With the arrival of 5G networks, cellular networks are moving in direction diversified, broadband, integrated, and intelligent networks. At same time, popularity various smart terminals has led to an explosive growth traffic. Accurate network traffic prediction become important part intelligence. In this context, paper proposes a deep learning method for space-time modeling communication First, we analyze temporal spatial characteristics from Telecom Italia. On basis, propose hybrid...

The Danjiangkou hydropower station is a water source project for the middle line of South-to-North Water Transfer Project in China. dam composed riverbed concrete and earth rock on both banks, with total length 3442 m. Once wrecked, it will yield disastrous consequences. Therefore, very important to evaluate safety behavior time. Based long-term short-term memory (LSTM) network, deformation prediction models embankment are constructed. contain two LSTM layers, adopt rectified linear unit...

Abstract To achieve an accurate comprehensive evaluation of smart distribution networks, it is necessary to scientifically select the key indicators for this evaluation. This paper proposes a method screening networks based on grey relational analysis. First, analysis large amount data state and operation 40 are selected from dimensions safety, reliability, intelligence, economy. Then, using analysis, importance each indicator in process network analyzed perspective correlation degree, while...

Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI

Circular synthetic aperture radar (CSAR) has the potential to form 3D images with single-pass single-channel data, which is very time-efficient. This article proposes a volumetric neural renderer that utilizes CSAR 2D amplitude reconstruct power distribution of imaged scene. The innovations are two-fold: Firstly, we propose new SAR image formation model establishes linear mapping relationship between multi-look amplitude-squared and real-valued 4D (spatial location (x, y, z) azimuth angle θ)...

The imputation of time series traffic flow data is great significance to the intelligent transportation, urban planning, and road emergency handling. This paper proposes a filling missing with Generative Adversarial Network (ST-FVGAN), which not only considers spatio-temporal correlation utilizes idea generating Network, but also external factors introduces more comprehensive loss function. Specifically, model firstly constructs Generator network composed convolutional layer, residual block,...

As the main raw material in glass manufacturing, quartz sand and its elementary contents play an important role quality. contains a variety of impurities such as metal oxides, which makes traditional chemical analysis methods for determining composition challenging cumbersome, results are often influenced by skill level analysts various reagent factors. In this study, we investigate feasibility employing calibration-free laser-induced breakdown spectroscopy (CF-LIBS) simultaneous...

Traffic flow prediction has great significance for improving road traffic capacity and safety. However, in a certain area is usually affected by some factors such as weather, holidays neighboring areas. So, situation complicated difficult. How to use existing data information predict future the key this problem. In paper, we develop an accurate model based on dilated convolution — ST-MINet (Deep Spatio-Temporal Modified-Inception with Dilated Networks). We fully consider complexity,...

The use of deep learning methods to predict traffic flow in transportation systems has become a hot research project. existing predictive model method faces problems such as long calculation time and difficult data pre‐processing, especially for the prediction effect high area. In this study, authors propose novel framework ST‐ESNet, spatio‐temporal expand‐and‐squeeze networks, that designs several effective strategies considering complexity, non‐linearity uncertainty flow, better captures...

The prediction for average Material Removal Rate (MRR) in Chemical Mechanical Planarization (CMP) process has been recognized to be critical working condition identification semiconductor manufacturing. This paper focuses on feature engineering virtual metrology (VM) MRR prediction. A statistical extraction method is proposed, and a two-step hybrid selection designed select the relevant subset. Enhanced performance achieved PHM 16 data set.

Based on the traditional Wifi positioning algorithm, offline fingerprint database has a large number of redundant APs, which is easy to cause dimensionality disaster. this situation, paper proposes algorithm based SVM-kmeans-ReliefF-SAE. First, wifi divided into building and floors by SVM intra-floor region K-means algorithm. Then feature selection constructed carried out. The area removed ReliefF reduced established. Finally constructing deep neural network for through stack auto encoder....

Support vector machine (SVM) has been widely applied in the classification of remotely sensed image. How to reduce support number SVM classifier so as time still an important open problem, especially case mass data. To obtain fast with high accuracy, active learning schema is proposed based image classification. Experimental results synthetic data and multispectral images show that, compare classifiers trained whole training sample set a time, obtained by selection instances have much fewer...

There are lots of challenges for deriving urban land cover types high resolution optical imagery because spectral similarity different objects, mixed pixels, shadows buildings and large tree crowns. In order to reduce these uncertainties, recently, it's a trend the classification from multi-source sensors in field remote sensing. this study, hierarchical support vector machine (SVM) method was applied mapping, using WorldView-2 airborne Light Detection Ranging (LiDAR) data. The results...

Precise detection and analysis of trace elements in materials are critical for various industrial applications. This study integrates laser-induced breakdown spectroscopy (LIBS) with advanced machine learning algorithms—random forest (RF) and...

Dongba characters are hieroglyphs still used by The Naxi nationality in Yunnan province, and have a long history which is called "living fossils" of characters. In view the neglect influence similar on recognition accuracy absence construction application relevant models existing studies characters, this paper proposes multi-scale feature fusion based character model for First all, due to lack data set, two sets (including DB1424, common DBS20, set about characters) constructed. Then, design...

In a certain vehicle of Dongfeng Passenger Vehicle Company (DFPV), new design integrated hot stamping B-pillar reinforcement panel is introduced to replace the original two cold parts connecting by spot-weld. A further light-weight thickness thinning also attempted. All schemes are validated process simulation and side impact simulation. The results show that using hot-stamping steel same as design, performance anti-vehicle improved; an equivalent anti-side can be achieved appropriate...