Robust multi-innovation full parameter identification for separable fractional-order systems based on online...

Area of Science:

• Control Systems Engineering
• Nonlinear Dynamics
• Signal Processing

Background:

• Estimating parameters and differential order in fractional-order systems is complex, especially with outlier-corrupted data.
• Existing robust principal component analysis methods are offline, limiting their use in real-time applications.

Purpose of the Study:

• To develop an online robust parameter estimation framework for fractional-order systems.
• To enable simultaneous real-time detection of outliers and adaptive parameter estimation.

Main Methods:

• Converted outlier detection into a matrix decomposition problem solved via a Sylvester equation for online information matrix recovery.
• Incorporated a multi-innovation strategy and a sliding window mechanism for efficient data utilization and real-time updates.
• Derived a robust multi-innovation gradient-based iterative (RMIGI) algorithm for simultaneous full parameter estimation.

Main Results:

• The proposed framework successfully detects outliers and estimates parameters online in real-time.
• The RMIGI algorithm demonstrated effectiveness and superiority in Monte Carlo simulations and a circuit case study.
• Theoretical analysis confirmed the convergence of the RMIGI method and characterized its computational complexity.

Conclusions:

• The developed online robust parameter estimation framework is suitable for real-time applications involving fractional-order systems.
• The RMIGI algorithm offers a significant advancement over existing offline methods for parameter identification in the presence of outliers.
• This work provides a robust and efficient solution for accurate system identification in challenging measurement conditions.