At the edge of network close to source data, computing deploys computing, storage and other capabilities provide intelligent services in proximity offers low bandwidth consumption, latency high security. It satisfies requirements transmission bandwidth, real-time security for Internet Things (IoT) application scenarios. Based on IoT architecture, an (EC-IoT) reference architecture is proposed, which contained three layers: The end edge, cloud edge. Furthermore, key technologies artificial...

With the development of 3D scanning technology, a huge volume point cloud data has been collected at lower cost. The set is main burden during processing clouds, so simplification critical. aim to reduce while preserving features. Therefore, this paper provides new method for simplification, named FPPS (feature-preserved simplification). In FPPS, entropy defined, which quantifies features hidden in clouds. According entropy, key points including majority geometric are selected. Then, based...

In this article, a multi-layer convolutional neural network (ResNet-18) and Long Short-Term Memory Networks (LSTM) model is proposed for dynamic gesture recognition. The Soli dataset based on the signals collected by millimeter-wave radar. As sensor radar, radar has high positional accuracy can recognize small movements, to achieve ultimate goal of Human-Computer Interaction (HCI). A set velocity-range Doppler images transformed from original signal used as input model. Especially, ResNet-18...

In this study, the machine learning algorithm, K-Nearest Neighbor (KNN) is introduced for human action recognition. A wearable sensor employed to collect acceleration signals, which correspond six types of actions, including walking, walking upstairs, downstairs, sitting, standing and laying. The signals are normalized randomly divided into a training set test set. classifier trained using KNN It aimed recognize actions when given signals. We apply studies relationship between value k in...

UAVs have entered various fields of life, and target tracking is one the key technologies for UAV applications. However, there are challenges in practical applications, such as scale change video images, motion blur too high shooting angle leading to tracked objects being small, resulting poor accuracy. To cope with problem that small targets poorly by due less effective information output from deep residual network, a SiamMFF method introduces an efficient multi-scale feature fusion...

The YOLO-B infrared target detection algorithm is proposed to address the problems of incomplete extraction detailed features and missed wrong targets by YOLOv5s. improves SPPF YOLOv5s feature network proposing CSPPF structure increase sensory field model. Bifusion Neck invoked fuse shallow location information with deep semantic enhance capability Taking fully into account different concern for classification localization, efficient decoupled head used as prediction this algorithm, which...

Many Intelligent Transportation System applications require continuous and accurate positioning information of the vehicles traveling on road network. The common devices used for vehicle location are based Global Position (GPS), dead-reckoning (DR) sensors, map matching (MM). paper develops a fuzzy logic algorithm. algorithm is tested different networks varying complexity. results show that provides significant improvement over existing algorithms both in terms identifying correct links...

The network topology structure has great impact on its traffic capacity. Most of the biological and technological networks have been proved to exhibit scale-free feature. In real networks, hub links, which usually bear large loads, restricted transport efficiency complex networks. this paper, we focus bandwidth congestion propose a rewiring link strategy optimize dynamics Two strategies are compared simulation results show that our more advantages, makes degree distribution uniform, reducing...

As two-terminal passive fundamental circuit elements with memory characteristics, memristors are promising devices for applications such as neuromorphic systems, in-memory computing, and tunable RF/microwave circuits. The increasingly complex electromagnetic interference (EMI) environment threatens the reliability of memristor systems. However, various EMI signals' effects on still unclear. This paper selects continuous waves (CWs) signals. It provides a deeper insight into effect CWs driven...

Electrostatic discharge is a common phenomenon in daily life, and the electromagnetic pulse radiation generated during can cause great harm to power, communication, sensing, other equipment, resulting systems not working properly. To verify safety reliability of unmanned vehicle handling complex environments, interference effect electrostatic discharge, source on was studied. According mechanism, typical mode, model, an system tested for according ISO10605-2008 standard. Based measured data,...

Learning from the robust mechanism of biological nervous system is critical for creating reliable neuromorphic hardware. The homeostatic inhibition plasticity rule a to balance Hebbian and resist external environmental disturbances local damage. It plays an essential role in maintaining sparse firing patterns system. This paper imitates this provides fast inhibitory circuit with memristive synapse. Firstly, design method principle are demonstrated. Secondly, function was verified PSpice©...

As COVID-19 spread globally in 2020, the interaction between traffic dynamics and of epidemic has attracted much attention. However, controlling remains a challenging issue. In this paper, we have investigated relationships link-closure strategies traffic-driven spreading. It is found that spreading can be suppressed by targeted closing links small-degree nodes. contrast, large-degree nodes accelerate outbreak epidemic. These findings significance for

Accurate endpoint detection is important for speech procession. The problem nontrivial non-stationary backgrounds where noises may be introduced by the speaker, recording environment and transmission system. In this paper, an effective algorithm proposed improving signal processing performance in noisy environment. speech/pause discrimination method based on box-counting dimension. At same time, a dynamically updated threshold adaptive window are used to improve of algorithm. It...

The network structure acquires great influence on the traffic capacity for complex network. Since nodes with high degree usually bear large load in process of packet transmission, we propose a new multilayer model which can balance low-speed and high-speed layers. simulation results show that compared randomly select model, makes more balanced, thereby enhancing reducing possibility congestion. This gives full play to transmission advantages layer network, reduce consumption resources while...

In order to study traffic dynamics on multilayer networks, it is of great significance build a network model which can more exactly reflect the actual layered structure characteristics. this paper, three-layer in two logical layers are mapped one physical layer established, and capacities three kinds networks with different combinations compared. Simulation results show that when same random network, whose has optimal capacity, scale-free better capacity than networks.

Abstract Improving the traffic capacity of multilayer networks has been subject extensive interest and research. This paper proposes an effective strategy to improve three‐layer via process removing links linking nodes with large degrees. In order capacity, two different link deletion rules are proposed remove a portion logical layer without change physical topology structure. It is found that can capacity; furthermore, rule layers delete same number at time better than only one deleted....

Multiple illuminators of opportunity (IOs) and a large rotation angle are often required for current passive radar imaging techniques. However, demands long observation time, which cannot be implemented actual system. To overcome this disadvantage, paper proposes super-resolution framework with sparsity-inducing compressed sensing (CS) technique, allows fewer IOs smaller angle. In the proposed framework, sparsity-based is modeled mathematically, spatial frequencies amplitudes different...

The indoor localization of people is the key to realizing "smart city" applications, such as smart homes, elderly care, and an energy-saving grid. method based on electrostatic information a passive label-free technique with better balance accuracy, system power consumption, privacy protection, environmental friendliness. However, physical each actual application scenario different, resulting in transfer function from human potential sensor signal not being unique, thus limiting generality...