Search

In this paper, a new speech feature fusion method is proposed for speaker recognition on the basis of the cross gate parallel convolutional neural network (CG-PCNN). The Mel filter bank features (MFBFs) of different frequency resolutions can be extracted from each speech frame of a speaker's speech by several Mel filter banks, where the numbers of the triangular filters in the Mel filter banks are different. Due to the frequency resolutions of these MFBFs are different, there are some complementaries for these MFBFs. The CG-PCNN is utilized to extract the deep features from these MFBFs, which applies a cross gate mechanism to capture the complementaries for improving the performance of the speaker recognition system. Then, the fusion feature can be obtained by concatenating these deep features for speaker recognition. The experimental results show that the speaker recognition system with the proposed speech feature fusion method is effective, and marginally outperforms the existing state-of-the-art systems.

A numerical study of the crushing of circular aluminium tubes with and without aluminium foam fillers has been carried out to investigate their buckling behaviours under axial compression. A crushing mode classification chart has been established for empty tubes. The influence of boundary conditions on crushing mode has also been investigated. The effect of foam filler on the crushing mode of tubes filled with foam was then examined. The predicted results would assist the design of crashworthy tube components with the preferred crushing mode with the maximum energy absorption., Comment: 6 pages, 5th Australasian Congress on Applied Mechanics (ACAM 2007), Brisbane : Australia (2007)

Some highly ordered compounds of graphene oxide (GO), e.g., the so-called clamped and unzipped GO, are shown to have piezoelectric responses via first-principles density functional calculations. By applying an electric field perpendicular to the GO basal plane, the largest value of in-plane strain and strain piezoelectric coefficient, d31 are found to be 0.12% and 0.24 pm/V, respectively, which are comparable with those of some advanced piezoelectric materials. An in-depth molecular structural analysis reveals that deformation of the oxygen doping regions in the clamped GO dominates its overall strain output, whereas deformation of the regions without oxygen dopant in the unzipped GO determines its overall piezoelectric strain. This understanding explains the observed dependence of d31 on oxygen doping rate, i.e., higher oxygen concentration giving rise to a larger d31 in the clamped GO whereas leading to a reduced d31 in the unzipped GO. As the thinnest two-dimensional piezoelectric materials, GO has a great potential for a wide range of MEMS/NEMS actuators and sensors., Comment: 10 pages, 4 figures

The finite element method and dimensional analysis have been applied in the present paper to study a hydraulic impact, which is utilized in a non-explosive rock breaking technology in mining industry. The impact process of a high speed piston on liquid water, previously introduced in a borehole drilled in rock, is numerically simulated. The research is focused on the influences of all the parameters involved in the technology on the largest principal stress in the rock, which is considered as one of the key factors to break the rock. Our detailed parametric investigation reveals that the variation of the isotropic rock material properties, especially its density, has no significant influence on the largest principal stress. The influences of the depth of the hole and the depth of the water column are also very small. On the other hand, increasing the initial kinetic energy of the piston can dramatically increase the largest principal stress and the best way to increase the initial kinetic energy of the piston is to increase its initial velocity. Results from the current dimensional analysis can be applied to optimize this non-explosive rock breaking technology.