Comparison of algorithms for tracking short-term changes in arterial circulation parameters

Area of Science:

• Biomedical Engineering
• System Identification
• Physiology

Background:

• The systemic arterial bed exhibits complex viscoelastic properties.
• Accurate tracking of these properties is crucial for understanding cardiovascular dynamics.
• Rapid parameter changes in biological systems pose identification challenges.

Purpose of the Study:

• To evaluate three recursive methods for tracking rapidly changing viscoelastic properties of the systemic arterial bed.
• To compare the performance and efficiency of Recursive Least Squares (RLS), Least Squares with Variable Forgetting Factor (LSVF), and a Constant Forgetting Factor with Covariance Modification (CFCM) algorithms.
• To assess the sensitivity of these methods to design variables using simulated noisy data.

Main Methods:

• Application of RLS, LSVF, and CFCM algorithms within a unified framework.
• System identification of viscoelastic properties using noisy computer simulation data.
• Investigation of parameter sensitivity and tracking accuracy for each method.

Main Results:

• All analyzed methods demonstrated satisfactory tracking of rapid changes in peripheral resistance.
• LSVF offered slightly better performance than RLS and is suitable when calculation efficiency is prioritized.
• CFCM exhibited the best tracking ability, even with varying noise sequences, while maintaining simplicity.

Conclusions:

• Recursive methods, particularly LSVF and CFCM, are effective for identifying systems with rapidly changing parameters like the arterial bed.
• LSVF provides a good balance between performance and computational efficiency.
• CFCM demonstrates superior tracking performance in the presence of noise, making it a robust choice for complex physiological systems.