Auxiliary model maximum likelihood gradient‐based iterative identification for feedback nonlinear systems

Optimal Control Applications and Methods, Год журнала: 2024, Номер 45(5), С. 2346 - 2363

Опубликована: Июнь 17, 2024

Abstract This article considers the iterative identification problems for a class of feedback nonlinear systems with moving average noise. The model contains both dynamic linear module and static module, which brings challenges to identification. By utilizing key term separation technique, unknown parameters from modules are included in parameter vector. Furthermore, an auxiliary maximum likelihood gradient‐based algorithm is derived estimate parameters. In addition, stochastic gradient as comparison. numerical simulation results indicate that can effectively get more accurate estimates than algorithm.

Язык: Английский

---

Interpretability Research of Deep Learning: A Literature Survey

Information Fusion, Год журнала: 2024, Номер 115, С. 102721 - 102721

Опубликована: Окт. 9, 2024

Язык: Английский

---

Auxiliary Model‐Based Maximum Likelihood Multi‐Innovation Forgetting Gradient Identification for a Class of Multivariable Systems

Optimal Control Applications and Methods, Год журнала: 2025, Номер unknown

Опубликована: Янв. 29, 2025

ABSTRACT Through dividing a multivariable system into several subsystems, this paper derives the sub‐identification model. Utilizing obtained model, an auxiliary model‐based maximum likelihood forgetting gradient algorithm is derived. Considering enhancing parameter estimation accuracy, multi‐innovation (AM‐ML‐MIFG) proposed taking advantage of identification theory. Simulation results test effectiveness algorithms, and confirm that AM‐ML‐MIFG has satisfactory performance in capturing dynamic properties system.

Язык: Английский

---

Parameter Estimation and Model-free Multi-innovation Adaptive Control Algorithms

International Journal of Control Automation and Systems, Год журнала: 2024, Номер 22(11), С. 3509 - 3524

Опубликована: Ноя. 1, 2024

Язык: Английский

---

Parameter estimation methods for time‐invariant continuous‐time systems from dynamical discrete output responses based on the Laplace transforms

International Journal of Adaptive Control and Signal Processing, Год журнала: 2024, Номер 38(9), С. 3213 - 3232

Опубликована: Июль 3, 2024

Summary In industrial process control systems, parameter estimation is crucial for controller design and model analysis. This article examines the issue of identifying parameters in continuous‐time models. presents a stochastic gradient algorithm recursive least squares continuous systems. It derives identification linear systems based on Laplace transforms input output To prove that techniques given here work, we have included simulated example.

Язык: Английский

---

Hierarchical Least Squares Identification for the Multivariate Input Nonlinear Controlled Autoregressive Moving Average Systems

International Journal of Adaptive Control and Signal Processing, Год журнала: 2025, Номер unknown

Опубликована: Март 28, 2025

ABSTRACT This article presents a decomposition‐based least squares estimation algorithm for the multivariate input nonlinear system. By using hierarchical identification principle, breaks down system into two subsystems, one containing parameters of linear dynamic block and other static block. Treating unknown variables contained in information vector model is to replace them with outputs an auxiliary model. The comparative results between recursive developed this are provided test proposed algorithms have lower computational cost higher accuracy. Furthermore, convergence analyzed, which can guarantee stability algorithm. simulation confirm efficacy derived effectively estimating systems.

Язык: Английский

---

Identification for Precision Mechatronics: An Auxiliary Model‐Based Hierarchical Refined Instrumental Variable Algorithm

International Journal of Robust and Nonlinear Control, Год журнала: 2025, Номер unknown

Опубликована: Апрель 4, 2025

ABSTRACT When the physical properties of mechanical systems align with structure model, continuous‐time (CT) can be effectively represented by an interpretable and parsimonious additive formal models. This article addresses parameter estimation challenges CT autoregressive moving average (ACTARMA) systems. Based on maximum likelihood principle, optimality conditions for proposed identification algorithms are formulated ACTARMA Additionally, auxiliary model‐based hierarchical refined instrumental variable (AM‐HRIV) iterative algorithm AM‐HRIV recursive developed means principle model idea. These establish a pseudo‐linear regression relationship involving optimal prefilters derived from unified model. The effectiveness methods is demonstrated numerical simulation, performance method in identifying modal representations verified experimental data.

Язык: Английский

---

Joint State and Parameter Estimation for the Fractional‐Order Wiener State Space System Based on the Kalman Filtering

International Journal of Adaptive Control and Signal Processing, Год журнала: 2025, Номер unknown

Опубликована: Май 7, 2025

ABSTRACT This paper mainly investigates the joint estimation of parameters and states for fractional‐order Wiener state space model. Based on Kalman filter principle, a generalized recursive least squares algorithm with forgetting factor is proposed. In addition, filtering‐based presented, which reduces influence colored noise parameter estimation. A gradient identification introduced to estimate order fractional‐order. Under persistent excitation conditions, analysis indicates that proposed can system. simulation example given confirm algorithms are effective.

Язык: Английский

---

Multi-object tracking using score-driven hierarchical association strategy between predicted tracklets and objects

Image and Vision Computing, Год журнала: 2024, Номер unknown, С. 105303 - 105303

Опубликована: Окт. 1, 2024

Язык: Английский

---

A Novel Filtering Based Maximum Likelihood Generalized Extended Gradient Method for Multivariable Nonlinear Systems

International Journal of Adaptive Control and Signal Processing, Год журнала: 2024, Номер unknown

Опубликована: Дек. 25, 2024

ABSTRACT This study proposes a filtering based maximum likelihood generalized extended gradient algorithm for multivariable nonlinear systems with autoregressive moving average noises. The parameter estimates are obtained by minimizing the half squared residual measurement which can approach true values. An auxiliary model is also established measurable information of system, and output used to replace unmeasurable variables so that approximates these variables, as obtain consistent estimation system parameters. A derived comparison numerical example provided show effectiveness proposed method converge actual values quickly.

Язык: Английский

---