Adaptive parameter estimation for the expanded sandwich model.

Area of Science:

• Control Systems Engineering
• Nonlinear System Identification
• Signal Processing

Background:

• Expanded-sandwich systems are advanced nonlinear block-oriented models crucial for describing complex industrial processes.
• Traditional system identification methods often rely on prediction error output, limiting their applicability.
• Recent research highlights the need for more robust identification techniques for these systems.

Purpose of the Study:

• To propose a novel recursive identification algorithm for expanded-sandwich systems.
• To develop an adaptive estimator based on parameter identification error data.
• To offer a new design framework for system identification algorithms.

Main Methods:

• A recursive identification algorithm is developed using parameter identification error data.
• A filter is employed to extract system information with a minimalist structure.
• Intermediate variables are designed using filtered vectors to obtain identification error data.
• An adaptive estimator is established by integrating the derived identification error data.

Main Results:

• The proposed algorithm converges to true parameter values under general continuous excitation conditions.
• Experimental results demonstrate the effectiveness and utility of the novel identification method.
• The new approach provides a valuable alternative to traditional prediction error-based estimators.

Conclusions:

• The developed recursive identification algorithm offers a novel and effective approach for expanded-sandwich systems.
• The method's reliance on parameter identification error data provides a new perspective in algorithm design.
• The study validates the algorithm's performance through simulations and experimental data, confirming its practical applicability.