Huan H.X.,Vietnam National University, Hanoi |
Hien D.T.T.,University of Transport and Communications |
Tue H.H.,Hanoi University
IEEE Transactions on Neural Networks | Year: 2011
This brief paper proposes a new algorithm to train interpolation Gaussian radial basis function (RBF) networks in order to solve the problem of interpolating multivariate functions with equally spaced nodes. Based on an efficient two-phase algorithm recently proposed by the authors, Euclidean norm associated to Gaussian RBF is now replaced by a conveniently chosen Mahalanobis norm, that allows for directly computing the width parameters of Gaussian radial basis functions. The weighting parameters are then determined by a simple iterative method. The original two-phase algorithm becomes a one-phase one. Simulation results show that the generality of networks trained by this new algorithm is sensibly improved and the running time significantly reduced, especially when the number of nodes is large. © 2011 IEEE.
Nguyen H.V.,University of Transport and Communications
Construction and Building Materials | Year: 2013
The quality of hot recycled asphalt mixture prepared by conventional laboratory procedures tends to be overestimated. This is because the long RAP preheating time, for instance 2 h at 110 c might entirely soften RAP and coincidentally enhance the mixing between RAP and virgin materials thus the quality of hot recycled mixture. In order to manufacture the laboratory specimens that better present the quality as produced in the industrial mixing plant, this research employed the mixing method that simulates the mixing mechanism actually occurs in the asphalt mixing plant in which RAP at ambient temperature is mixed with superheated virgin aggregate at 215 c for 2, 4, 6 and 8 min before this blend is mixed with virgin binder. The proportion of RAP in the recycled mixture is 40% and bitumen 160/220 penetration grade is used as virgin binder. Two sizes of RAP, 20 and 40 mm are used. The stiffness of recycled asphalt specimens composed of different sizes of RAP and prepared by different mixing procedure including the conventional approach is recorded and analysed. The result indicates with short mixing time, RAP tends to work as inert black rock and the complete blending situation only approached if the mixing duration significantly increased. This finding is against the premise that RAP and virgin binder are completely blending in the recycled mixture design literature. However, longer RAP/virgin aggregate mixing times could result in recycled hot asphalt mixtures with higher stiffness modulus and better homogeneity. In addition, RAP size seriously affects the level of blending hence the stiffness variation of recycled mixtures. © 2013 Elsevier Ltd. All rights reserved.
Nguyen H.C.,University of Transport and Communications
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2016
Digital images can easily be tampered because of the popularity and power editing software. In order to create a persuasive forged image, the image is usually exposed to several geometric transformations, such as rescaling and rotating. Since the manipulations require a resampling step, uncovering traces of resampling became an important approach for detecting image forgeries. In this paper, we propose a new technique to reveal image resampling artifacts. The technique employs specific features of the linear dependencies of neighboring image samples for discriminating resampled images from original images. A machine learning method is utilized for classification. Experimental results in a large dataset show that the proposed technique is good in detecting resampled images, even when the manipulated images were slightly transformed. © Springer-Verlag Berlin Heidelberg 2016.
Nguyen V.H.,University of Transport and Communications |
Kim W.-J.,Texas A&M University
IEEE Transactions on Energy Conversion | Year: 2012
This paper presents the electromagnetic design and force calculation of a compact multiaxis precision positioner. A six-coil single-moving-part platen moves over a superimposed concentrated-field permanent-magnet matrix. With a rectangular coil placed in the magnetic field generated by the superimposed concentrated-field magnet matrix, the force acting on the coil is calculated by volume integration based on the Lorentz force law. The distance between the long sides and that between the short sides of a rectangular coil are designed to be a half pitch and one pitch of the magnet matrix, respectively. This allows for the simplification of force generation and calculation, compact size, and light mass (0.64kg) of the moving platen. Six coils are divided into three two-phase linear-motor armatures with 270 or 450 phase differences. The complete force-current relation for the entire platen with the six coils is derived. Experimental results are presented to verify the working principle of the positioner designed in this paper. The positioner can be employed for the stepping and scanning applications that require 3-DOF planar motions with long travel ranges in two horizontal directions and small rotational motions about the vertical axis. © 2011 IEEE.
Nguyen X.H.,University of Transport and Communications
International Journal of Structural Stability and Dynamics | Year: 2014
This paper presents a simplified modeling strategy for simulating the nonlinear behavior of reinforced concrete (RC) structures under seismic loadings. A new type of Euler-Bernoulli multifiber beam element with axial force and bending moment interaction is introduced. To analyze the behavior of RC structures in the axial direction, the interpolation of the axial strain is enriched using the incompatible modes method. The model uses the constitutive laws based on plasticity for steel and damage mechanics for concrete. The proposed multifiber element is implemented in the finite element Code-Aster to simulate the nonlinear behavior of two different RC structures. One structure is a building tested on a shaking table; the other is a column subjected to cyclic loadings. The comparison between the simulation and experimental results shows that the performance of this approach is quite good. The proposed model can be used to investigate the behavior of a wider variety of configurations which are impossible to study experimentally. © 2014 World Scientific Publishing Company.