Indoor localization plays an essential role in enabling location-based services (LBSs) for sensor networks wherever global navigation satellite systems (GNSSs) are unreachable. For deep-learning-based indoor methods, radio data extremely important accurate fingerprinting-based but often incomplete or not temporally and spatially sufficient acquisition. Motivated by the goal of leveraging existing fingerprints to further improve accuracy, this article, we propose a graph-based fingerprint...

Optimizing the deployment of large language models (LLMs) in edge computing environments is critical for enhancing privacy and computational efficiency. Toward efficient wireless LLM inference computing, this study comprehensively analyzes impact different splitting points mainstream open-source LLMs. On basis, introduces a framework taking inspiration from model-based reinforcement learning (MBRL) to determine optimal point across user equipment (UE). By incorporating reward surrogate...

This letter proposed a new configuration of neutral point-based common mode (CM) active electromagnetic interference (EMI) filter (NP-CM-AEF). It is achieved by setting the reference filtering system to point equipment under test. Therefore, CM noise could be sensed across and ground. Meanwhile, compensation signal built op-amp re-injected cancel initial noise. The main advantage NP-CM-AEF lies in that no more separation such as bulky passive transformer needed get pure whole becomes...

Nowdays, many Video-on-Demand (VoD) providers offer multiple-quality video streaming services to heterogeneous clients, called as multi-version VoD. Some researches focus on transcoding in real-time or layered encoding/decoding, but they are not widely used VoD industry. Storing multiple versions of the same is an easy solution, it consumes lots storage space. Although there also a few works about trading-off between and storage, did utilize relationships among different took popularity into...

Cloud service providers can offer users virtual machines (VMs) on-demand as a over the Internet. VMs running on top of physical machine (PM) share resources (CPU, memory, and/or bandwidth), and there may be great resource contention among them, which results in performance degradation. To prevent this, cloud need to study how place PMs efficiently. However, existing placement (VMP) methods mainly tried optimize instead VM performance. In this paper, we propose VMP method based models cloud....

Mental health problems of various populations during the COVID-19 pandemic have received high attention, but there is little research on mental Chinese civil servants. The present study investigated occupational stress, (i.e., anxiety, depression, and insomnia), social support, work-family conflict in servants pandemic. A total 327 Wenzhou city, China, participated an online survey, which collected data socio-demographic characteristics, conflict. Data were analyzed using non-parametric...

Media cloud has become a promising paradigm for deploying large-scale streaming media applications at reduced cost. Due to dynamic and diverse demands of users, presents two crucial characteristics: high resource consumption traffic among servers. Consequently, Virtual Machine (VM) migration in is highly required suit varying requirements the patterns. Moreover, such bandwidth-intensive needs cautious handling, especially internal Data Center Networks (DCN). However, existing management...

In medium-to-high power converters, full-bridge LLC converter has been a common choice due to its high efficiency and density. However, today, the EMI noise remains significant obstacle dv/dt large parasitic capacitors in transformers MOSFETs. Through eliminating displacement currents, presented technique can help maintain low-profile by avoiding bulky CM chokes, so as further increase This paper analyses main sources of FB presents cancellation eliminate without other passive components. A...

Long-term visual localization is an essential problem in robotics and computer vision, but remains challenging due to the environmental appearance changes caused by lighting seasons. While many existing works have attempted solve it directly learning invariant sparse keypoints descriptors match scenes, these approaches still struggle with adverse changes. Recent developments image transformations such as neural style transfer emerged alternative address gaps. In this work, we propose combine...

Lane detection is an important enabling or enhancing technology for many intelligent applications. A marker line can be segmented into several image blocks, each of which contains lane marking in the centre. This study proposes a learning-based method localisation via detecting and grouping such blocks. The authors model class using regionlet representation, block regarded as region represented by group regionlets. feature composed features extracted from regionlets contributes weak...

The global impact of the Covid-19 outbreak has threatened social and economic fabric communities, altered nature lives. Today, as major health risks COVID-19 seem to be receding, it is crucial pause, ponder plan progress after crisis better prepared for future. In view rising physical retail transactions, noticed that shopping crowd returned since people loathed feeling being cooped up in house during movement control orders. As such, a contactless intelligent cart could an ideal invention....

Fringing and proximity effects are the main reasons for serious ac copper loss in high-frequency planar inductors. In this article, a novel air-gapless core structure based on magnetic building blocks (MBBs) with magnetoresistance parallel symmetrically distributed relative to windings is proposed, which two printed circuit board (PCB) made of metal soft material low permeability three perpendicular ferrite high permeability. Equivalent models theoretical analysis proposed reveal that even...

The explosively-growing Internet of Things (IoT) is generating massive amount multivariate time series data. Anomaly detection for great importance in detecting and locating system failures, device malfunctions malicious attacks IoT. In this paper, we propose a new anomaly model, called Graph Attention-based Gated Recurrent Unit (GAGRU), to exploit the complex spatiotemporal features series. proposed GAGRU basically an attention-based (GRU), which embeds graph attention mechanism into GRU...

Planar PCB magnetic elements have become increasingly popular in the design of wide-band-gap converters based on SiC and GaN to reach a higher power density. The parasitic capacitance planar transformers often cannot be ignored because large area copper very thin dielectric between layers. A multi-conductor distributed equivalent model for transformer is proposed this paper establish more accurate high frequency which transformer. 3D winding structure can converted into 2D circuit through...

This paper presents a detailed electromagnetic interference (EMI) analysis and evaluation for three-phase three-level flying capacitor diode clamped dual active bridge (DAB) converter. Principle of the DAB is firstly discussed. Then potential EMI noise source as well coupling path was analyzed. It includes both common mode (CM) differential (DM) path. A hardware prototype built to verify simulation results. demonstrates theoretical in good accordance with experimental

Planar transformers (PTs) have attracted tremendous attentions in many applications requiring low-profile power converters, such as server center and telecom. Today, traditional PTs still face large common-mode (CM) noise due to inter-winding capacitances lager winding losses eddy current solid windings. Previous works show that there may exist a compromise between AC reduction CM reduction. The paper proposes novel high-performance PT based on currents cancellation full-bridge (FB) LLC...

A optimized magnetic core structure without air gaps and made of two different permeability materials for planar inductors is proposed. In this paper, the effect filling yokes with on improving field distribution around PCB windings discussed. order to optimize magneto motive force in inductors, magnetoresistance multi-permeability designed. The theoretical analysis based Ampere's loop theorem 3D FEA simulation results demonstrate effectiveness proposed reducing copper losses. And efficiency...

Abstract Due to the technological advancement in smart buildings and grid, there is increasing desire of managing energy demand achieve efficiency. In this context, building prediction has become an essential approach for measuring performance, assessing system efficiency, developing management strategies. study, two artificial intelligence techniques (i.e., ANN = neural networks SVR support vector regression) are examined used predict peak estimate usage office on a university campus based...

In this paper, the switching oscillations of GaN power devices can be predicted accurately by Ansys Q3D high frequency parameter simulation. A ringing optimization method for GaN-Based Driving Circuit is proved to feasible on half-bridge standard boards with and low input voltage. And will used hard scenarios.

Long-term visual localization is an essential problem in robotics and computer vision, but remains challenging due to the environmental appearance changes caused by lighting seasons. While many existing works have attempted solve it directly learning invariant sparse keypoints descriptors match scenes, these approaches still struggle with adverse changes. Recent developments image transformations such as neural style transfer emerged alternative address gaps. In this work, we propose combine...

Data acquisition in the large-scale Internet of Things (IoT) demands high-cost sensing and communication. In this paper, we develop a data-driven deep learning method for low-complexity sensor selection high-accuracy data reconstruction IoT. The proposed model is designed with sampling network neural (DNN) reconstruction, where Gumbel-max trick applied to ensure that loss function differentiable. With design an interconnected network, matrix its corresponding can be trained simultaneously....