检索结果-内蒙古大学图书馆

A DATA-DRIVEN FAULT DETECTION APPROACH WITH PERFORMANCE OPTIMIZATION

CANADIAN JOURNAL OF CHEMICAL ENGINEERING 2018年第2期96卷 507-514页

作者： Li, Linlin Ding, Steven X. Peng, Kaixiang Han, Huayun Yang, Ying Yang, Xu Univ Sci & Technol Beijing Sch Automat & Elect Engn Key Lab Knowledge Automat Ind Proc Minist Educ Beijing 100083 Peoples R China Univ Duisburg Essen Inst Automat Control & Complex Syst AKS D-47057 Duisburg Germany Peking Univ Dept Mech & Engn Sci Coll Engn State Key Lab Turbulence & Complex Syst Beijing 100871 Peoples R China

This paper is concerned with the data-driven realization of fault detection approach with performance optimization. For our purpose, the datadriven realization form of linear kernel representations is studied first, which is essential in our work. It is followed by a data-driven realization of kernel representation and its implementation in the design scheme of fault detection systems. Nevertheless, the basic idea behind this approach lies in the one-step identification of kernel representation using LQ-decomposition. Then, the recursive kernel representation is introduced and the so-called gradient descent algorithm is applied to optimize the performance of the proposed data-driven fault detection system. The effectiveness of the proposed approaches is illustrated by a numerical example and a case study on laboratory setup of a three-tank system.

关键词： data-driven methods observer-based fault detection gradient descent algorithm

来源：评论

学校读者我要写书评

暂无评论

Efficient classification algorithm and a new training mode for the adaptive radial basis function neural network equaliser

引用

IET COMMUNICATIONS 2012年第2期6卷 125-137页

作者： Assaf, R. El Assad, S. Harkouss, Y. Zoaeter, M. Team Image Site Nantex HDR IETR Lab Intitut Elect & Telecommun Rennes IETR UMR CNRS 6164 Rennes France

The study presents a new classification algorithm and a new online training mode used for learning the parameters of a Bayesian RBFNN (radial basis function neural network) equaliser in a non-linear time-varying channel. The classification algorithm is used to determine the centres of the hidden layer neurons that are equal to the channel states. This proposed unsupervised classification algorithm is based on both the K-means and the rival penalised competitive algorithms. Its main advantage is neither an initialisation phase nor a knowledge of the channel states number is required. The connections of weights and the spread of the hidden neurons are learned by the gradient descent algorithm, which applies a new proposed training mode. This training mode combines the advantages of both the online and the offline training modes such as stability and good speed of convergence. The performances of the RBFNN equaliser trained by the proposed method are shown in comparison with the performances of the optimal Bayesian equaliser and those of the same equaliser trained by other known training modes. All these performances are studied by using different types of channels.

关键词： hidden layer neuron online training mode Communications computing Neural computing techniques offline training mode optimal Bayesian adaptive RBFNN equaliser gradient descent algorithm Knowledge engineering techniques adaptive equalisers belief networks time-varying channels parameter learning radial basis function networks gradient methods optimal adaptive Bayesian radial basis function neural network equaliser telecommunication computing unsupervised classification algorithm unsupervised learning K-means algorithm Interpolation and function approximation (numerical analysis) nonlinear time-varying channel Communication channel equalisation and identification rival penalised competitive algorithm

来源：评论

学校读者我要写书评

暂无评论

AdaRW training optimization algorithm for deep learning model of marine target detection based on SAR

引用

INTERNATIONAL JOURNAL OF REMOTE SENSING 2022年第1期43卷 120-131页

作者： Li, Wanwu Liu, Lin Zhang, Jixian Shandong Univ Sci & Technol Coll Geodesy & Geomat Qingdao 266590 Peoples R China Natl Qual Inspect & Testing Ctr Surveying & Mappi Beijing Peoples R China

Deep learning adjusts parameters to optimize the model through model training. Training algorithm is the key to model optimization and implementation. Therefore, the improvement of model training algorithm is of great significance to deep learning. Based on the original gradient algorithm, the paper proposes a new gradient descent algorithm AdaGrad Restricted by Windows (AdaRW) for Deep learning model training optimization. Aiming at the defects of AdaGrad algorithm, the new algorithm uses the subset of window to limit historical accumulation, so as to slow down the attenuation of learning rate, and improve the speed of model training. The paper constructs OceanTDA9, a Deep learning model of marine target detection for Synthetic Aperture Radar (SAR) data, and adopts the proposed AdaRW algorithm to train the model based on SAR data with a resolution of 10 m in the Bohai Sea. Experiment shows that the accuracy and loss of the algorithm are better than those of AdaGrad and Stochastic gradient descent (SGD) algorithms, and the standard deviation is better than that of Adam algorithm.

关键词： Marine target detection Deep learning model gradient descent algorithm AdaGrad algorithm

来源：评论

学校读者我要写书评

暂无评论

Filtering-Based Accelerated Estimation Approach for Generalized Time-Varying Systems With Disturbances and Colored Noises

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS 2023年第1期70卷 206-210页

作者： Ji, Yan Jiang, Anning Qingdao Univ Sci & Technol Coll Automat & Elect Engn Qingdao 266061 Peoples R China

This brief studies the parameter estimation problem of a generalized time-varying system by using a novel gradient descent algorithm. The time-varying parameters are estimated by the accelerating gradient descent algorithm under considering the measurable disturbance factors. The proposed algorithm increases the convergence rate by adding a momentum term to the correction of the parameter updated estimates. In addition, a data filter is introduced to filter the input and output data, so as to eliminate the interference of colored noise to the accuracy of parameter estimation. A numerical example shows the effectiveness of the proposed algorithm.

关键词： Generalized time-varying system gradient descent algorithm parameter estimation

来源：评论

学校读者我要写书评

暂无评论

RAPIDO: a rejuvenating adaptive PID-type optimiser for deep neural networks

引用

ELECTRONICS LETTERS 2019年第16期55卷 899-901页

作者： Kim, S. Park, D. J. Chang, D. E. Korea Adv Inst Sci & Technol Sch Elect Engn Daejeon 34141 South Korea

The authors present a novel gradient descent algorithm called RAPIDO for deep learning. It adapts over time and performs optimisation using current, past and future information similar to the PID controller. The proposed method is suited for optimising deep neural networks that consist of activation functions such as sigmoid, hyperbolic tangent and ReLU functions because it can adapt appropriately to sudden changes in gradients. They experimentally study the authors' method and show the performance results by comparing with other methods on the quadratic objective function and the MNIST classification task. The proposed method shows better performance than the other methods.

关键词： pattern classification optimisation learning (artificial intelligence) neural nets gradient methods three-term control RAPIDO rejuvenating adaptive PID-type optimiser deep neural networks deep learning PID controller activation functions sigmoid functions hyperbolic tangent functions ReLU functions quadratic objective function gradient descent algorithm MNIST classification task

来源：评论

学校读者我要写书评

暂无评论

MODELING AND PREDICTIVE CONTROL USING HYBRID INTELLIGENT TECHNIQUES FOR A NONLINEAR MULTIVARIABLE PROCESS

引用

INSTRUMENTATION SCIENCE & TECHNOLOGY 2011年第2期39卷 211-230页

作者： Subathra, B. Radhakrishnan, T. K. Natl Inst Technol Dept Chem Engn Tiruchirappalli 620015 Tamil Nadu India

A recurrent neuro fuzzy network (RNFN) model-based multistep ahead predictive control strategy is proposed in this article. The fuzzy logic (FL) and neural networks (NN) are intelligent system approaches, and they complement each other. Hybridization of FL and NN utilizes the concepts of human cognitive capabilities and biological systems, respectively. Dynamic processes necessitate past information about the process input/output variables. In order to store the information, a memory unit is introduced between the fuzzy inference layer and the fuzzification layer. This recurrent structure enhances the prediction capability;hence, this RNFN model can be used to develop the multistep ahead predictive controller. The objective function of model based controller (MPC) minimizes the future control moves. The gradient descent (GD) algorithm is used to the optimize control moves. The proposed RNFN model is used to develop a model predictive controller. The performance of the RNFN-MPC is compared with that of a neuro fuzzy network (NFN)-based MPC for a laboratory scale quadruple tank process.

关键词： gradient descent algorithm hybrid soft computing intelligent techniques modeling MPC recurrent neuro fuzzy network

来源：评论

学校读者我要写书评

暂无评论

Standalone CMAC Control System With Online Learning Ability

引用

IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS PART B-CYBERNETICS 2010年第1期40卷 43-53页

作者： Yeh, Ming-Feng Tsai, Cheng-Hung Lunghwa Univ Sci & Technol Dept Elect Engn Tao Yuan 33327 Taiwan Univ Sci & Technol China Dept Elect Engn Taipei 11581 Taiwan

A cerebellar model articulation controller (CMAC) control system, which contains only one single-input controller implemented by a differentiable CMAC, is proposed in this paper. In the proposed scheme, the CMAC controller is solely used to control the plant, and no conventional controller is needed. Without a preliminary offline learning, the single-input CMAC controller can provide the control effort to the plant at each online learning step. To train the differentiable CMAC online, the gradient descent algorithm is employed to derive the learning rules. The sensitivity of the plant, with respect to the input, is approximated by a simple formula so that the learning rules can be applied to unknown plants. Moreover, based on a discrete-type Lyapunov function, conditions on the learning rates guaranteeing the convergence of the output error are derived in this paper. Finally, simulations on controlling three different plants are given to demonstrate the effectiveness of the proposed controller.

关键词： Cerebellar model articulation controller (CMAC) dead-zone function gradient descent algorithm online learning ability signed distance

来源：评论

学校读者我要写书评

暂无评论

Online learning of mixture experts for real-time tracking

引用

IET COMPUTER VISION 2016年第6期10卷 585-592页

作者： Gu, S. Ma, Z. Xie, M. Chen, Z. Univ Elect Sci & Technol China Sch Commun & Informat Engn Chengdu Peoples R China Univ Elect Sci & Technol China Sch Elect Engn Chengdu Peoples R China Mem Univ Newfoundland Fac Comp Sci St John NF Canada

Template tracking has been extensively investigated in computer vision to track objects for various applications. Tracking based on gradient descent algorithm using image gradient is one of the most popular object tracking method. However, it is difficult to define the relationship between the observed data set and the warping function due to the unobserved heterogeneity of the data set which inevitably results in poor tracking performance. This study proposes a novel method based on hierarchical mixture of expert to perform robust, real-time tracking from stationary cameras. By extending the idea of hyperplane approximation, the proposed approach establishes a hierarchical mixture of generalised linear regression model instead of a single model which reduces the non-linear error. The experiments' results show significant improvement over the traditional hyperplane approximation (HA) approach.

关键词： object tracking computer vision learning (artificial intelligence) gradient methods approximation theory regression analysis image sensors mixture models online mixture experts learning template tracking computer vision gradient descent algorithm object tracking method observed data set warping function stationary cameras hyperplane approximation generalised linear regression model

来源：评论

学校读者我要写书评

暂无评论

A Low Noise Floor 3 x 3 Coupler Phase Demodulation Scheme

引用

JOURNAL OF LIGHTWAVE TECHNOLOGY 2024年第1期42卷 470-476页

作者： Zhang, Jihao Wu, Xuqiang Zhou, Wen Shi, Jinhui Guang, Dong Zuo, Cheng Mu, Shengquan Liu, Yangzhou Lin, Zhiwei Fang, Chongxu Yu, Benli Anhui Univ Informat Mat & Intelligent Sensing Lab Anhui Prov Hefei 230601 Peoples R China Anhui Univ Key Lab Optoelect Informat Acquisit & Manipulat Minist Educ Hefei 230601 Peoples R China

Refining the noise floor and enhancing reliability are crucial for optical fiber sensing systems employing 3 x 3 coupler phase demodulation technology. In this article, we achieve the intensity noise cancellation by dividing the interference signal directly by the laser intensity. Experimental results show an average noise floor reduction of 11.2 dB and a maximum reduction of 17.3 dB at 50 kHz. We then optimize the commonly used ellipse fitting algorithm (EFA) in the 3 x 3 coupler demodulation system to overcome its inability to work properly in demodulating small signals. The improved algorithm ensures the equality of direct current (DC) values of the two interference signals by introducing gradient descent. Additionally, it rotates the coordinate system to transform the Lissajous figure into a standard ellipse with fixed phase difference and limited center. This transformation reduces the five parameters of the EFA model to three. Experimental results demonstrate that this demodulation scheme achieves an 99.98% linearity and 1.56 mu rad/root Hz phase measurement resolution. Notably, it offers several advantages, including low noise floor, high reliability, and the absence of phase modulation requirements, making it highly applicable in 3 x 3 coupler demodulation systems.

关键词： Demodulation Couplers Interference Standards Fitting Floors Phase measurement gradient descent algorithm low noise floor phase demodulation standard ellipse

来源：评论

学校读者我要写书评

暂无评论

Hierarchical wavelet packet fuzzy inference system for pattern classification and system identification

引用

INTERNATIONAL JOURNAL OF SYSTEMS SCIENCE 2013年第1期44卷 109-126页

作者： Sharifi, A. Shoorehdeli, M. Aliyari Teshnehlab, M. Islamic Azad Univ Sci & Res Branch Dept Comp Tehran Iran KNT Univ Technol Mechatron Dept Tehran Iran KNT Univ Technol Dept Elect Engn Tehran Iran

This study presents a hierarchical Takagi-Sugeno-Kang type fuzzy system called hierarchical wavelet packet fuzzy inference system. In the proposed method, wavelet packet transform is applied on the input data to produce approximation and detail sub-bands of the input data and the output is used as the input vector of the proposed network. This network uses a hierarchical structure same as wavelet packet decomposition tree, in which adaptive network-based fuzzy inference system is used as sub-model. Also, gradient descent algorithm is chosen for training the parameters of antecedent and conclusion parts of the sub-models. In order to evaluate the capability of the proposed method, its applications in pattern classification, system identification and time-series prediction have been studied. The results show that the proposed method performs better than the other conventional models.

关键词： wavelet transform hierarchical fuzzy model gradient descent algorithm pattern classification system identification time-series prediction

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：