Search results

Conventional image super-resolution reconstruction by the conventional deep learning architectures suffers from the problems of hard training and gradient disappearing. In order to solve such problems, the purpose of this paper is to propose a novel image super-resolution algorithm based on improved generative adversarial networks (GANs) with Wasserstein distance and gradient penalty.

The proposed algorithm first introduces the conventional GANs architecture, the Wasserstein distance and the gradient penalty for the task of image super-resolution reconstruction (SRWGANs-GP). In addition, a novel perceptual loss function is designed for the SRWGANs-GP to meet the task of image super-resolution reconstruction. The content loss is extracted from the deep model’s feature maps, and such features are introduced to calculate mean square error (MSE) for the loss calculation of generators.

To validate the effectiveness and feasibility of the proposed algorithm, a lot of compared experiments are applied on three common data sets, i.e. Set5, Set14 and BSD100. Experimental results have shown that the proposed SRWGANs-GP architecture has a stable error gradient and iteratively convergence. Compared with the baseline deep models, the proposed GANs models have a significant improvement on performance and efficiency for image super-resolution reconstruction. The MSE calculated by the deep model’s feature maps gives more advantages for constructing contour and texture.

Compared with the state-of-the-art algorithms, the proposed algorithm obtains a better performance on image super-resolution and better reconstruction results on contour and texture.

This paper aims to propose a novel hand to bridge the gap between the traditional rigid robot hands and the soft hands to obtain a better grasping performance.

The proposed hand consists of three fingers. Each finger has 15 degrees of freedom and three phalanxes, which can bend in one direction when load is applied, but they are rigid toward the opposite direction at the initial position. The grasping process and simulations of the fingers are discussed in this paper. Both kinematic and dynamics analyses are performed to predict the performance of the hand. Subsequently, a prototype of the hand is developed for experiments.

Both kinematics and dynamics analyses indicate good grasping performance of the hand. Simulations and experiments confirm the feasibility of the finger design. The hand can execute hybrid grasping modes with more uniform force distribution and a larger workspace than traditional rigid fingers. The proposed hand has much potential in the industrial sector.

A new method to obtain better grasping performance and to bridge the gap between the rigid finger and the soft finger has been presented and verified. The hand combines the advantages of both the rigid phalanxes and the soft fingers. Compared with some traditional rigid fingers, the proposed design has a more uniform force distribution and a bigger workspace.

Surface roughness has a serious impact on the fatigue strength, wear resistance and life of mechanical products. Realizing the evolution of surface quality through theoretical modeling takes a lot of effort. To predict the surface roughness of milling processing, this paper aims to construct a neural network based on deep learning and data augmentation.

This study proposes a method consisting of three steps. Firstly, the machine tool multisource data acquisition platform is established, which combines sensor monitoring with machine tool communication to collect processing signals. Secondly, the feature parameters are extracted to reduce the interference and improve the model generalization ability. Thirdly, for different expectations, the parameters of the deep belief network (DBN) model are optimized by the tent-SSA algorithm to achieve more accurate roughness classification and regression prediction.

The adaptive synthetic sampling (ADASYN) algorithm can improve the classification prediction accuracy of DBN from 80.67% to 94.23%. After the DBN parameters were optimized by Tent-SSA, the roughness prediction accuracy was significantly improved. For the classification model, the prediction accuracy is improved by 5.77% based on ADASYN optimization. For regression models, different objective functions can be set according to production requirements, such as root-mean-square error (RMSE) or MaxAE, and the error is reduced by more than 40% compared to the original model.

A roughness prediction model based on multiple monitoring signals is proposed, which reduces the dependence on the acquisition of environmental variables and enhances the model's applicability. Furthermore, with the ADASYN algorithm, the Tent-SSA intelligent optimization algorithm is introduced to optimize the hyperparameters of the DBN model and improve the optimization performance.

The purpose of this paper is to study the effect of doping, annealing temperature and visible optical excitation on CuO-ZnO nanocomposites’ acetone sensing properties and introduce an attractive candidate for acetone detection at about room temperature.

ZnO nanoparticles doped with CuO were prepared by sol-gel method, and the structure and morphology were characterized via X-ray diffraction, scanning electron microscope, energy dispersive spectroscopy and Brunauer-Emmett-Teller. The photoelectric responses of CuO-ZnO nanocomposites to cetone under the irradiation of visible light were investigated at about 30°C. The photoelectric response mechanism was also discussed with the model of double Schottky.

The doping of CuO enhanced performance of ZnO nanoparticles in terms of the photoelectric responses and the gas response and selectivity to acetone of ZnO nanoparticles, in addition, decreasing the operating temperature to about 30ºC. The optimum performance was obtained by 4.17% CuO-ZnO nanocomposites. Even at the operating temperature, about 30ºC, the response to 1,000 ppm acetone was significantly increased to 579.24 under the visible light irradiation.

The sensor fabricated by 4.17% CuO-ZnO nanocomposites exhibited excellent acetone-sensing characteristics at about 30ºC. It is promising to be applied in low power and miniature acetone gas sensors.

In the present research, a new nanocomposite material of CuO-ZnO was prepared by Sol-gel method. The optimum gas sensing properties to acetone were obtained by 4.17% CuO-ZnO nanocomposites at about 30ºC operating temperature when it was irradiated by visible light with the wavelength more than 420 nm.

Molecular dynamics is an emerging simulation technique in the field of machining in recent years. Many researchers have tried to simulate different processing methods of various materials with the theory of molecular dynamics (MD), and some preliminary conclusions have been obtained. However, the application of MD simulation is more limited compared with traditional finite element model (FEM) simulation technique due to the complex modeling approach and long computation time. Therefore, more studies on the MD simulations are required to provide a reliable theoretical basis for the nanoscale interpretation of grinding process. This study investigates the crystal structures, dislocations, force, temperature and subsurface damage (SSD) in the grinding of iron-nickel alloy using MD analysis.

In this study the simulation model is established on the basis of the workpiece and single cubic boron nitride (CBN) grit with embedded atom method and Morse potentials describing the forces and energies between different atoms. The effects of grinding parameters on the material microstructure are studied based on the simulation results.

When CBN grit goes through one of the grains, the arrangement of atoms within the grain will be disordered, but other grains will not be easily deformed due to the protection of the grain boundaries. Higher grinding speed and larger cutting depth can cause greater impact of grit on the atoms, and more body-centered cubic (BCC) structures will be destroyed. The dislocations will appear in grain boundaries due to the rearrangement of atoms in grinding. The increase of grinding speed results in the more transformation from BCC to amorphous structures.

This study is aimed to study the grinding of Fe-Ni alloy (maraging steel) with single grit through MD simulation method, and to reveal the microstructure evolution within the affected range of SSD layer in the workpiece. The simulation model of polycrystalline structure of Fe-Ni maraging steel and grinding process of single CBN grit is constructed based on the Voronoi algorithm. The atomic accumulation, transformation of crystal structures, evolution of dislocations as well as the generation of SSD are discussed according to the simulation results.

The conventional pedestrian detection algorithms lack in scale sensitivity. The purpose of this paper is to propose a novel algorithm of self-adaptive scale pedestrian detection, based on deep residual network (DRN), to address such lacks.

First, the “Edge boxes” algorithm is introduced to extract region of interests from pedestrian images. Then, the extracted bounding boxes are incorporated to different DRNs, one is a large-scale DRN and the other one is the small-scale DRN. The height of the bounding boxes is used to classify the results of pedestrians and to regress the bounding boxes to the entity of the pedestrian. At last, a weighted self-adaptive scale function, which combines the large-scale results and small-scale results, is designed for the final pedestrian detection.

To validate the effectiveness and feasibility of the proposed algorithm, some comparison experiments have been done on the common pedestrian detection data sets: Caltech, INRIA, ETH and KITTI. Experimental results show that the proposed algorithm is adapted for the various scales of the pedestrians. For the hard detected small-scale pedestrians, the proposed algorithm has improved the accuracy and robustness of detections.

By applying different models to deal with different scales of pedestrians, the proposed algorithm with the weighted calculation function has improved the accuracy and robustness for different scales of pedestrians.

In order to improve the weak recognition accuracy and robustness of the classification algorithm for brain-computer interface (BCI), this paper proposed a novel classification algorithm for motor imagery based on temporal and spatial characteristics extracted by using convolutional neural networks (TS-CNN) model.

According to the proposed algorithm, a five-layer neural network model was constructed to classify the electroencephalogram (EEG) signals. Firstly, the author designed a motor imagery-based BCI experiment, and four subjects were recruited to participate in the experiment for the recording of EEG signals. Then, after the EEG signals were preprocessed, the temporal and spatial characteristics of EEG signals were extracted by longitudinal convolutional kernel and transverse convolutional kernels, respectively. Finally, the classification of motor imagery was completed by using two fully connected layers.

To validate the classification performance and efficiency of the proposed algorithm, the comparative experiments with the state-of-the-arts algorithms are applied to validate the proposed algorithm. Experimental results have shown that the proposed TS-CNN model has the best performance and efficiency in the classification of motor imagery, reflecting on the introduced accuracy, precision, recall, ROC curve and F-score indexes.

The proposed TS-CNN model accurately recognized the EEG signals for different tasks of motor imagery, and provided theoretical basis and technical support for the application of BCI control system in the field of rehabilitation exoskeleton.

The purpose of this article is about the design of controllers for conical two-tank noninteracting level (CTTNL) system in simulation. Local linearization around the equilibrium point has been done for the nonlinear CTTNL system to obtain a linearized model transfer function.

This article deals with the design of novel optimal fractional-order tilt-integral-derivative (TID) controller using type-1 fuzzy set for the CTTNL prototype system. In this study, type-1 fuzzy TID controller parameters have been optimized through genetic algorithm (GA) and those set of values have been employed for the design of proportional-integral-derivative (PID) controller.

A performance comparison between FTID and PID controller is then investigated. The analysis shows the superiority of FTID controller over PID controller in terms of integral absolute error (IAE), integral square error (ISE), integral of time multiplied absolute error (ITAE) and integral of time multiplied squared error (ITSE) integral errors. The transient and steady state performance of the FTID controller are superior as compared to conventional PID controller. In future, the FTID controller fault-tolerance capability tested on CTTNL system subject to actuator and system component (leak) faults. The detailed study of robustness in presence of model uncertainties will be incorporated as a scope of further research.

A performance comparison between FTID and PID controller is then investigated. The analysis shows the superiority of FTID controller over PID controller in terms of IAE, ISE, ITAE and ITSE integral errors. Additionally, fault-tolerant performance of the proposed controller evaluated with fault-recovery time (F_rt) parameter. The transient and steady state performance of the FTID controller are superior as compared to conventional PID controller.

Aiming at the shortcomings of EEG signals generated by brain's sensorimotor region activated tasks, such as poor performance, low efficiency and weak robustness, this paper proposes an EEG signals classification method based on multi-dimensional fusion features.

First, the improved Morlet wavelet is used to extract the spectrum feature maps from EEG signals. Then, the spatial-frequency features are extracted from the PSD maps by using the three-dimensional convolutional neural networks (3DCNNs) model. Finally, the spatial-frequency features are incorporated to the bidirectional gated recurrent units (Bi-GRUs) models to extract the spatial-frequency-sequential multi-dimensional fusion features for recognition of brain's sensorimotor region activated task.

In the comparative experiments, the data sets of motor imagery (MI)/action observation (AO)/action execution (AE) tasks are selected to test the classification performance and robustness of the proposed algorithm. In addition, the impact of extracted features on the sensorimotor region and the impact on the classification processing are also analyzed by visualization during experiments.

The experimental results show that the proposed algorithm extracts the corresponding brain activation features for different action related tasks, so as to achieve more stable classification performance in dealing with AO/MI/AE tasks, and has the best robustness on EEG signals of different subjects.

Millennials, a cohort of young consumers, are the primary group of shoppers via e-commerce. This study aims to identify Chinese individual cultural values in millennials and examine the role of perceived shopping (hedonic and utilitarian) values in creating e-loyalty.

Data were collected from 296 Chinese millennials who use online shopping websites and apps. Confirmatory factor analysis was used to examine the validity and reliability of the methodology. Structural equation modeling was employed to test the proposed hypotheses.

All hypotheses were supported except one: the respondents rejected the impact of the face value on utilitarianism. The findings confirmed that Chinese individual cultural characteristics (face and Yuan) impact perceived shopping values (hedonic and utilitarian). Furthermore, these shopping values significantly influence e-loyalty.

The study's findings suggest some implications that academicians and market practitioners should consider. Additional implications for business managers focus on cultural characteristics, strong local teams, market-based approaches and long-term strategies.

The present work highlights the online shopping behavior of Chinese millennials by exploring e-loyalty, considering its two dimensions.