SVM-based Embedded Soft Sensor System

Hu Shao

2005-01-01

Abstract:In the past, artificial neural networks such as radial basis function neural networks have been successfully used in soft sensor modeling. However, ANNs have many weakness compared with SVMs in regression analysis. An SVM-based soft sensor package via embedded system, which lays solid base for advanced control system is developed.

What problem does this paper attempt to address?

Research on SVM integration for intelligent sensor systen

Gui-long BIAN,Yi DING,Hai-bin SHEN

DOI: https://doi.org/10.13873/J.1000-9787(2014)08-0044-04

2014-01-01

Abstract:Support vector machine( SVM)as the representative of artificial intelligent techniques has been widely used in the intelligent sensor system,however,traditional SVM suffers from the catastrophic forgetting phenomenon,which results in loss of previously learned information,so it is unable to meet the requirements of real-time intelligent sensor system. The strength of Learn++ lies in its ability to learn new data without forgetting previously acquired knowledge,even when the new data introduce new classes. In order to solve the above problem,a Learn++ integration method based on hull vectors is proposed. Experimental results show that the algorithm not only has the ability of incremental learning,improve training speed and reduce the storage size,but also can ensure the classification precision,which meets the current demand of intelligent sensor systems for online learning.
Soft Sensor Modeling Based on Support Vector Machines Incremental Learning

ZHANG Ying,SU Hong-ye,CHU Jian

DOI: https://doi.org/10.3969/j.issn.1000-3932.2005.03.006

2005-01-01

Abstract:In order to overcome model failure problem,a soft sensor method based on support vector machines incremental learning is presented.Non-support vectors in working set are replaced by the samples which violate the Karush-Kuhn-Tucker conditions in the incremental set.The proposed soft sensor model is applied to predict the purity of para-xylene (PX) in a PX fractionation by adsorption process.Simulation results indicate that the proposed soft sensor model increases the adaptive abilities to various operation conditions and solves the model failure problem.
A Soft Sensor Modeling Approach Using Support Vector Machines

Rui Feng,Wei Shen,Huihe Shao

DOI: https://doi.org/10.1109/acc.2003.1240410

2003-01-01

Abstract:Artificial Neural Networks (ANNs) such as radial basis function networks (RBF NNs) have been successfully used in soft sensor modeling.. However, the generalization ability of conventional ANNs is not very well. For this reason, we present a novel soft sensor modeling approach based on Support Vector Machines (SVMs). Since standard SVMs have the limitation of speed and size in training large data set, we hereby propose Least Squares Support Vector Machines (LS_SVMs) and apply it to soft sensor modeling. Systematic analysis is performed and indicates that the proposed method provides satisfactory performance with excellent approximation and generalization property. Monte Carlo simulations show that our soft sensor modeling approach achieves superior performance to the conventional method based on RBF NNs.
A Novel Soft Sensor Modeling Approach Based on Least Squares Support Vector Machines

Song Chunlin,Zhang Yanzhu,Shao Huihe

DOI: https://doi.org/10.3772/j.issn.1006-6748.2004.04.009

2004-01-01

Abstract:Artificial Neural Networks (ANNs) such as radial basis function neural networks (RBFNNs) have been successfully used in soft sensor modeling. However, the generalization ability of conventional ANNs is not very well. For this reason, we present a novel soft sensor modeling approach based on Support Vector Machines (SVMs). Since standard SVMs have the limitation of speed and size in training large data set, we hereby propose Least Squares Support Vector Machines (LS_SVMs) and apply it to soft sensor modeling. Systematic analysis is performed and the result indicates that the proposed method provides satisfactory performance with excellent approximation and generalization property. Monte Carlo simulations show that our soft sensor modeling approach achieves performance superior to the conventional method based on RBFNNs.
Realization of SVM Soft Measurement Platform Based on Embedded System

WANG Jian-wei,SONG Zhi-huan

DOI: https://doi.org/10.3969/j.issn.1671-7848.2011.02.036

2011-01-01

Abstract:To the soft measurement problem in complex petrochemical industrial, a soft measurement platform is designed based on OMAP-L137 (an ARM and floating point DSP double-core processor). For real-time purpose, by analyzing the penalty factors of sup- port vector machine SVM, gamma of RBF, epsilon-insensitive and stop criterion of training, the optimized SVM is realized in the double-core platform that supports on-line training of soft measurement model. The powerful and easy-to-use instrument is convenient for real-time detection and dynamic display. The testing of datasets demonstrates the excellent high-precision and real-time performance of the instrument. The platform provides a low-cost, portable and real-time solution for the soft measurement of complex petrochemical industrial process.
SOFT SENSOR MODELING BASED ON SUPPORT VECTOR MACHINE

Rui Feng

IF: 1.24

2002-01-01

Information and Computation

Abstract:Support vector machine (SVM) is a new learning machine based on the statistical learning theory. This paper presents a soft sensor model based on the SVM. Theoretical and simulation analysis indicates that this method features high learning speed, good approximation, well generalization ability, and little dependence on the sample set. It has the better performance than the soft sensor modeling based on the RBF neural network.
Soft Sensor Modeling Based on Multiple Support Vector Machines

YUAN Ping,MAO Zhi-zhong,WANG fu-li

DOI: https://doi.org/10.3969/j.issn.1004-731x.2006.06.012

2006-01-01

Abstract:Soft sensor which solved the difficult problem of measuring the un-measurable output variables has been widely used in industrial process control.The core problem of soft sensor is to construct an appropriate mathematic model.Support vector machine(SVM)which has high generalization is adopted to establish soft sensor models.Based on the idea that the accuracy of model could be significantly improved by combining several sub-models,a multiple support vector machine(MSVM)modeling approach was proposed to build the soft sensor model.In this method,subtractive clustering was adopted to divide the input space into several sub-spaces,and sub-models were built by Least Square SVM(LS SVM)in every sub-space.In order to minimize the severe correlation among sub-models,to improve the accuracy and robustness of the model,the sub-models were combined by principal components regression(PCR).The software sensor model accuracy is perfectly improved.The simulation results demonstrate the efficiency of the method.
Soft-Sensor Modeling Method Based On Support Vector Machines

Mg Zhang,Ww Yan

2005-01-01

Abstract:Soft-sensor has been widely used in industrial process control to improve the quality of product and assure safety in production. The core problem of soft-sensor is to construct appropriate mathematic model. Support vector machines (SVM) is a novel machine learning method based on structure risk minimizing principle, which is powerful for the problem characterized by small sample, nonlinearity, high dimension and local minima, and has high generalization. In this paper, soft-sensor modeling method and algorithm based on Support Vector Machines (SVM) is proposed, and cross validation method is used to select hyper-parameters of SVM model. Soft-sensor model based on SVM is applied to estimation nonlinear time-vary system. Experiment results show that SVM provides a new effective method for soft-sensor modeling and has promising application in industrial process applications.
Application of Soft Sensor Based on Embedded System ARM

Hu Shao

2008-01-01

Abstract:A universal software of soft sensor on Windows CE of ARM embedded system is implemented.The ARM embedded system is used as hardware platform,and the software is programmed with Embedded visual C++ on personal computer,and a double RBF neural network is used as the mathematical model of soft sensor.The software adopts the object oriented architecture,and can be compartmentalized as six modules,such as system attemperation modules etc.according to the principle of clarity.This method is used to estimate the concentration of carbinol in modeling data and good results are obtained.
Soft Sensor Based on Embedded System and Its Application

Li Xiaofeng,Shi Jinfei,Yan Weiwu

DOI: https://doi.org/10.3969/j.issn.1000-0380.2008.08.006

2008-01-01

Abstract:The design and developing frame of the soft sensor based on embedded system are introduced,and the general data structure of soft sensor software package is proposed.The double RBF neural network is used as the core algorithm of soft sensing and two-phase neural network training is adopted to reduce the error of soft sensing.With the application for FCC as example,the method of on-line soft sensing of freezing point of light diesel by using this system is presented.The system features generalization,open,and real-time performance,it can be widely used in various situations for offering high applicable value and bright future.
Soft Sensors of Nonlinear Industrial Processes Based on Self-Learning Kernel Regression Model

Chihang Wei,Junghui Chen,Zhihuan Song,Chun-I Chen

DOI: https://doi.org/10.1109/ascc.2017.8287444

2017-01-01

Abstract:The traditional kernel data-driven regression methods would not guarantee a good prediction performance as the kernel space of the traditional kernel function is defined without considering the regression relationships between the output variables and the input variables. This paper proposes a self-learning kernel regression model. It learns a specific kernel space from input and output data, simultaneously considering the manifold structure of the input variables for the dimension reduction and the regression relationship between the projected input variables and the output variable. The model is casted to an instance of semi-definite programming with the newly designed objective function. A data-dependent kernel is then utilized to scale kernel function approximation and it can be applied on-line whenever a new data point is available. Two case studies, including a numerical example and a real semiconductor process, are provided to demonstrate the effectiveness of the proposed algorithm.
Soft Sensor Modeling Based on PCA and Support Vector Machines

ZHENG Xiao-xia,QIAN Feng

DOI: https://doi.org/10.3969/j.issn.1004-731x.2006.03.052

2006-01-01

Abstract:Soft sensor is an effective method to estimate variables which are difficult to be measured on-line in industrial processes.Support vector machine(SVM)is a novel machine learning method based on the statistical learning theory.A soft sensor based on principal component analysis(PCA)and Least Square SVM was proposed.The PCA method could not only solve the linear correlation of the input and compress data but also simply the SVM structure.Cross validation method was used to select parametrs of LS-SVM model.Soft sensor was applied to prediction of 4-CBA.Results indicates that this method features high learning speed,good approximation and well generalization ability.It provides convenice for on-line 4-CBA measurement.
Soft Sensor Modeling Based on KPCA and Least Square SVM

XU Ye,DU Wen-li,QIAN Feng

DOI: https://doi.org/10.3969/j.issn.1004-731x.2007.17.002

2007-01-01

Abstract:Soft sensor is necessary for industrial process control and analysis, and the core problem is how to construct appropriate model having fast convergence speed and good generalization performance. A kind of soft sensor method was proposed based on kernel principle component analysis (KPCA) and least square support vector machine (LSSVM). KPCA was applied to choose the nonlinear principal component of the model input data space, and LSSVM was applied to proceed regression modelling, which could not only reduce the complexity of calculation but could improve the generalization ability. The proposed KPCA-LSSVM was applied to predict the granularity of PTA. Simulation indicates that this method features high learning speed, good approximation and good generalization ability compared with SVM and PCA-SVM, and is proved to be an efficient modeling method.
Nonlinear Soft Sensor Development Based on Relevance Vector Machine

Zhiqiang Ge,Zhihuan Song

DOI: https://doi.org/10.1021/ie101146d

2010-01-01

Industrial & Engineering Chemistry Research

Abstract:This paper proposes an effective nonlinear soft sensor based on relevance vector machine (RVM), which was originally proposed in the machine learning area. Compared to the widely used support vector machine (SVM) and least-squares support vector machine (LSSVM) based soft sensors, RVM gives a more sparse model structure, which can greatly reduce computational complexity for online prediction. While SVM/LSSVM can only provide a point estimation of the prediction result, RVM gives a probabilistic prediction result, which is more sophisticated for the soft sensor application. Furthermore, RVM can successfully avoid several drawbacks of the traditional support vector machine type method, such as kernel function limitation, parameter tuning complexity, and etc. Due to the advantages of RVM, a practical application of this method is made for soft sensor modeling in this paper. To evaluate the performance of the developed soft sensor, two case studies are demonstrated, which both support that RVM performs much better than other methods for soft sensing.
An approach to soft sensor modeling based on wavelets and a least square support vector machine

XIANG Zheng-rong,CHEN Qing-wei

DOI: https://doi.org/10.3969/j.issn.1673-4785.2010.01.010

2010-01-01

Abstract:Some industrial process variables are very difficult to measure.To overcome this problem,a soft sensor modeling,based on wavelets and a least square support vector machine (LS-SVM),was proposed.Initially,a stream of sample data was decomposed into sub-sequences with different frequences.This was done on the basis of wavelet transform.Then the respective sub-sequences were modeled by appropriate SVMs.Finally,estimated values for the primary variables were obtained by wavelet reconstruction.A quantum particle swarm optimization (QPSO) algorithm was employed to select parameters for the LS-SVM and the kernel function.Simulation results confirmed that the proposed method has high precision and good generalization ability.
Soft Sensor Modeling Based on the Soft Margin Support Vector Regression Machine

Tao Ye,Xuefeng Zhu,Daoping Huang,Xiangyang Li

DOI: https://doi.org/10.1109/icca.2007.4376464

2007-01-01

Abstract:This paper focuses on regression applications of the Support Vector Machine (SVM) in the process industry. The support vector regression machines are employed to build soft sensing models in the paper. Soft sensor modeling, in a sense, is a kind of regression problems in industrial processes. First we review the development history of the Vapnik Chervonenkis (VC) theory and SVM. And then, the basic idea behind the SVM is introduced and some famous SVM regression algorithms are talked about. After that, the standard QP and SMO implementations to Vapnik's soft margin epsiv-SVM regression algorithm are discussed in detail. Using these two implementing methods, we perform some experiments, to predict pulp Kappa numbers, over a real-life dataset retrieved from a kraft pulp cooking process. Some useful conclusions are drawn finally.
SVM Based Soft Sensor for Antibiotic Fermentation Process

WM Zhong,DY Pi,YX Sun

DOI: https://doi.org/10.1109/icsmc.2003.1243808

2003-01-01

Abstract:In this paper, we introduce a SVM based soft sensor method for resultant concentration of antibiotic fermentation. Non-linear black-box models for resultant concentration are established used SVM with poly and Gaussian kernels respectively. Some treatments such as the outliers' elimination and data smoothness are done before modeling. Simulation results with Matlab toolbox show that the SVM with poly and Gaussian kernels both can do the soft sensor work. Compare to BP neural network, the SVM with Gaussian kernel based soft sensor has better performance in generalization ability and training speed. According to the analysis, the method of SVM based soft sensor for resultant concentration is feasible.
Supervised Nonlinear Dynamic System for Soft Sensor Application Aided by Variational Auto-Encoder

Bingbing Shen,Zhiqiang Ge

DOI: https://doi.org/10.1109/tim.2020.2968162

IF: 5.6

2020-01-01

IEEE Transactions on Instrumentation and Measurement

Abstract:Dynamic data modeling has been attracting much attention from researchers and has been introduced into the probabilistic latent variable model in the process industry. It is a huge challenge to extend these dynamic probabilistic latent variable models to nonlinear forms. In this article, a supervised nonlinear dynamic system (NDS) based on variational auto-encoder (VAE) is introduced for processes with dynamic behaviors and nonlinear characteristics. Based on the framework of VAE, which has a probabilistic data representation and a high fitting ability, the supervised NDS can extract effective nonlinear features for latent variable regression. The feasibility of the proposed supervised NDS is tested on two numerical examples and an industrial case. Detailed comparisons verify the effectiveness and superiority of the proposed model.
Soft sensor modeling via support vector regression with KICA-based feature extraction

Yihua Liu,Hongying Deng,Zengliang Gao,Ping Li

DOI: https://doi.org/10.4028/www.scientific.net/AMR.186.560

2011-01-01

Abstract:A novel two-level integrated soft sensor modeling method using kernel independent component analysis (KICA) and support vector regression (SVR) is proposed for chemical processes. In the first level, the KICA approach is adopted to extract information of input variables in the high dimensional feature space. Based on this strategy, the correlation of input variables can be eliminated and thus the complexity is reduced. Then, the model is established using SVR in the second level. The KICA-SVR soft sensor modeling method is applied to estimate product compositions in the Tennessee Eastman process. The obtained results show that it can exhibit better performance, compared to the traditional ICA, principal component analysis (PCA) and kernel PCA based information extraction methods, under different operating conditions.
Simulation and Implement of RBFNN Soft Sensor Based on Embedded System

Hu Shao

2006-01-01

Abstract:Implement problem of the RBF neural network soft sensor on WinCE embedded system is discussed.By using K-means procedure,the cluster center is established and recursion algorithm is given to train the network weight value.A Double network structure is built to improve the network accuracy.One is used to train sample data and the other is used to train network error.To implement this soft sensor instrument, industrial embedded computer is used for hardware platform,and the software is programmed with Embedded visual C++.This method is used to estimate the freeze point of diesel oil in modeling data and good results are obtained.

SVM-based Embedded Soft Sensor System

Research on SVM integration for intelligent sensor systen

Soft Sensor Modeling Based on Support Vector Machines Incremental Learning

A Soft Sensor Modeling Approach Using Support Vector Machines

A Novel Soft Sensor Modeling Approach Based on Least Squares Support Vector Machines

Realization of SVM Soft Measurement Platform Based on Embedded System

SOFT SENSOR MODELING BASED ON SUPPORT VECTOR MACHINE

Soft Sensor Modeling Based on Multiple Support Vector Machines

Soft-Sensor Modeling Method Based On Support Vector Machines

Application of Soft Sensor Based on Embedded System ARM

Soft Sensor Based on Embedded System and Its Application

Soft Sensors of Nonlinear Industrial Processes Based on Self-Learning Kernel Regression Model

Soft Sensor Modeling Based on PCA and Support Vector Machines

Soft Sensor Modeling Based on KPCA and Least Square SVM

Nonlinear Soft Sensor Development Based on Relevance Vector Machine

An approach to soft sensor modeling based on wavelets and a least square support vector machine

Soft Sensor Modeling Based on the Soft Margin Support Vector Regression Machine

SVM Based Soft Sensor for Antibiotic Fermentation Process

Supervised Nonlinear Dynamic System for Soft Sensor Application Aided by Variational Auto-Encoder

Soft sensor modeling via support vector regression with KICA-based feature extraction

Simulation and Implement of RBFNN Soft Sensor Based on Embedded System