ECG denoising using angular velocity as a state and an observation in an Extended Kalman Filter framework

Area of Science:

• Biomedical Engineering
• Signal Processing
• Computational Medicine

Background:

• Synthetic electrocardiogram (ECG) signal generation often requires robust filtering techniques.
• Existing methods may not fully capture the dynamic characteristics of ECG signals.
• The Extended Kalman Filter (EKF) is a powerful tool for state estimation in nonlinear systems.

Purpose of the Study:

• To propose an efficient filtering procedure for synthetic ECG signals using a modified nonlinear dynamic model.
• To incorporate the angular velocity of the ECG signal as a state within the EKF framework.
• To evaluate the performance of the proposed filtering method on a standard ECG database.

Main Methods:

• Development of a modified nonlinear dynamic model for ECG signal generation.
• Implementation of an Extended Kalman Filter (EKF) incorporating ECG angular velocity as a state.
• Consideration of two observation equation scenarios: with and without direct observation of angular velocity.
• Quantitative evaluation using the MIT-BIH Normal Sinus Rhythm Database (NSRDB).

Main Results:

• The proposed EKF-based filtering procedure demonstrates significant performance.
• An average Signal-to-Noise Ratio (SNR) improvement of 8 dB was achieved for an input signal with -4 dB SNR.
• The inclusion of angular velocity as a state in the EKF proved effective.

Conclusions:

• The proposed EKF-based filtering method offers an efficient approach for enhancing synthetic ECG signals.
• The method effectively utilizes the dynamic properties of ECG signals, including angular velocity.
• The results indicate a substantial improvement in signal quality, making it valuable for ECG analysis and research.