Zhang S.-H.,Xiamen University of Technology
Zidonghua Xuebao/Acta Automatica Sinica | Year: 2017
Deep belief nets consist of several-layered unsupervised restricted Boltzmann machines and one-layered su- pervised feedback neural network. It digs the inner structure and pattern of the complex input data through effective information abstraction layer by layer, which can well reflect the input mode. However, the hidden layer numbers and the feature dimension are difficult to determine. The feedback process exhibits the vanishing gradient problem, which results in ineffective structural parameters training for lower layers. Moreover, noise affects the recognition results directly. To aim at the problem, a transformation matrix between samples and labels is made for each layer to realize adaptive adjustment of the parameter of hidden layer, and the feature of the hidden layer is de-noised for improving recognition accuracy and calculation efficiency. Simulation experiments on fault diagnosis of a gearbox in complex working conditions have proved the effectiveness of the proposed method. Copyright © 2017 Acta Automatica Sinica. All rights reserved.
Xu L.,University of Waterloo |
Li J.,University of Waterloo |
Li J.,Xiamen University of Technology |
Brenning A.,University of Waterloo
Remote Sensing of Environment | Year: 2014
The discrimination of oil spills and look-alike phenomena (e.g., low wind area, wind front area and natural slicks) on Synthetic Aperture Radar (SAR) images is a crucial task in marine oil spill detection. Many classification techniques can be employed for this purpose. In order to make the best use of the large variety of statistical and machine learning classification methods, it is necessary to assess their performance differences and make recommendations for classifier selection and improvement. The objective of this paper is to compare different classification techniques for oil-spill detection in RADARSAT-1 imagery. The data of this study consists of 15 features of 192 oil spills and look-alikes identified by Canadian Ice Service between 2004 and 2008 off Canada's east and west coastal areas. The studied classifiers include the Support Vector Machine (SVM), Artificial Neural Network (ANN), tree-based ensemble classifiers (bagging, bundling and boosting), Generalized Additive Model (GAM) and Penalized Linear Discriminant Analysis (PLDA). Two performance measures, the specificity at fixed sensitivity (80%) and the area under the Receiver Operating Characteristic (ROC) curve (AUC), were estimated using cross-validation to evaluate the performance of classifiers at a high sensitivity. Overall, the bundling technique which achieved a median specificity of 90.7%, at sensitivity of 80%, significantly outperformed the second best (i.e. bagging) by 1.5 percentage points, and the worst (i.e. ANN) by 15 percentage points. The median values of AUC measure indicated consistent results. Bundling and bagging achieved comparable median AUC values of about 92%, followed by GAM and PLDA, with ANN yielding the smallest. Most classifiers (SVM, bundling and especially PLDA and ANN) performed significantly better on datasets pre-processed by log-transformation and standardization than on the original dataset. These results demonstrate the importance and benefit of selecting the optimal classifiers for oil spill classification, and configuring the classifiers by proper feature construction techniques. © 2013 Elsevier Inc.
Yang Y.,Xiamen University of Technology |
Xu X.,Fuzhou University
Computational Materials Science | Year: 2012
In this work a series of carbon allotropes related to graphene, called graphyne, graphdiyne, gaphene-3, graphene-4 and graphene-5 are constructed by connecting two adjacent hexagonal rings with different number of acetylenic linkages. Mechanical properties of these monolayer networks are investigated through acting tensile loads on the architectures and molecular dynamics simulations are performed to calculate the fracture strains and associated ultimate stresses. In the armchair loading case, the fracture strain remains nearly unchanged whereas the ultimate strength degrades gradually with longer acetylenic chains. In the zigzag loading situation, the ultimate strength remains nearly the same whereas the fracture strain improves by a little amount with longer acetylenic chains. Furthermore, Young's moduli of all the investigated architectures are computed to analyze the material stiffness at the near equilibrium regime. The obtained results show that these structures are mechanically stable with high strength and stiffness. The unique mechanical property variations of graphyne family against armchair and zigzag loads suggest flexible designations towards functional use of this novel material, especially in the direction-dependent nanomechanical applications. © 2012 Elsevier B.V. All rights reserved.
Zheng X.Q.,Xiamen University of Technology
Advanced Materials Research | Year: 2014
Thanks to the development of microprocessors, hybrid stepping motors have been widely used in many areas where they perform positioning operations. However, the stepping motor suffers from system variations, low performance and lack of adaptability to load variations, which slow down their responding speed of high-precision positioning operations. In this paper, a computational verb PID controller is proposed to control the position of a stepping motor drive. The simulation results show that the computational verb PID controller has better performances than conventional and fuzzy PID controllers. The simulation results also show that the responding speed and positioning accuracy of the controlled hybrid stepping motor were greatly improved. Computational verb PID controller has much less computational complexity than fuzzy PID controller. © (2014) Trans Tech Publications, Switzerland.
Wang C.,Xiamen University of Technology
PLoS ONE | Year: 2016
In this paper, control of uncertain fractional-order financial chaotic system with input saturation and external disturbance is investigated. The unknown part of the input saturation as well as the system's unknown nonlinear function is approximated by a fuzzy logic system. To handle the fuzzy approximation error and the estimation error of the unknown upper bound of the external disturbance, fractional-order adaptation laws are constructed. Based on fractional Lyapunov stability theorem, an adaptive fuzzy controller is designed, and the asymptotical stability can be guaranteed. Finally, simulation studies are given to indicate the effectiveness of the proposed method. © 2016 Chenhui Wang. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Zuo J.,Xiamen University of Technology
Applied Surface Science | Year: 2010
Ag nanostructures on TiO2 films were deposited by RF magnetron sputtering under variable deposition parameters, such as DC potential, RF-power and total pressure. The concentration, shape, and distribution of the deposited nanostructures and continuous Ag films on thin films of TiO2 can be tailored by careful variation of the deposition parameters. Controllable clusterlike, islandlike and film Ag structures on TiO2 film were obtained, respectively. DC potential was found as an appropriate parameter to tailor the change of Ag nanostructure and the overall Ag amount. The compositions, nanostructures and morphologies of nanocomposite films appreciably influence the optical response. © 2010 Elsevier B.V.
Xu J.,Xiamen University of Technology
International Journal of Heat and Technology | Year: 2015
Having the premise of the certain acoustic performance, a muffler should make the pressure loss as small as possible. A simulation model of a fork truck muffler with a complex structure is established. Based on the finite volume method, multi-dimensional numerical simulation regarding velocity field and pressure field of steady flows for a muffler is performed using CFD (computational fluid dynamic method). Flow characteristics and pressure distribution of the muffler are analyzed. It is found that the vortex inside the muffler creates a great pressure loss. With the increases of inlet gas flow rate, the pressure loss of the muffler increases gradually. The internal structure of the muffler is redesigned for obtaining the optimized structure on the basis of analysis. The influences of the inner tube length on the flow and pressure loss of muffler are researched. The study will provide a theoretical basis for designing a complex muffler.
Mi X.,Xiamen University of Technology
International Journal of Advancements in Computing Technology | Year: 2012
In the paper, the rectangular column sandwich plate protective device is developed to resist the impact loading without the bridge's damage. And its behavior is studied. The ANSYS/LS-DYNA, which is non-linear dynamics finite element software, is applied to simulate the crashworthiness performance of the rectangular column steel sandwich plate. The crashworthiness performance factors of the sandwich plate are analyzed so as to optimize the behavior of the sandwich plate. Through the structural optimization, it can be gotten that the rectangular column sandwich plates are better than traditional steel plates in crashworthiness performance. Then, the sandwich plate is used to replace the outside steel plates of the traditional box-protective device. Using the ANSYS/LS-DYNA simulation, the behavior of the rectangular column steel sandwich protective devices is obtained. The results show that he rectangular column steel sandwich protective device is better than the traditional one.
Cui J.F.,Xiamen University of Technology |
Chae H.S.,Pusan National University
Information and Software Technology | Year: 2011
Context: Component identification, the process of evolving legacy system into finely organized component-based software systems, is a critical part of software reengineering. Currently, many component identification approaches have been developed based on agglomerative hierarchical clustering algorithms. However, there is a lack of thorough investigation on which algorithm is appropriate for component identification. Objective: This paper focuses on analyzing agglomerative hierarchical clustering algorithms in software reengineering, and then identifying their respective strengths and weaknesses in order to apply them effectively for future practical applications. Method: A series of experiments were conducted for 18 clustering strategies combined according to various similarity measures, weighting schemes and linkage methods. Eleven subject systems with different application domains and source code sizes were used in the experiments. The component identification results are evaluated by the proposed size, coupling and cohesion criteria. Results: The experimental results suggested that the employed similarity measures, weighting schemes and linkage methods can have various effects on component identification results with respect to the proposed size, coupling and cohesion criteria, so the hierarchical clustering algorithms produced quite different clustering results. Conclusions: According to the experimental results, it can be concluded that it is difficult to produce perfectly satisfactory results for a given clustering algorithm. Nevertheless, these algorithms demonstrated varied capabilities to identify components with respect to the proposed size, coupling and cohesion criteria. © 2011 Elsevier B.V. All rights reserved.
Tang D.,Xiamen University of Technology
Differential Geometry and its Application | Year: 2011
In this paper, we study a new non-Riemannian quantity H defined by the S-curvature. We find that the non-Riemannian quantity is closely related to S-curvature. We characterize Randers metrics of almost isotropic S-curvature if and only if they have almost vanishing H-curvature. Furthermore, the Randers metrics actually have zero S-curvature if and only if they have vanishing H-curvature. © 2010 Elsevier B.V.