Runtime Monitoring and Statistical Approaches for Correlation Analysis of ECG and PPG
This work addresses the need for improved accuracy and robustness in health monitoring systems, such as through sensor fusion and attack detection, but it is incremental as it builds on existing methods without introducing a new paradigm.
The paper tackled the problem of establishing precise correlations between ECG and PPG signals, which are used separately for cardiac monitoring, by presenting the first formal approach combining runtime monitoring and statistical analysis to define key relationships.
Biophysical signals such as Electrocardiogram (ECG) and Photoplethysmogram (PPG) are key to the sensing of vital parameters for wellbeing. Coincidentally, ECG and PPG are signals, which provide a "different window" into the same phenomena, namely the cardiac cycle. While they are used separately, there are no studies regarding the exact correction of the different ECG and PPG events. Such correlation would be helpful in many fronts such as sensor fusion for improved accuracy using cheaper sensors and attack detection and mitigation methods using multiple signals to enhance the robustness, for example. Considering this, we present the first approach in formally establishing the key relationships between ECG and PPG signals. We combine formal run-time monitoring with statistical analysis and regression analysis for our results.