Simulating Motion Artifact Using an Autoregressive Model for Research in Biomedical Signal Quality Analysis

Area of Science:

• Biomedical Engineering
• Signal Processing
• Wearable Technology

Background:

• Motion artifact contamination significantly impacts biological signal interpretation.
• Existing methods for motion artifact mitigation rely on limited datasets or poorly described simulations.
• Developing diverse and realistic motion artifact simulations is essential for robust algorithm development.

Purpose of the Study:

• To develop and evaluate an autoregressive (AR) model for simulating motion artifact.
• To assess the model's ability to replicate key properties of experimental motion artifact data.
• To identify limitations of the AR model in capturing complex motion artifact characteristics.

Main Methods:

• Collected motion artifact data from 6 subjects walking at varying paces (slow, medium, fast).
• Developed an autoregressive (AR) model to simulate motion artifact based on collected data.
• Evaluated the simulated data against experimental data for time and frequency domain properties, morphology, and probability distribution.

Main Results:

• The AR model successfully replicated the mean, variance, and power spectrum of the experimental motion artifact data.
• The model demonstrated limitations in imitating the morphology and probability distribution (kurtosis) of the motion artifact.
• Kurtosis error ranged from 100.9% to 103.6%, indicating significant deviation from experimental data.

Conclusions:

• The developed autoregressive model provides a basis for simulating motion artifact but has limitations.
• More advanced modeling approaches are required to accurately capture the complexity of motion artifact morphology and distribution.
• Improved motion artifact simulations are necessary for the development of more effective signal processing algorithms.