The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
Self-adaptive back propagation neural network (BPNN) models based on hierarchical clustering were developed to classify corn kernels. To generate the sample sets, randomly selected kernels were divided into seven classes using multiple clusters, including three classes of flat kernels, three classes of round kernels and abnormal class. Further, the stepwise discriminant analysis was conducted to select...
The contract force between the fingertip and object need to be controlled when the underwater dexterous hand grasp the object. The original impedance control method lacks robustness because of the inaccurate dynamic model, position error and unknown stiffness of the object, thus, the position-based neural network impedance control method was studied. The effect of position error was analyzed, the...
Taking FOG SINS (fiber-optic gyroscope strapdown inertial system) as an object, a new fault diagnostic scheme based on BP (back-propagation) neural network is proposed. Being capable of training and simulating data off-line, neural networks provide a solution to overcome some drawbacks of the quantitative fault diagnosis. The fault tree of FOG SINS is analyzed, which is the basis of the study of neural...
In order to detect features of protuberant characters, a novel stroke detection method based on Gabor filters is proposed. First, the gray images of protuberant characters were preprocessed using morphological algorithm. Next, a set of Gabor filters is used to break down an image of protuberant characters into four directional images, which contain the stroke information of four directions. Then,...
Backpropagation learning algorithm for multilayer perceptrons (MLPs) has disadvantages of slow convergence and easily being trapped into local optimum. Inspired by efficient global searching ability of particle swarm optimization (PSO), a novel PSO based backpropagation learning algorithm (PSO-BP) is proposed. At first, training procedure for MLPs is formulated as nonlinear optimization problem that...
This paper evaluates the performance in knowledge-based companies by using back propagation (BP) neural networks. Based on the theory and methods on performance, this paper presents the variables to evaluate performance in knowledge-based companies. With the application of BP neural networks, this paper constructs a model and describes the steps to use BP neural networks in performance assessment...
A PID neural-network-based space-vector pulse width modulation (SVPWM) for a three-level inverter is proposed in this paper. A three-level inverter has lots of switching states about the vectors, and the implementation of modulation algorithm is considerably complex. In the proposed design, fast implementation of SVPWM algorithm is realized based on PID neural network instead of conventional neural...
This paper proposed an algorithm of feature selection used in fusion of soft computing and based on the chain of data-information-cognition. The algorithm is as follow: Firstly, the weights wij from input layer to hidden layer are obtained when the training accuracy of BP neural network (BPNN) is got. Where i denotes the i th feature and j denotes the j th node in hidden layer of BPNN. Secondly, zeta...
In the paper, the mapping relation between static characteristics of electro-hydraulic servo valve and fault pattern is analyzed. The method of multi-parameter fault pattern recognition of electro-hydraulic servo valve based on BP neural networks is introduced, which is researched based on tests. The result shows that the accuracy rate of fault pattern recognition is higher by adopting this method...
A new image restoration method was presented and investigated based on genetic algorithm BP neural network. The method combined the characteristics of global optimization of genetic algorithm with local optimization of BP neural network. The mapping relationship between degenerated image and clear image was established by training genetic algorithm BP neural network. Experimental results show that...
Based on pattern recognition theory and least squares support vector machine(LS-SVM) technology, automatic detection, location, segmentation and recognition of vehicles and license plates characters are discussed. A new multi-sorts classification method-binary exponent classification is proposed. By comparing LS-SVM with BP neural network in vehicle and license plates pattern recognition and classification...
Based on rough sets reducts, a new neural network ensemble method is proposed. Reducts with robustness and good generalization ability are achieved by a dynamic reduction technology. Then according to different reducts, multiple BP neural networks are designed as base classifiers. And with the idea of selective ensemble, the best neural network ensemble can be found by some search strategies. Finally,...
The IT projects risk assessment is a focus problem of the practice and theories research of IT project management. In this paper, a set of index system of IT project risk assessment is established. Based on the index system, an integration classifier with GCPSO-based artificial neural network is established to assess IT projects risk in project management. At last, an experiment is given and BP neural...
In recent years, back-propagation (BP) neural network has been widely applied to the remote sensing image classification. However, the BP method based on the gradient descent principle suffers from the problem of getting stuck at local minimum. In addition, only using spectral information for multispectral remote sensing image classification could not get the ideal result. In this paper, a new method...
A novel anti-collision fuzzy model is proposed to decide whether anti-collision action should be taken. To build a supporting system to ship anti-collision in multiple ships encounter, a last time to take action (LTTA) concept is proposed without considering any psychological factors. To obtain a LTTA capability in the anti-collision model, an innovative neurofuzzy network is proposed and applied...
According to the traffic flow features of urban intersections, a multi-phase adaptive control algorithm is given. The structure of network and program of realizing fuzzy control based on improved multi-layer BP neural networks are obtained. Results of simulation research show that with the abilities of learning and generation, the fuzzy neural controller can cope with the fast changing of arriving...
Partners selection and evaluation is one of the most vital actions of companies in a product design chain. Moreover, decision support is an important part of cooperative design. Selecting the right partner is the assurance of cooperative design. In order to improve reliability and scientificalness, a method of partner evaluation based on BP neural network and PCA is presented. An evaluation index...
Desulfurization of molten iron is very complex. The model based on traditional technology cannot achieve optimal desulfurization effect. A multi BP NN predictor model for desulfurizing molten iron is presented in this paper to predict three important operation parameters including consumption of desulfurizer, stirring speed and time. The results demonstrate that the model has good performance and...
The control strategy of genetic neural network (GNN) combines the good performance of back-propagation (BP) in weight learning and genetic algorithm (GA) in gaining global optimum. Firstly, the control strategy optimizes initial samples of the control system by GA, and then, weights and thresholds of the neuron are trained by applying a GNN approach, so that the performance parameter of the controllerpsilas...
Considering the issues that the sewage treatment process is a complicated and nonlinear system, it is very difficult to found the process model to describe it, and the key parameters of sewage treatment quality can not be detected on-line, a soft measurement modeling method based on high speed and precise genetic algorithm neural network is presented in this paper. The high speed and precise genetic...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.