首页 | 本学科首页   官方微博 | 高级检索  
相似文献
 共查询到20条相似文献，搜索用时 125 毫秒
1.
This paper presents a discrete-time decentralized neural identification and control for large-scale uncertain nonlinear systems, which is developed using recurrent high order neural networks (RHONN); the neural network learning algorithm uses an extended Kalman filter (EKF). The discrete-time control law proposed is based on block control and sliding mode techniques. The control algorithm is first simulated, and then implemented in real time for a two degree of freedom (DOF) planar robot.  相似文献   

2.
This paper introduces an alternative method artificial neural networks (ANN) used to obtain numerical solutions of mathematical models of dynamic systems, represented by ordinary differential equations (ODEs) and partial differential equations (PDEs). The proposed trial solution of differential equations (DEs) consists of two parts: The initial and boundary conditions (BCs) should be satisfied by the first part. However, the second part is not affected from initial and BCs, but it only tries to satisfy DE. This part involves a feedforward ANN containing adjustable parameters (weight and bias). The proposed solution satisfying boundary and initial condition uses a feedforward ANN with one hidden layer varying the neuron number in the hidden layer according to complexity of the considered problem. The ANN having appropriate architecture has been trained with backpropagation algorithm using an adaptive learning rate to satisfy DE. Moreover, we have, first, developed the general formula for the numerical solutions of nth-order initial-value problems by using ANN.For numerical applications, the ODEs that are the mathematical models of linear and non-linear mass-damper-spring systems and the second- and fourth-order PDEs that are the mathematical models of the control of longitudinal vibrations of rods and lateral vibrations of beams have been considered. Finally, the responses of the controlled and non-controlled systems have been obtained. The obtained results have been graphically presented and some conclusion remarks are given.  相似文献   

3.
The paper deals with a method of constructing orthonormal bases of coordinates which maximize, through redundant dictionaries (frames) of biorthogonal bases, a class separability index or distances among classes. The method proposes an algorithm which consists of biorthogonal expansions over two redundant dictionaries. Embedded classes are often present in multiclassification problems. It is shown how the biorthogonality of the expansion can really help to construct a coordinate system which characterizes the classes. The algorithm is created for training wavelet networks in order to provide an efficient coordinate system maximizing the Cross Entropy function between two complementary classes. Sine and cosine wavelet packets are basis functions of the network. Thanks to their packet structure, once selected the depth of the tree, an adaptive number of basis functions is automatically chosen. The algorithm is also able to carry out centering and dilation of the basis functions in an adaptive way. The algorithm works with a preliminary extracted feature through shrinkage technique in order to reduce the dimensionality of the problem. In particular, our attention is pointed out for time-frequency monitoring, detection and classification of transients in rail vehicle systems and the outlier problem. In the former case the goal is to distinguish transients as inrush current and no inrush current and a further distinction between the two complementary classes: dangerous inrush current and no dangerous inrush current. The proposed algorithm is used on line in order to recognize the dangerous transients in real time and thus shut-down the vehicle. The algorithm can also be used in a general application of the outlier detection. A similar structure is used in developed algorithms which are currently integrated in the inferential modeling platform of the unit responsible for Advanced Control and Simulation Solutions within ABB's (Asea Brown Boveri) industry division. It is shown how impressive and rapid performances are achieved with a limited number of wavelets and few iterations. Real applications using real measured data are included to illustrate and analyze the effectiveness of the proposed method.  相似文献   

4.
In this paper, a novel decentralized adaptive neural control approach based on the backstepping technique is proposed to design a decentralized H∞ adaptive neural controller for a class of stochastic large-scale nonlinear systems with external disturbances and unknown nonlinear functions. RBF neural networks are utilized to approximate the packaged unknown nonlinearities. A novel concept with regard to bounded-H∞ performance is proposed. It can be applied to solve an H∞ control problem for a class of stochastic nonlinear systems. The constant terms appeared in stability analysis are dealt with by using Gronwall inequality, so that H∞ performance criterion is satisfied. The assumption that the approximation errors of neural networks must be square-integrable in some literature can be eliminated. The design process for decentralized bounded-H∞ controllers is given. The proposed control scheme guarantees that all the signals in the resulting closed-loop large-scale system are uniformly ultimately bounded in probability, and each subsystem possesses disturbance attenuation performance for external disturbances. Finally, the simulation results are provided to illustrate the effectiveness and feasibility of the proposed approach.  相似文献   

5.
This paper proposes a fuzzy neural network (FNN) based on wavelet support vector regression (WSVR) approach for system identification, in which an annealing robust learning algorithm (ARLA) is adopted to adjust the parameters of the WSVR-based FNN (WSVR-FNN). In the WSVR-FNN, first, the WSVR method with a wavelet kernel function is used to determine the number of fuzzy rules and the initial parameters of FNN. After initialization, the adjustment for the parameters of FNNs is performed by the ARLA. Combining the self-learning ability of neural networks, the compact support of wavelet functions, the adaptive ability of fuzzy logic, and the robust learning capability of ARLA, the proposed FNN has the superiority among the several existed FNNs. To demonstrate the performance of the WSVR-FNN, two nonlinear dynamic plants and a chaotic system taken from the extant literature are considered to illustrate the system identification. From the simulation results, it shows that the proposed WSVR-FNN has the superiority over several presented FNNs even the number of training parameters is considerably small.  相似文献   

6.
This paper proposes a new wave digital filter derived from doubly terminated LC-ladder networks by replacing each series or shunt arm element of the ladder by its equivalent digital two-port. It is shown that such two-ports may be cascaded without the use of adapters defined by Fettweis (1). A number of realizations of the wave digital two-ports, which are canonic with respect to both multipliers and delays, have been obtained. Also a realization which is canonic with respect to multipliers only is given and an example considered using this realization. The sensitivity of this filter with respect to the multiplier coefficient changes due to finite word length is compared with the conventional cascaded digital filter and also the one proposed by Renner and Gupta. It is found that the proposed filter appears to be a more desirable form of implementation than the conventional cascade form and comparable to that of Renner and Gupta (2).  相似文献   

7.
The decentralized tracking control methods for large-scale nonlinear systems are investigated in this paper. A backstepping-based robust decentralized adaptive neural H∞ tracking control method is addressed for a class of large-scale strict feedback nonlinear systems with uncertain disturbances. Under the condition that the nonlinear interconnection functions in subsystems are unknown and mismatched, the decentralized adaptive neural network H∞ tracking controllers are designed based on backstepping technology. Neural networks are used to approximate the packaged multinomial including the unknown interconnections and nonlinear functions in the subsystems as well as the derivatives of the virtual controls. The effect of external disturbances and approximation errors is attenuated by H∞ tracking performance. Whether the external disturbances occur or not, the output tracking errors of the close-loop system are guaranteed to be bounded. A practical example is provided to show the effectiveness of the proposed control approach.  相似文献   

8.
This study uses data mining techniques to examine the effect of various demographic, cognitive and psychographic factors on Egyptian citizens’ use of e-government services. Data mining uses a broad family of computationally intensive methods that include decision trees, neural networks, rule induction, machine learning and graphic visualization. Three artificial neural network models (multi-layer perceptron neural network [MLP], probabilistic neural network [PNN] and self-organizing maps neural network [SOM]) and three machine learning techniques (classification and regression trees [CART], multivariate adaptive regression splines [MARS], and support vector machines [SVM]) are compared to a standard statistical method (linear discriminant analysis [LDA]). The variable sets considered are sex, age, educational level, e-government services perceived usefulness, ease of use, compatibility, subjective norms, trust, civic mindedness, and attitudes. The study shows how it is possible to identify various dimensions of e-government services usage behavior by uncovering complex patterns in the dataset, and also shows the classification abilities of data mining techniques.  相似文献   

9.
In this study, an adaptive fractional order sliding mode controller with a neural estimator is proposed for a class of systems with nonlinear disturbances. Compared with traditional sliding mode controller, the new proposed fractional order sliding mode controller contains a fractional order term in the sliding surface. The fractional order sliding surface is used in adaptive laws which are derived in the framework of Lyapunov stability theory. The bound of the disturbances is estimated by a radial basis function neural network to relax the requirement of disturbance bound. To investigate the effectiveness of the proposed adaptive neural fractional order sliding mode controller, the methodology is applied to a Z-axis Micro-Electro-Mechanical System (MEMS) gyroscope to control the vibrating dynamics of the proof mass. Simulation results demonstrate that the proposed control system can improve tracking performance as well as parameter identification performance.  相似文献   

10.
In this paper we focus on the problem of question ranking in community question answering (cQA) forums in Arabic. We address the task with machine learning algorithms using advanced Arabic text representations. The latter are obtained by applying tree kernels to constituency parse trees combined with textual similarities, including word embeddings. Our two main contributions are: (i) an Arabic language processing pipeline based on UIMA—from segmentation to constituency parsing—built on top of Farasa, a state-of-the-art Arabic language processing toolkit; and (ii) the application of long short-term memory neural networks to identify the best text fragments in questions to be used in our tree-kernel-based ranker. Our thorough experimentation on a recently released cQA dataset shows that the Arabic linguistic processing provided by Farasa produces strong results and that neural networks combined with tree kernels further boost the performance in terms of both efficiency and accuracy. Our approach also enables an implicit comparison between different processing pipelines as our tests on Farasa and Stanford parsers demonstrate.  相似文献   

11.
Influence maximization (IM) has shown wide applicability in immense fields over the past decades. Previous researches on IM mainly focused on the dyadic relationship but lacked the consideration of higher-order relationship between entities, which has been constantly revealed in many real systems. An adaptive degree-based heuristic algorithm, i.e., Hyper Adaptive Degree Pruning (HADP) which aims to iteratively select nodes with low influence overlap as seeds, is proposed in this work to tackle the IM problem in hypergraphs. Furthermore, we extend algorithms from ordinary networks as baselines. Results on 8 empirical hypergraphs show that HADP surpasses the baselines in terms of both effectiveness and efficiency with a maximally 46.02% improvement. Moreover, we test the effectiveness of our algorithm on synthetic hypergraphs generated by different degree heterogeneity. It shows that the improvement of our algorithm effectiveness increases from 2.66% to 14.67% with the increase of degree heterogeneity, which indicates that HADP shows high performance especially in hypergraphs with high heterogeneity, which is ubiquitous in real-world systems.  相似文献   

12.
In this paper, the discrete-time fuzzy cellular neural network with variable delays and impulses is considered. Based on M-matrix theory and analytic methods, several simple sufficient conditions checking the global exponential stability and the existence of periodic solutions are obtained for the neural networks. Moreover, the estimation for exponential convergence rate index is proposed. The obtained results show that the stability and periodic solutions still remain under certain impulsive perturbations for the neural network with stable equilibrium point and periodic solutions. Some examples with simulations are given to show the effectiveness of the obtained results.  相似文献   

13.
The present work aims to develop a novel adaptive iterative learning control(AILC) method for nonlinear multiple input multiple output (MIMO) systems that execute various control missions with iteration-varying magnitude-time scales. In order to reduce the variations of the systems, this work proposes a series of time scaling transformations to normalize the iteration-varying trial lengths. An AILC scheme is then developed for the transformed control systems on a uniform trial length, which is shown to be capable of ensuring the asymptotic convergence of the tracking error. In other words, the proposed AILC algorithm is able to relax the constraint in conventional ILC where the control task must remain the same in the iteration domain. Additionally, the basic assumption in classic ILC that the control system must repeat on a fixed finite period is also removed. The convergence analysis of the AILC is derived rigorously according to the composite energy function (CEF) methodology. It is shown that the newly developed learning control strategy works well for control plants with either time-invariant or time-varying parametric uncertainties. To show the effectiveness of the AILC, three examples are illustrated in the end. Meanwhile, the proposed learning method is also implemented to a traditional X–Y table system.  相似文献   

14.
This paper focuses on the problem of direct adaptive neural network (NN) tracking control for a class of uncertain nonlinear multi-input/multi-output (MIMO) systems by employing backstepping technique. Compared with the existing results, the outstanding features of the two proposed control schemes are presented as follows. Firstly, a semi-globally stable adaptive neural control scheme is developed to guarantee that the ultimate tracking errors satisfy the accuracy given a priori, which cannot be carried out by using all existing adaptive NN control schemes. Secondly, we propose a novel adaptive neural control approach such that the closed-loop system is globally stable, and in the meantime the ultimate tracking errors also achieve the tracking accuracy known a priori, which is different from all existing adaptive NN backstepping control methods where the closed-loop systems can just be ensured to be semi-globally stable and the ultimate tracking accuracy cannot be determined a priori by the designers before the controllers are implemented. Thirdly, the main technical novelty is to construct three new nth-order continuously differentiable switching functions such that multiswitching-based adaptive neural backstepping controllers are designed successfully. Fourthly, in contrast to the classic adaptive NN control schemes, this paper adopts Barbalat׳s lemma to analyze the convergence of tracking errors rather than Lyapunov stability theory. Consequently, the accuracy of ultimate tracking errors can be determined and adjusted accurately a priori according to the real-world requirements, and all signals in the closed-loop systems are also ensured to be uniformly ultimately bounded. Finally, a simulation example is provided to illustrate the effectiveness and merits of the two proposed adaptive NN control schemes.  相似文献   

15.
Global dissipativity of stochastic neural networks with time delay   总被引：1，自引：0，他引：1  
Liao and Wang [Global dissipativity of continuous-time recurrent neural networks with time delay, Phys. Rev. E 68 (2003) 016118] firstly studied the dissipativity of neural networks. In this paper, the neural network model is generalized to a stochastic case, and the global dissipativity in mean of such stochastic system is investigated. By constructing several proper Lyapunov functionals combining with Jensen's inequality, Itô's formula and some analytic techniques, several sufficient conditions for the global dissipativity in mean of such stochastic neural networks are derived in LMIs forms, which can be easily verified in practice. Three numerical examples are provided to demonstrate the effectiveness of our criteria.  相似文献   

16.
A novel H∞ filter design methodology has been presented for a general class of nonlinear systems. Different from existing nonlinear filtering design, the nonlinearities are approximated using neural networks, and then are modeled based on linear difference inclusions, which makes the structure of the desired filter simpler and parameter turning easier and has the advantages of guaranteed stability, numeral robustness, bounded estimation accuracy. A unified framework is established to solve the addressed H∞ filtering problem by exploiting linear matrix inequality (LMI) approach. A numerical example shows that the filtering error systems will work well against bounded error between a nonlinear dynamical system and a multilayer neural network.  相似文献   

17.
The paper is concerned with similarity search at large scale, which efficiently and effectively finds similar data points for a query data point. An efficient way to accelerate similarity search is to learn hash functions. The existing approaches for learning hash functions aim to obtain low values of Hamming distances for the similar pairs. However, these methods ignore the ranking order of these Hamming distances. This leads to the poor accuracy about finding similar items for a query data point. In this paper, an algorithm is proposed, referred to top k RHS (Rank Hash Similarity), in which a ranking loss function is designed for learning a hash function. The hash function is hypothesized to be made up of l binary classifiers. The issue of learning a hash function can be formulated as a task of learning l binary classifiers. The algorithm runs l rounds and learns a binary classifier at each round. Compared with the existing approaches, the proposed method has the same order of computational complexity. Nevertheless, experiment results on three text datasets show that the proposed method obtains higher accuracy than the baselines.  相似文献   

18.
In this paper, the development and experimental validation of a novel double two-loop nonlinear controller based on adaptive neural networks for a quadrotor are presented. The proposed controller has a two-loop structure: an outer loop for position control and an inner loop for attitude control. Similarly, both position and orientation controllers also have a two-loop design with an adaptive neural network in each inner loop. The output weight matrices of the neural networks are updated online through adaptation laws obtained from a rigorous error convergence analysis. Thus, a training stage is unnecessary prior to the neural network implementation. Additionally, an integral action is included in the controller to cope with constant disturbances. The error convergence analysis guarantees the achievement of the trajectory tracking task and the boundedness of the output weight matrix estimation errors. The proposed scheme is designed such that an accurate knowledge of the quadrotor parameters is not needed. A comparison against the proposed controller and two other well-known schemes is presented. The obtained results showed the functionality of the proposed controller and demonstrated robustness to parametric uncertainty.  相似文献   

19.
In this paper, the distributed optimization problem over multi-cluster networks is considered. Different from the existing works, this paper studies the optimization algorithm under uncoordinated step sizes. More specifically, by combining a random sleep strategy and the round-robin communication among clusters, a new hierarchical algorithm is developed to solve the considered problem. In the proposed algorithm, the random sleep strategy enables each agent to independently choose either performing the projected subgradient descent, or keeping the previous estimate by a Bernoulli decision, based on which the step size of each agent is selected as an uncoordinated form that only relates to the independent Bernoulli decision variable. Technically, by introducing a key definition on the algorithm history, it is proven that the estimates of the proposed algorithm can converge to the optimal solution even with the uncoordinated step sizes. In addition, we also study the convergence performance of the proposed algorithm with simpler constant step sizes. In this case, it is proven that the random sleep strategy can efficiently improve the convergence accuracy of the algorithm. Finally, the theoretical findings are verified via simulation examples.  相似文献   

20.
Due to the harmful impact of fabricated information on social media, many rumor verification techniques have been introduced in recent years. Advanced techniques like multi-task learning (MTL), shared-private models suffer from many strategic limitations that restrict their capability of veracity identification on social media. These models are often reliant on multiple tasks for the primary targeted objective. Even the most recent deep neural network (DNN) models like VRoC, Hierarchical-PSV, StA-HiTPLAN etc. based on VAE, GCN, Transformer respectively with improved modification are able to perform good on veracity identification task but with the help of additional auxiliary information, mostly. However, their rise is still not substantial with respect to the proposed model even though the proposed model is not using any additional information. To come up with an improved DNN model architecture, we introduce globally Discrete Attention Representations from Transformers (gDART). Discrete-Attention mechanism in gDART is capable of capturing multifarious correlations veiled among the sequence of words which existing DNN models including Transformer often overlook. Our proposed framework uses a Branch-CoRR Attention Network to extract highly informative features in branches, and employs Feature Fusion Network Component to identify deep embedded features and use them to make enhanced identification of veracity of an unverified claim. Moreover, to achieve its goal, gDART is not dependent on any costly auxiliary resource but on an unsupervised learning process. Extensive experiments reveal that gDART marks a considerable performance gain in veracity identification task over state-of-the-art models on two real world rumor datasets. gDART reports a gain of 36.76%, 40.85% on standard benchmark metrics.  相似文献   

设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号