Comparative Analysis of Kalman Filtering and Machine Learning Based Cardiovascular Signal Processing Algorithms

Cardiovascular (CV) disorder is one of the critical health problem that cause economical and social impacts, even death to lots of peoples globally. Electrocardiogram (ECG) signal is the signal taken from the human body to study the status of CV and heart conditions. Before the introduction of computers, those tasks were done by the experts that cause various mistakes. Currently, the use of advancing signal processing devices manage to reduce these effects. Besides, it allows to develop various signal detection and parameter estimation algorithms. By studying the parameters of ECG signals, it is possible to determine whether the person is in critical condition or not. This helps to take an appropriate action. In the last decades, both classical and machine learning methods have been used to study and characterize the essential properties and parameters of CV signals.

In this work, we study different algorithms that are useful for ECG based CV parameters estimation. We evaluate the performance of both classical (Kalman Filtering) and machine learning algorithms with Butterworth low pass filter, wavelet transform and linear regression for parameter estimation. Besides, we proposed an algorithm that combines adaptive Kalman filter (AKF) and discrete wavelet transform (DWT). In this algorithm, the ECG signal is filtered using AKF. Then segmentation is performed and features are extracted using DWT algorithm. Numerical simulation is done to validate the performances of these algorithms. The results show that the proposed algorithm gives better performance than Kalman filtering and has nearly the same performance with machine learning methods.