Robust recursive estimation for the errors-in-variables nonlinear systems with impulsive noise.

Area of Science:

• Control Systems Engineering
• Signal Processing
• Nonlinear System Identification

Background:

• Non-Gaussian external disturbances present significant challenges in system modeling and identification.
• Impulsive noise corrupts data, complicating accurate estimation in nonlinear systems.

Purpose of the Study:

• To develop robust recursive estimation algorithms for errors-in-variables nonlinear systems affected by impulsive noise.
• To enhance system modeling and identification accuracy in the presence of non-Gaussian disturbances.

Main Methods:

• Formulation of a robust recursive algorithm by minimizing a continuous logarithmic mixed p-norm criterion.
• Development of bias correction techniques for estimating nonlinear monomials from noisy input data.
• Derivation of a hierarchical estimation algorithm using the continuous logarithmic mixed p-norm to reduce computational complexity.

Main Results:

• The proposed algorithms provide robust estimations against impulsive noise via adjustable weight gain.
• Recursive expressions with bias correction effectively estimate nonlinear monomials.
• A hierarchical approach successfully reduces computational loads while maintaining robustness.

Conclusions:

• The developed algorithms are feasible and effective for robust estimation in nonlinear systems with impulsive noise.
• The continuous logarithmic mixed p-norm criterion offers a robust approach to handling non-Gaussian disturbances.
• The hierarchical algorithm provides a computationally efficient solution for robust system identification.