologies are increasingly favoured for health and fitness tracking, yet ensuring the reliability of the data they provide, particularly in medical contexts, is paramount for a correct interpretation. This project focuses on two electrophysiological measurements, namely heart rate variability (HRV) derived from optical sensors collecting photoplethysmography (PPG) signals, and electrodermal activity (EDA), both offering insights into autonomic nervous system activity, including sympathetic and vagal responses. However, these signals are prone to variation due to environmental factors such as ambient light and temperature, physiological factors like age, sex, and skin colour, as well as recording conditions including body position and sensor placement. Consequently, PPG signals frequently misidentify heartbeats, rendering them unreliable for HRV analysis which necessitates accurate and continuous heartbeat detection. Similarly, EDA is highly susceptible to motion artifacts, demanding meticulous analysis for dependable conclusions. The project endeavours to employ robust signal processing and estimation techniques to derive accurate vital and cardiovascular data from HRV and EDA measurements, while also quantifying their uncertainties. Currently, the lack of accompanying uncertainty estimates in wearable device data impedes its interpretation, underscoring the significance of this research endeavour. This holistic approach to physiological monitoring holds promise for advancing personalized healthcare and facilitating early detection of health conditions.