Conventional Network Structure Research Articles

An intelligent technique for network resource management and analysis of 5G-IoT smart healthcare application

<p>The transformation of traditional hospitals to a smart healthcare system is a patient-centric, reliable, and smarter way to enhance healthcare facilities for individuals. These healthcare devices are based on the Internet of Things (IoT) and conventional network structures such as 3rd generation (3G) or 4th generation (4G). That did not provide the patient health data in the real-time scenario. Therefore, the integrated 5G-IoT system offers distinct benefits such as remote monitoring, surgery, and real-time data analysis of healthcare system that is operated with slice of dedicated network. It generates huge data with the billions of healthcare equipment that can be evaluated for decision making. This enormous data requires low latency, high-security, capacity, and more reliable techniques for evaluation. In this scenario, network slicing is one of the key solutions that can be considered for isolated end-to-end network structures. Network slicing enables logical and independent network virtual networks to be multiplexed for the same physical network. In this article, we propose an intelligent approach for network resource optimization and analysis of a 5G-IoT based smart healthcare network. We examine the performance of the machine learning algorithm by an automatic model, named A Fast Library for automated machine learning (FLAML). In addition, we perform network slicing to obtain resource optimization in 5G-IoT-based healthcare application.<strong></strong></p>

TANet: Transformer‐based asymmetric network for RGB‐D salient object detection

AbstractExisting RGB‐D salient object detection methods mainly rely on a symmetric two‐stream Convolutional Neural Network (CNN)‐based network to extract RGB and depth channel features separately. However, there are two problems with the symmetric conventional network structure: first, the ability of CNN in learning global contexts is limited; second, the symmetric two‐stream structure ignores the inherent differences between modalities. In this study, a Transformer‐based asymmetric network is proposed to tackle the issues mentioned above. The authors employ the powerful feature extraction capability of Transformer to extract global semantic information from RGB data and design a lightweight CNN backbone to extract spatial structure information from depth data without pre‐training. The asymmetric hybrid encoder effectively reduces the number of parameters in the model while increasing speed without sacrificing performance. Then, a cross‐modal feature fusion module which enhances and fuses RGB and depth features with each other is designed. Finally, the authors add edge prediction as an auxiliary task and propose an edge enhancement module to generate sharper contours. Extensive experiments demonstrate that our method achieves superior performance over 14 state‐of‐the‐art RGB‐D methods on six public datasets. The code of the authors will be released at https://github.com/lc012463/TANet.

Deep Multiscale Detail Networks for Multiband Spectral Image Sharpening.

We introduce a new deep detail network architecture with grouped multiscale dilated convolutions to sharpen images contain multiband spectral information. Specifically, our end-to-end network directly fuses low-resolution multispectral and panchromatic inputs to produce high-resolution multispectral results, which is the same goal of the pansharpening in remote sensing. The proposed network architecture is designed by utilizing our domain knowledge and considering the two aims of the pansharpening: spectral and spatial preservations. For spectral preservation, the up-sampled multispectral images are directly added to the output for lossless spectral information propagation. For spatial preservation, we train the proposed network in the high-frequency domain instead of the commonly used image domain. Different from conventional network structures, we remove pooling and batch normalization layers to preserve spatial information and improve generalization to new satellites, respectively. To effectively and efficiently obtain multiscale contextual features at a fine-grained level, we propose a grouped multiscale dilated network structure to enlarge the receptive fields for each network layer. This structure allows the network to capture multiscale representations without increasing the parameter burden and network complexity. These representations are finally utilized to reconstruct the residual images which contain spatial details of PAN. Our trained network is able to generalize different satellite images without the need for parameter tuning. Moreover, our model is a general framework, which can be directly used for other kinds of multiband spectral image sharpening, e.g., hyperspectral image sharpening. Experiments show that our model performs favorably against compared methods in terms of both qualitative and quantitative qualities.

Polymerizable Microsphere-Induced High Mechanical Strength of Hydrogel Composed of Acrylamide.

Polymerizable microspheres are introduced into acrylamide to prepare the high mechanical strength hydrogels with a novel three-dimensional pore structure. Rheological properties, compressive stress–strain, tensile property, and compression strength of three different types of hydrogels were investigated. Moreover, a scanning electron microscope (SEM) was adopted to observe the three-dimension network structure of three different types of hydrogels. The test results illustrated that viscous moduli (G″) and elastic moduli (G′) of a hydrogel containing polymerizable microspheres (P) reached maximum values, compared to the normal hydrogel (N) and the composite hydrogel containing ordinary microspheres (O). When the hydrogels were squeezed, the N was easily fractured under high strain (99%), whereas the P was not broken, and quickly recovered its initial morphology after the release of load. The P showed excellent tensile properties, with an elongation at break up to 90% and a tensile strength greater than 220 g. The compression strength of the N was 100.44 kPa·m−1, while the resulting strength of P was enhanced to be 248.00 kPa·m−1. Therefore, the various performances of N were improved by adding polymerizable microspheres. In addition, the SEM images indicated that N has a general three-dimensional network structure; the conventional network structure did not exist in the P, which has a novel three-dimensional pore structure in the spherical projection and very dense channels, which led to the compaction of the space between the three-dimensional pore network layers and reduced the flowing of free water wrapped in the network. Therefore, the mechanical strength of hydrogel was enhanced.

Modified Betweenness-Based Measure for Prediction of Traffic Flow on Urban Roads

The conventional network structure measure, betweenness centrality (BC), has been used to predict roadway traffic flow; however, without consideration of variable travel demand on the roadway network, the measure’s prediction capability is limited by its static nature. With the objective of explicitly addressing the effects of travel pattern [e.g., origin–destination (O-D) distribution] on the prediction of urban roadway traffic flow, this paper proposes a modified measure that integrates the BC index and the traditional travel demand metrics (i.e., the O-D demand and the ratios of total demand). In the case study, roadway networks of two cities—San Francisco, California, and Nanjing, China—were selected to demonstrate the effectiveness of the proposed method. Taxi GPS trace data in the two cities were retrieved and applied to calibrate and validate the proposed measure. Correlation analyses were conducted to compare the predictions of the conventional and modified betweenness measures against the observed taxi traffic flows. The results show that, compared with the traditional approach, the modified betweenness measure produces predictions that have better correlations with the observed taxi traffic flows in both case studies. More accurate predictions for the future year traffic flow can be expected with the application of the new modified measure.

New MILP-based iterative approach for retrofitting heat exchanger networks with conventional network structure modifications

Conventional network structure modifications for retrofitting heat exchanger networks (HENs) usually require a large network superstructure, which might lead to very large scale retrofit models, and thus increases the computational difficulties of optimization, especially for industrial applications. This paper presents a novel optimization approach of retrofitting HENs for both small scale and industrial scale problems. To overcome the drawback of the existing retrofit research, a simple but efficient method is proposed to determine the network retrofit superstructure based on the concept of pinching match. Then two optimization stages (network structure optimization and investment optimization) are solved with the MILP-based iteration method that has been developed. The proposed approach has many benefits for solving HEN retrofit problems. First, smaller size of retrofit superstructure is provided based on the pinching match analysis. Second, the MILP-based iteration method developed converts conventional a mixed integer non-linear programming (MINLP) to a mixed integer linear programming (MILP), and significantly reduce the computational difficulties of nonlinear programming associated with configurational changes or structural modifications for existing heat exchanging arrangements, which is upgraded based on the MILP-based iterative method proposed by Pan et al. (2012a) for retrofitting HENs with heat transfer intensifications, as the earlier work was only applicable to the cases without network topology modification. In case studies, the new proposed approach is compared with the existing methods in two literature examples and one industrial problem, demonstrating that the new proposed approach is able to obtain better and more realistic solutions for HEN retrofitting considering the economic trade-offs between energy savings and investments.

Neutron diffraction observations of interstitial protons in dense ice

The motif of distinct H2O molecules in H-bonded networks is believed to persist up to the densest molecular phase of ice. At even higher pressures, where the molecule dissociates, it is generally assumed that the proton remains localized within these same networks. We report neutron-diffraction measurements on D2O that reveal the location of the D atoms directly up to 52 GPa, a pressure regime not previously accessible to this technique. The data show the onset of a structural change at ∼13 GPa and cannot be described by the conventional network structure of ice VII above ∼26 GPa. Our measurements are consistent with substantial deuteron density in the octahedral, interstitial voids of the oxygen lattice. The observation of this "interstitial" ice VII form provides a framework for understanding the evolution of hydrogen bonding in ice that contrasts with the conventional picture. It may also be a precursor for the superionic phase reported at even higher pressure with important consequences for our understanding of dense matter and planetary interiors.

Field Data Recovery of Tidal Level Using a Neuro-Genetic Algorithm

This paper presents validation results of the neuro-genetic algorithm with conventional network structure and recursive implementation. The aim of the research is to recover the missing data of tidal level in the vicinity of the Chao Phraya river mouth, in Thailand. This data recovery system (DRS) is based on the use of transfer function approach where the system response is constructed by the concept of learning from experiences. The genetic algorithm (GA) was used to find the optimum number of units in the hidden layer. Sensitivity test of input units was performed by trials. A self recovery and a spatial recovery are investigated at two tidal stations. It is found that the obtained network outperforms the harmonic analysis and can be used in real practice. In general, the efficiency index of the design network is found more than 0.90. Overall, the NN model reproduces the time series tidal level data in the missing window. The use of conventional network structure with known data set at the neighboring station in the spatial recovery gives better results than the use of recursive architecture in the self recovery.

Modified neural networks for rapid recovery of tokamak plasma parameters for real time control

Two modified neural network techniques are used for the identification of the equilibrium plasma parameters of the Superconducting Steady State Tokamak I from external magnetic measurements. This is expected to ultimately assist in a real time plasma control. As different from the conventional network structure where a single network with the optimum number of processing elements calculates the outputs, a multinetwork system connected in parallel does the calculations here in one of the methods. This network is called the double neural network. The accuracy of the recovered parameters is clearly more than the conventional network. The other type of neural network used here is based on the statistical function parametrization combined with a neural network. The principal component transformation removes linear dependences from the measurements and a dimensional reduction process reduces the dimensionality of the input space. This reduced and transformed input set, rather than the entire set, is fed into the neural network input. This is known as the principal component transformation-based neural network. The accuracy of the recovered parameters in the latter type of modified network is found to be a further improvement over the accuracy of the double neural network. This result differs from that obtained in an earlier work where the double neural network showed better performance. The conventional network and the function parametrization methods have also been used for comparison. The conventional network has been used for an optimization of the set of magnetic diagnostics. The effective set of sensors, as assessed by this network, are compared with the principal component based network. Fault tolerance of the neural networks has been tested. The double neural network showed the maximum resistance to faults in the diagnostics, while the principal component based network performed poorly. Finally the processing times of the methods have been compared. The double network and the principal component network involve the minimum computation time, although the conventional network also performs well enough to be used in real time.

Optimal hierarchical structure of broadcast network

With the realization of broadband integrated service digital networks (B-ISDN), broadcast services will occupy a large part of the network traffic. In these services, the same information is distributed to many subscribers through the point-to-multipoint connection paths. This connection form is quite different from that of conventional point-to-point services. Since the conventional network structure is designed for point-to-point services, the current network topology may lower the quality of service or may increase construction costs. In this paper, the hierarchical broadcast network topology is optimized under certain assumptions: the cost functions, the population distribution, and the percentage of broadcast services are treated as parameters. The result shows that the ring-star-star or ring-ring-star topologies can be better than the conventional mesh-star topology in some cases, and can be much better when the viewing ratio of broadcast services increases. © 1998 Scripta Technica, Electron Comm Jpn Pt 1, 82(2): 48–56, 1999