THE GLOBAL LINEARIZATION OF NARMAX MODELS VIA ARMAX MODELS

Botao Qin,Zhengzhi Han

1997-01-01

Abstract:In this paper, a linearization method for complex and unknown nonlinear systems represented by NARMAX models is presented by using approximation ARMAX models. The limitation and construction method are discussed at the same time. The simulation result shows the efficiency of this method.

What problem does this paper attempt to address?

Nonlinear Identification for Control by Using NARMAX Models

Dan Stefanoiu,Janetta Culita,Andreea-Cristina Voinea,Vasilica Voinea

DOI: https://doi.org/10.3390/math12142252

IF: 2.4

2024-07-19

Mathematics

Abstract:The identification (and control) of nonlinear systems is one of the most important and actual research directions. Moreover, many systems are multivariable. Different from linear system identification (where only a few classes of models are available), in the case of nonlinear systems, the class set of models is quite diverse. One of the most appealing nonlinear models belongs to the nonlinear ARMAX (NARMAX) class. This article focusses on the identification of such a model, which can be compared with other models (such as nonlinear ARX (NARX) and linear ARMAX) in an application based on the didactical installation ASTANK2. The mathematical foundation of NARMAX models and their identification method are described at length within this article. One of the most interesting parts is concerned with the identification of optimal models not only in terms of numerical parameters but also as structure. A metaheuristic (namely, the Cuckoo Search Algorithm) is employed with the aim of finding the optimal structural indices based on a special cost function, referred to as fitness. In the end, the performances of all three models (NARMAX, NARX, and ARMAX) are compared after the identification of the ASTANK2 installation.

mathematics
Informative Conditions for Identification of MISO ARMAX Model in Closed-loop Systems

Wengang Yan,Yucai Zhu

DOI: https://doi.org/10.1016/j.ifacol.2018.09.187

2018-01-01

IFAC-PapersOnLine

Abstract:In system identification, a data set needs to be informative to ensure that the identification criterion has a unique global minimum asymptotically, and the parameter estimation is consistent. In this work, meaningful sufficient and necessary informative conditions for a data set in multi-input single-output (MISO) closed-loop systems are obtained for the identification of ARMAX model. Compared with the results in the existing literature, our theoretical results provide precisely quantitative conditions on the complexity of controllers to generate informative experiment without using external excitation. It is shown that the newly-proposed theoretical results are in agreement with the simulation results.
Multi-loop nonlinear internal model controller design under nonlinear dynamic PLS framework using ARX-neural network model

Bin Hu,Zhao Zhao,Jun Liang

DOI: https://doi.org/10.1016/j.jprocont.2011.09.002

IF: 3.951

2012-01-01

Journal of Process Control

Abstract:In this paper, a novel multi-loop nonlinear internal model control (IMC) strategy for multiple-input multiple-output (MIMO) systems is presented under the partial least squares (PLS) framework, which automatically decomposes the system into several univariate subsystems in the latent space. To formulate a nonlinear dynamic PLS framework, we propose an ARX-neural network (ARX-NN) cascaded structure, and incorporate it into PLS inner model. A gradient-based optimization approach is then provided to identify the parameter sets of the ARX-NN PLS model so that the plant-model mismatch is minimized. Furthermore, with perfect model, we show that the response of the closed loop system can be reduced to a simple linear IMC filter with the original system delay. The simulation results of a methylcyclohexane (MCH) distillation column from Aspen Dynamic Module, demonstrate the effectiveness of our approach in terms of disturbance rejection and tracking performance. © 2011 Elsevier Ltd. All Rights Reserved.
Modified bidiagonalization algorithm for NARMAX model identification

Xiao Wang,Chongzhao Han,Bai?u Wan

1997-01-01

Systems Analysis Modelling Simulation

Abstract:The main contribution of this paper is: (1) A new effective least squares algorithm is proposed, which is referred to as the Modified Bidiagonalization Least Squares (MBLS) one. Under the condition that round-off errors exist, the convergence and numerical stability of the MBLS algorithm has been discussed in detail, and a convergence proof has also been obtained. Due to its optimal mechanism, round-off errors have little influence to its convergence. The numerical results indicate that it is superior to the Singular Value Decomposition (SVD) method, which is well known for its well numerical stability; (2) Based on the MBLS and SVD algorithm, a new nonlinear system identification algorithm has been proposed in this paper. It selects the model terms directly from the assumed NARMAX full model, avoids the orthogonalization process which may be severely damaged by round-off errors, thus is numerically more stable than the orthogonal algorithms, such as the Classical Gram-Schmidt (CGS) algorithm, the Modified Gram-Schmidt (MGS) algorithm, the Householder Transformation (HT) and Givens method, etc. The simulation results indicate their efficiency.
System Identification Based on NARMAX Model Using Hopfield Networks

Shi Hong-li,Cai Yuan-li,Qiu Zu-lian

DOI: https://doi.org/10.1007/s11741-006-0122-6

2006-01-01

Journal of Shanghai University (English Edition)

Abstract:An approach is proposed to avoid model structure determination in system identification using NARMAX (nonlinear auto-regressive moving average with exogenous inputs) model. Identification procedure is formulated as an optimization procedure of a special class of Hopfield network in the proposed approach. The particular structure of these Hopfield networks can avoid the local optimum problem. Training of these Hopfield network achieves model structure determination and parameter estimation. Convergence of Hopfield networks guarantees that a NARMAX model of random initial state will approach a valid identification model with accurate state parameters. Results of two simulation examples illustrate that this approach is efficient and simple.
Control of Nonaffine Nonlinear Discrete-Time Systems Using Reinforcement-Learning-Based Linearly Parameterized Neural Networks

Qinmin Yang,Jonathan Blake Vance,S. Jagannathan

DOI: https://doi.org/10.1109/tsmcb.2008.926607

2008-01-01

Abstract:A nonaffine discrete-time system represented by the nonlinear autoregressive moving average with eXogenous input (NARMAX) representation with unknown nonlinear system dynamics is considered. An equivalent affinelike representation in terms of the tracking error dynamics is first obtained from the original nonaffine nonlinear discrete-time system so that reinforcement-learning-based near-optimal neural network (NN) controller can be developed. The control scheme consists of two linearly parameterized NNs. One NN is designated as the critic NN, which approximates a predefined long-term cost function, and an action NN is employed to derive a near-optimal control signal for the system to track a desired trajectory while minimizing the cost function simultaneously. The NN weights are tuned online. By using the standard Lyapunov approach, the stability of the closed-loop system is shown. The net result is a supervised actor-critic NN controller scheme which can be applied to a general nonaffine nonlinear discrete-time system without needing the affinelike representation. Simulation results demonstrate satisfactory performance of the controller.
Parameter identification based on adaptive density estimation

ZHONG Lu-sheng,SONG Zhi-huan

DOI: https://doi.org/10.3969/j.issn.1000-8152.2007.05.031

2007-01-01

Abstract:The problem of estimating parameters of ARMAX-model subject to unknown noise density distribution is addressed in this paper. Firstly, a method for adaptive nonparametric noise density estimation by using the Gaussian kernel is proposed. The global bandwidth and local bandwidth of the Gaussian kernel are dynamically computed according to the estimation error. Secondly, an iterative algorithm for parameter estimation is presented by minimizing the maximum likelihood function and the algorithm is based on the nonparametric estimate of the noise density. Thirdly, the convergence of the algorithm is analyzed and the convergent condition is given. Finally, simulation results show the proposed technique offers improved performance over existing methods when the noise density distribution is unknown.
An Adaptive Linear Element Based Method for Identification of Linear Parameter Varying Models

Fei Ye,Yong Sun,Qinmin Yang

DOI: https://doi.org/10.3182/20130902-3-cn-3020.00074

2013-01-01

IFAC Proceedings Volumes

Abstract:This paper is concerned with the identification of nonlinear processes which are parameterized as linear parameter varying (LPV) models. The LPV models are represented as blended linear models by using ADAptive LINear Element (ADALINE) structure. Firstly, ADALINE is employed to represent local models describing the system's behaviour at each fixed working point. Then, a novel method is proposed to estimate parameters of weight functions such that the global LPV model can be built by interpolating the local ones. The adaptation algorithm is discussed and extensive simulation studies are conducted to verify the effectiveness of this method.
Parameter Estimation of Arx/Narx Model: A Neural Network Based Method

D Wang,KY Lum,GH Yang

DOI: https://doi.org/10.1109/iconip.2002.1202794

2002-01-01

Abstract:In this paper, we consider the parameter-estimation problem of time domain models in the form of ARX or NARX. Instead of Least-Square based methods and the Maximum Likelihood method, the parameters of the time domain model will be estimated from test data using feed-forward neural networks. In the applications of neural networks, commonly used activation functions for function approximation are nonlinear transfer functions (e.g., Logsigmoid function, Hyperbolic tangent function and polynomial function) as well as linear transfer function. In our study, we investigate the equivalence of linear/nonlinear Autoregressive exogenous (ARX/NARX) models with feed-forward neural network models, in which the activation function of hidden neuron is one of these most frequently used functions. The coefficients of ARX/NARX models are given by neural network weight values for different activation functions. We show that for pure linear systems, ARX models can be obtained from feed-forward neural networks with commonly used hidden neuron activation functions such as log-sigmoid function, hyperbolic tangent function, polynomial function, exponential function as well as pure linear function. For nonlinear systems, ARX and NARX models can be obtained from neural networks with polynomial or exponential hidden neuron activation functions to describe the linear contribution and nonlinear contribution respectively.
Convergence of Recursive Identification for ARMAX Process with Increasing Variances

Jin Ya,Luo Guiming

DOI: https://doi.org/10.1016/s1007-0214(07)70006-8

2007-01-01

Abstract:The autoregressive moving average exogenous （ARMAX） model is commonly adopted for describing linear stochastic systems driven by colored noise. The model is a finite mixture with the ARMA component and external inputs. In this paper we focus on a parameter estimate of the ARMAX model. Classical modeling methods are usually based on the assumption that the driven noise in the moving average （MA） part has bounded variances, while in the model considered here the variances of noise may increase by a power of log n. The plant parameters are identified by the recursive stochastic gradient algorithm. The diminishing excitation technique and some results of martingale difference theory are adopted in order to prove the convergence of the identification. Finally, some simulations are given to show the reliability of the theoretical results.
NARMAX Model Identification Application in Dynamic Analysis of a Mooring Line

Heng-teng JI,Ju FAN,Xiang-lu HUANG

DOI: https://doi.org/10.3321/j.issn:1006-2467.2008.04.036

2008-01-01

Shanghai Jiaotong Daxue Xuebao/Journal of Shanghai Jiaotong University

Abstract:NARMAX model is a non-linear system model. A system model of a mooring line was constructed according to the NARMAX model theory. To system identification, the motion trace of the upper end of the mooring line is treated as the input of a dynamic system, and the time trace of the mooring line tension as the output of the system. The system expression obtained by system identification can be used to predict future tension trace of the mooring line. The predicted results by the NARMAX model were compared with the original data. It can be seen that this model has a good performance in predicting tension trace of a mooring line.
NARMAX Model Based on Wavelet Network with Application in Ship Control

Wenjun Zhang,Zhengjiang Liu,Wei Li

DOI: https://doi.org/10.1109/ccdc.2012.6243061

2012-01-01

Abstract:A wavelet neural network (WNN) is introduced to realize the system identification with nonlinear auto-regressive moving average with exogenous inputs (NARMAX) model. The method takes both advantages of the wavelet network and NARMAX model, and the multi-objective optimization method is used to real-time estimate the network coefficient. The wavelet-network-based NARMAX identification model is used as online system identifier, and the experimental result of ship course control proves the efficiency of the proposed neural identification model.
A Method to Model Nonlinear Systems by Neural Networks

Xifan Yao,Dongyuan Ge,Zhaotong Lian

DOI: https://doi.org/10.1109/icicic.2009.27

2009-01-01

Abstract:Many processes in reality exhibit nonlinear characteristics and in most of cases they cannot be treated satisfactorily using linearized approach in a large operating range. In this paper, an approximate approach is introduced to overcome the inaccuracy and inconsistency between the linearized model and the real process, due to linear representation of the nonlinear system, such as using Taylor series expansion by treating the nonlinear system as a linear uncertain system, that consists of a linear part, and an uncertain part. A neural network with Gaussian radial basis function in the hidden layer is employed to approximate the uncertain system. The approach can incorporate prior knowledge in its framework and provide a more transparent insight than the neural "black box" approach. The simulation results reveal that the proposed modeling approach to nonlinear systems is effective.
Integrated Identification of the Nonlinear Autoregressive Models With Exogenous Inputs (NARX) for Engineering Systems Design

Anastasia Kadochnikova,Yunpeng Zhu,Zi-Qiang Lang,Visakan Kadirkamanathan

DOI: https://doi.org/10.1109/tcst.2022.3171130

IF: 4.8

2022-12-31

IEEE Transactions on Control Systems Technology

Abstract:This brief presents a new framework for the identification of nonlinear autoregressive (AR) models with exogenous inputs (NARX) model for design (NARX-M-for-D), which represents NARX of engineering systems where the model coefficients are represented explicitly as a function of the physical parameters that can be adjusted for the system design. The framework is concerned with identifying a common structure of the NARX model which is shared by all design configurations, and with identifying the nonlinear static maps that link these design parameters with NARX coefficients. The problem of the common structure identification is solved via extended forward orthogonal regression, after which a joint regression problem is formulated to determine the explicit relationships between NARX coefficients and physical parameters for the system. Using sparse regression methods allows simultaneous detection of a compact structure of the NARX model and design parameter maps. The reduced structure improves model generalization in design parameter space which is instrumental for the application of the identified model in the system design. The performance of the framework is evaluated on a benchmark model and on the experimental data from dynamic testing of auxetic foams. An example of evaluating the output frequency response from the identified model demonstrates how the proposed framework can be used to assess dynamical properties of engineered systems in the design process.

automation & control systems,engineering, electrical & electronic
Identification of ARMAX based on genetic algorithm

SH He,XY Li,J Zhong

IF: 3.752

2002-01-01

Transactions of Nonferrous Metals Society of China

Abstract:On the basis of genetic algorithm, an intelligent search approach to determination of parameters of ARMAX (Autor Regressive Moving Average model with external input) processes was proposed. By representing the system with pole and zero pairs and repairing illegal chromosomes, the search space is limited to stable schemes. In calculation of objective function the "shifted data window" was designed, so that every input-output pair is used to guide the evolution and the "Data Saturation" is avoided. To prevent premature convergence, the adaptive fitness function was introduced, the conventional crossover and mutation operator was modified and the "catastrophic mutation" which is based on Metropolis mechanism was adopted. So the performance of convergence to the global optimum is improved. The validity and efficiency of proposed algorithm were illustrated by simulated results.
Augmented Space Linear Model.

Zhengda Qin,Badong Chen,Nanning Zheng,Jose C. Principe

DOI: https://doi.org/10.1109/tsp.2020.2987053

2018-01-01

Abstract:The linear model uses the space defined by the input to project the target or desired signal and find the optimal set of model parameters. When the problem is nonlinear, the adaption requires nonlinear models for good performance, but it becomes slower and more cumbersome. In this paper, we propose a linear model called Augmented Space Linear Model (ASLM), which uses the full joint space of input and desired signal as the projection space and approaches the performance of nonlinear models. This new algorithm takes advantage of the linear solution, and corrects the estimate for the current testing phase input with the error assigned to the input space neighborhood in the training phase. This algorithm can solve the nonlinear problem with the computational efficiency of linear methods, which can be regarded as a trade off between accuracy and computational complexity. Making full use of the training data, the proposed augmented space model may provide a new way to improve many modeling tasks.
Research on ARMAX energy modeling based on rough set

Guangjie Liu,Tao Xu,Fengsen Zhang

2013-01-01

Abstract:For modeling and forecasting of ARMAX energy process, we usually adopt nonlinear identification algorithm for modeling estimation, and then do orthogonal decomposition or take conditional expectation for forecasting according to fitting model. But nonlinear estimation algorithm is time-consuming and complex to calculate. This article tries to use estimation algorithm based on rough set for initial energy modeling: it simplifies modeling workload, it has the characteristics of real-time estimation and real-time forecasting, and it can meet the requirements of forecasting and modeling at the same time. © Sila Science.
Modeling for a Class of Nonstationary ARMA Series 1

Cao Xin,Sheng Zhaohan,Xu Nanrong

DOI: https://doi.org/10.1016/s1474-6670(17)51991-6

1990-01-01

IFAC Proceedings Volumes

Abstract:This paper is devoted to the parameter estimation problem for a class of nor-stationary ARMA series. From the viewpoint of instrumental variable method. the modified Yule-Walker estimator is extended to nonstationary situation. The strong consistence of the estimator is proved by means of martingale convergence theory
A Method for Nonlinear System Model in Fermentation Process

WANG Bin,WANG Su-nan,LIU Shu-guang

DOI: https://doi.org/10.3969/j.issn.1006-8341.2005.01.021

2005-01-01

Abstract:To solve the problem of modeling in the fermentation process, the NN-NARMAX (Neural Network Nonlinear Auto Regressive Moving Average) model for nonlinear system has been proposed. In the model the nonlinear mapping relation of NARMAX model is expressed with RBF (Radial Basis Function) neural network, the genetic algorithms is used to optimize the parameters (such as delay of NARMAX model, coefficients of filter). A group of contrasting experiments ,which are on modeling for temperature control in beer fermentation process by using NN-NARMAX model and Elman neural network respectively, have been done and the results show that NN-NARMAX model is more accurate than Elman neural network, and more capable in solving the problem of modeling the temperature control of fermentation process.
Nonlinear Additive ARX Models

Rong Chen,R. Tsay

DOI: https://doi.org/10.1080/01621459.1993.10476363

1993-09-01

Abstract:Abstract We consider in this article a class of nonlinear additive autoregressive models with exogenous variables for nonlinear time series analysis and propose two modeling procedures for building such models. The procedures proposed use two backfitting techniques (the ACE and BRUTO algorithms) to identify the nonlinear functions involved and use the methods of best subset regression and variable selection in regression analysis to determine the final model. Simulated and real examples are used to illustrate the analysis.

Mathematics

THE GLOBAL LINEARIZATION OF NARMAX MODELS VIA ARMAX MODELS

Nonlinear Identification for Control by Using NARMAX Models

Informative Conditions for Identification of MISO ARMAX Model in Closed-loop Systems

Multi-loop nonlinear internal model controller design under nonlinear dynamic PLS framework using ARX-neural network model

Modified bidiagonalization algorithm for NARMAX model identification

System Identification Based on NARMAX Model Using Hopfield Networks

Control of Nonaffine Nonlinear Discrete-Time Systems Using Reinforcement-Learning-Based Linearly Parameterized Neural Networks

Parameter identification based on adaptive density estimation

An Adaptive Linear Element Based Method for Identification of Linear Parameter Varying Models

Parameter Estimation of Arx/Narx Model: A Neural Network Based Method

Convergence of Recursive Identification for ARMAX Process with Increasing Variances

NARMAX Model Identification Application in Dynamic Analysis of a Mooring Line

NARMAX Model Based on Wavelet Network with Application in Ship Control

A Method to Model Nonlinear Systems by Neural Networks

Integrated Identification of the Nonlinear Autoregressive Models With Exogenous Inputs (NARX) for Engineering Systems Design

Identification of ARMAX based on genetic algorithm

Augmented Space Linear Model.

Research on ARMAX energy modeling based on rough set

Modeling for a Class of Nonstationary ARMA Series 1

A Method for Nonlinear System Model in Fermentation Process

Nonlinear Additive ARX Models