University of Transport and Communications
Bac Giang, Vietnam
Time filter
Source Type

Son L.H.,Vietnam National University, Hanoi | Cuong B.C.,Vietnam Academy of Science and Technology | Long H.V.,University of Transport and Communications
Knowledge-Based Systems | Year: 2013

In this paper, we introduce a novel model so-called Spatial Interaction - Modification Model (SIM2), serving for the classification of spatially-referenced demographic data. It is integrated with the main part of the best fuzzy clustering algorithm for geo-demographic analysis problem - IPFGWC to form the new method named as MIPFGWC. Theoretical and experimental analyses show that MIPFGWC achieves better clustering quality than IPFGWC and other available algorithms. © 2013 Elsevier B.V. All rights reserved.

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.

Son L.H.,Vietnam National University, Hanoi | Linh N.D.,Vietnam National University, Hanoi | Long H.V.,University of Transport and Communications
Engineering Applications of Artificial Intelligence | Year: 2014

In this paper, we propose an integrated approach between fuzzy C-means (FCM) and multi-active neuro fuzzy inference system (MANFIS) for the lossless DEM compression for fast retrieval (DCR) problem, aiming to compress digital elevation model (DEM) data with the priority of fast retrieval from the client machine over the Internet environment. Previous researches of this problem either used the float wavelet transforms integrated with the SPIHT coding or constructed a predictor model using statistical correlation of DEM data in local neighborhoods; thus giving large-sized compressed data and slow transferring time of data between the server and the client. Based on the observation that different non-linear transforms for predictive values in the sliding windows may increase the compression ratio, we herein present a novel approach for DCR problem and validated it experimentally on the benchmark DEM datasets. The comparative results show that our method produces better compression ratio than the relevant ones. © 2013 Elsevier Ltd. All rights reserved.

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 V.H.,Texas A&M University | Nguyen V.H.,University of Transport and Communications | Kim W.-J.,Texas A&M University
IEEE/ASME Transactions on Mechatronics | Year: 2013

In this paper, a single-moving-part planar positioner with six coils is designed and implemented. A concentrated-field permanent-magnet matrix is employed as the stationary part. The moving platen has a compact size (185.4 mm × 157.9 mm), light mass (0.64 kg) and low-center-of-gravity. The moving platen carries three planar-motor armatures with two phases per motor. Force calculation is based on the Lorentz force law and conducted by volume integration. In order to deal with the nonlinearity due to trigonometric dependencies in the force-current relation, modified proportional-integral- derivative (PID) and lead-and-PI compensators are designed with computed currents to close the control loop and obtain the desired performances. Experimental results verified the commutation law and the force calculation. The new design with only six coils allows for simplification of the control algorithm and reduced power consumption of the positioner. The maximum travel ranges in x, y, and the rotation about the vertical axis are 15.24 cm, 20.32 cm, and 12.03°, respectively. The positioning resolution in x and y is 8 μm with the rms position noise of 6 μm. The positioning resolution in rotations about the vertical axis is 100 μrad. © 1996-2012 IEEE.

Long T.V.,University of Transport and Communications
Proceedings - 2013 RIVF International Conference on Computing and Communication Technologies: Research, Innovation, and Vision for Future, RIVF 2013 | Year: 2013

Multidimensional data visualization is an interesting research field with many applications in ubiquitous all fields of sciences. Star coordinates are one of the most common information visualization techniques for visualizing multidimensional data. A star coordinate system is a linear transformation that maps a multidimensional data space into a two-dimensional visual space, unfortunately, involving a loss of information. In this paper, we proposed to improve standard star coordinates by developing the concept of Laplacian star coordinates for visualizing multidimensional data. The Laplacian star coordinate system is based on dimension axes placement according to their similarity, which improves the quality of data representation. We prove the efficiency and robustness of our methods by measuring the quality of the representations for several data sets. © 2013 IEEE.

Van Long T.,University of Transport and Communications
International Conference on Advanced Technologies for Communications | Year: 2016

High-dimensional data visualization is a changing task with many applications in a various fields of sciences. Parallel coordinates is one of the most widely used information visualization technique for multivariate data analysis and high-dimensional geometry. The dimension ordering is an original problem for exploring structures in a high-dimensional data space. In this paper, we propose a new metric for measuring distance between two line-segment on the parallel coordinates. The metric is suitable and effective on the parallel coordinates. We use our metric distance for finding an optimal dimension ordering on the parallel coordinates. Finally, we demonstrate our method can be applied to visualize clusters in high-dimensional data on the parallel coordinates. © 2015 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.

Dao T.T.,University of Transport and Communications
IEEE Joint Conference - International Conference on IC Design and Technology, ICICDT 2016 and Solid State Systems Symposium, 4S 2016 | Year: 2016

Author presents a fully organic complementary amplifier circuit with a gate dielectric layer of P(VDF-TrFE):PS composite. Thin-film morphologies of dielectric and semiconducting layers are examined vie AFM topographic images. Good balance between low and high noise margins of the circuit is obtained by using different transistor geometry parameters. Experimental measurements indicate that at a supply voltage of 5 V and an input signal frequency of 10 Hz, the maximum gain factor is about 34. The active sensor was made by wiring the amplifier and the printed collision detector with a PVDF piezoelectric material. Collision events can be well detected with the built active sensor. © 2016 IEEE.

Loading University of Transport and Communications collaborators
Loading University of Transport and Communications collaborators