A Robust Kernel‐Based Global Identification Method for Hammerstein Parameters Varying System With Non‐Gaussian Output Noises
Authors:
- 31 August 2025
- journal article
- Published by Wiley in International Journal of Robust and Nonlinear Control
- Vol. 36 (2) , 830-844
Abstract
The Hammerstein parameters varying system (HPVS) with output noises is studied for dynamic processes in the paper. It has a cascaded structure of Hammerstein models, and its parameters are dependent on a time‐varying variable known as the scheduling variable, while also considering non‐Gaussian output noises. We present a robust kernel‐based global identification method (RKGIM) by using kernel methods and expectation maximization variational inference (EMVI) algorithm. Firstly, a Gaussian process (GP) with a radial basis kernel is considered to model the dependence of HPVS's parameters on scheduling variables, and a noise‐like term is introduced for numerical reasons. Their union designs a prior distribution for the system's noise‐free output, providing a possible description of the HPVS's output on scheduling variables. Then, to ensure the robustness of the identification, the measurement noise is described as a parametric Student's t distribution rather than the traditional Gaussian distribution. Furthermore, in the EMVI framework, the E‐step estimates the posterior estimations of the noise parameters and noise‐free output by using VI, while the M‐step estimates the hyperparameters that determine the aforementioned kernel by maximizing the likelihood. Finally, the effectiveness of the proposed method is demonstrated by simulation experiments.Keywords
Funding Information
- National Natural Science Foundation of China (61873208)
- National Natural Science Foundation of China (61573287)
- National Natural Science Foundation of China (61203234)
- National Natural Science Foundation of China (61135001)
- National Natural Science Foundation of China (61374023)
- Chinese Aeronautical Establishment (2016ZC53018)
- National Key Research and Development Program of China (H8634001AA0406)
This publication has 42 references indexed in Scilit:
- Data Filtering-Based Maximum Likelihood Gradient-Based Iterative Algorithm for Input Nonlinear Box–Jenkins Systems with Saturation NonlinearityCircuits, Systems, and Signal Processing, 2024
- Online identification of Hammerstein systems with B‐spline networksInternational Journal of Adaptive Control and Signal Processing, 2024
- Hierarchical gradient‐ and least‐squares‐based iterative estimation algorithms for input‐nonlinear output‐error systems from measurement information by using the over‐parameterizationInternational Journal of Robust and Nonlinear Control, 2023
- Two‐stage and three‐stage recursive gradient identification of Hammerstein nonlinear systems based on the key term separationInternational Journal of Robust and Nonlinear Control, 2023
- Filtered auxiliary model recursive generalized extended parameter estimation methods for Box–Jenkins systems by means of the filtering identification ideaInternational Journal of Robust and Nonlinear Control, 2023
- Feature extraction of mechanical faults from phase variation in transformer vibrationApplied Acoustics, 2022
- A new enabling variational inference model for approximating measurement likelihood in filtering nonlinear systemInternational Journal of Robust and Nonlinear Control, 2021
- Nonlinear output feedback and periodic disturbance attenuation for setpoint tracking of a combustion engine test benchAutomatica, 2016
- Refined instrumental variable methods for identification of LPV Box–Jenkins modelsAutomatica, 2010
- Identification of linear parameter varying modelsInternational Journal of Robust and Nonlinear Control, 2002