Nothing Special   »   [go: up one dir, main page]

Next Issue
Volume 11, May-2
Previous Issue
Volume 11, April-2
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 

Electronics, Volume 11, Issue 9 (May-1 2022) – 224 articles

Cover Story (view full-size image): Since aviation contributes to rising CO2 emissions, there is increasing interest in more electric aircraft. The concept aims at replacing mechanically, pneumatically, and hydraulically driven parts as well as the primary power sources with electrically driven components. Building efficient DC–DC converters is the key to unlocking the full potential of more electric aircraft. The converters must have high reliability, preferably with redundancies, in addition to fault tolerance. This work describes the design, hardware setup, and testing of an LLC-based converter that can be reconfigured in the case of faults. The converter in a reconfigured state remains operational and provides sufficient voltage and power without considerable addition of component numbers nor overstressing of components. View this paper
  • Issues are regarded as officially published after their release is announced to the table of contents alert mailing list.
  • You may sign up for e-mail alerts to receive table of contents of newly released issues.
  • PDF is the official format for papers published in both, html and pdf forms. To view the papers in pdf format, click on the "PDF Full-text" link, and use the free Adobe Reader to open them.
Order results
Result details
Section
Select all
Export citation of selected articles as:
24 pages, 10837 KiB  
Article
Analysis of the Applicable Range of the Standard Lambertian Model to Describe the Reflection in Visible Light Communication
by Xiangyang Zhang, Xiaodong Yang, Nan Zhao and Muhammad Bilal Khan
Electronics 2022, 11(9), 1514; https://doi.org/10.3390/electronics11091514 - 9 May 2022
Cited by 5 | Viewed by 2335
Abstract
The existing visible light communication simulation research on reflection is mainly based on the standard Lambertian model. In recent years, some papers have mentioned that the standard Lambertian model is too simplified and approximate to meet the actual situation. To solve this problem, [...] Read more.
The existing visible light communication simulation research on reflection is mainly based on the standard Lambertian model. In recent years, some papers have mentioned that the standard Lambertian model is too simplified and approximate to meet the actual situation. To solve this problem, a variety of more complex reflection models have been proposed. However, the more complex models require more computation. To balance computation and simulation accuracy, by consulting the literature, this study found that the standard Lambertian model has a certain requirement of the incident angle range to describe reflection on a wall covered in plaster. In this paper, the inappropriate index Q of the standard Lambertian model is defined, and then the relationship between Q and the light-emitting diode position with only the first reflection considered is determined through a preliminary calculation. The calculation shows that, in an empty room with plaster walls, and when the distance is greater than 0.685 m, the standard Lambertian model can be used; when the distance is less than 0.685 m, other, more complex models need to be adopted according to the actual situation. Full article
(This article belongs to the Section Microwave and Wireless Communications)
Show Figures

Figure 1

Figure 1
<p>Schematic diagram of normalized luminous intensity of the standard Lambertian light source.</p>
Full article ">Figure 2
<p>Schematic diagram of the normalized luminance of the standard Lambertian light source.</p>
Full article ">Figure 3
<p>Diffuse reflection polar diagram of a plaster wall.</p>
Full article ">Figure 4
<p>Schematic diagram of the calculation analysis of one infinitely long wall.</p>
Full article ">Figure 5
<p>Calculation results of the relationship between the <span class="html-italic">Q</span> and <span class="html-italic">x</span> in the situation of a single infinitely long wall.</p>
Full article ">Figure 6
<p>Schematic diagram of the calculation analysis of a corner between two infinitely long walls.</p>
Full article ">Figure 7
<p>Calculation results of the relationship between the <span class="html-italic">Q</span> and <span class="html-italic">x</span> in the situation of a corner between two infinitely long walls.</p>
Full article ">Figure 8
<p>Differences in the <span class="html-italic">Q</span> between <a href="#electronics-11-01514-f007" class="html-fig">Figure 7</a> and <a href="#electronics-11-01514-f005" class="html-fig">Figure 5</a>.</p>
Full article ">Figure 9
<p>Schematic diagram of a square empty room with an LED installed in the center of the ceiling.</p>
Full article ">Figure 10
<p>Calculation results of the relationship between the <span class="html-italic">Q</span> and <span class="html-italic">x</span> in the situation of a square empty room with an LED installed in the center of the ceiling.</p>
Full article ">Figure 11
<p>Calculation results of the relationship between the <span class="html-italic">Q</span> and LED position in a room size of 1 m × 1 m: (<b>a</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 10°; (<b>b</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 20°; (<b>c</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 30°; (<b>d</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 40°; (<b>e</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 50°; (<b>f</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 60°; (<b>g</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 70°.</p>
Full article ">Figure 12
<p>Calculation results of the relationship between the <span class="html-italic">Q</span> and LED position at a room size of 2 m × 2 m: (<b>a</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 10°; (<b>b</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 20°; (<b>c</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 30°; (<b>d</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 40°; (<b>e</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 50°; (<b>f</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 60°; (<b>g</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 70°.</p>
Full article ">Figure 12 Cont.
<p>Calculation results of the relationship between the <span class="html-italic">Q</span> and LED position at a room size of 2 m × 2 m: (<b>a</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 10°; (<b>b</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 20°; (<b>c</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 30°; (<b>d</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 40°; (<b>e</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 50°; (<b>f</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 60°; (<b>g</b>) <span class="html-italic">θ</span><sub>1/2</sub> = 70°.</p>
Full article ">Figure 13
<p>Contour map of <span class="html-italic">Q</span> in a room size of 2 m × 2 m: (<b>a</b>) <span class="html-italic">Q</span> = −10 dB; (<b>b</b>) <span class="html-italic">Q</span> = −20 dB; (<b>c</b>) <span class="html-italic">Q</span> = −30 dB.</p>
Full article ">Figure A1
<p>Symbolic schematic diagram.</p>
Full article ">Figure A2
<p>Ω<span class="html-italic"><sub>W</sub></span> geometric sketch.</p>
Full article ">Figure A3
<p>Ω<span class="html-italic"><sub>L</sub></span> geometric sketch 1.</p>
Full article ">Figure A4
<p>Ω<span class="html-italic"><sub>L</sub></span> geometric sketch 2.</p>
Full article ">Figure A5
<p>Ω<span class="html-italic"><sub>W</sub></span> geometric sketch.</p>
Full article ">Figure A6
<p>Ω<span class="html-italic"><sub>L</sub></span> geometric sketch 1.</p>
Full article ">Figure A7
<p>Ω<span class="html-italic"><sub>L</sub></span> geometric sketch 2.</p>
Full article ">Figure A8
<p>Sub-block sketch.</p>
Full article ">Figure A9
<p>Schematic diagram of block 8 <span class="html-italic">θ<sub>W</sub></span><sub>1</sub> and <span class="html-italic">θ<sub>W</sub></span><sub>2</sub>.</p>
Full article ">Figure A10
<p>Ω<span class="html-italic"><sub>W</sub></span> geometric sketch.</p>
Full article ">Figure A11
<p>Ω<span class="html-italic"><sub>L</sub></span> geometric sketch 1.</p>
Full article ">Figure A12
<p>Ω<span class="html-italic"><sub>L</sub></span> geometric sketch 2.</p>
Full article ">
14 pages, 596 KiB  
Article
Feedback ARMA Models versus Bayesian Models towards Securing OpenFlow Controllers for SDNs
by Wael Hosny Fouad Aly, Hassan Kanj, Nour Mostafa and Samer Alabed
Electronics 2022, 11(9), 1513; https://doi.org/10.3390/electronics11091513 - 9 May 2022
Cited by 3 | Viewed by 2336
Abstract
In software-defined networking (SDN), the control layers are moved away from the forwarding switching layers. SDN gives more programmability and flexibility to the controllers. OpenFlow is a protocol that gives access to the forwarding plane of a network switch or router over the [...] Read more.
In software-defined networking (SDN), the control layers are moved away from the forwarding switching layers. SDN gives more programmability and flexibility to the controllers. OpenFlow is a protocol that gives access to the forwarding plane of a network switch or router over the SDN network. OpenFlow uses a centralized control of network switches and routers in and SDN environment. Security is of major importance for SDN deployment. Transport layer security (TLS) is used to implement security for OpenFlow. This paper proposed a new technique to improve the security of the OpenFlow controller through modifying the TLS implementation. The proposed model is referred to as the secured feedback model using autoregressive moving average (ARMA) for SDN networks (SFBARMASDN). SFBARMASDN depended on computing the feedback for incoming packets based on ARMA models. Filtering techniques based on ARMA techniques were used to filter the packets and detect malicious packets that needed to be dropped. SFBARMASDN was compared to two reference models. One reference model was Bayesian-based and the other reference model was the standard OpenFlow. Full article
(This article belongs to the Special Issue Next Generation Networks and Systems Security)
Show Figures

Figure 1

Figure 1
<p>Block diagram for the feedback control system for the SFBARMA<sub>SDN</sub> model.</p>
Full article ">Figure 2
<p>Feedback control system for the DFBCP model.</p>
Full article ">Figure 3
<p>Modeling the SFBARMA<sub>SDN</sub> using feedback system.</p>
Full article ">Figure 4
<p>Root locus for the SFBARMA<sub>SDN</sub> model.</p>
Full article ">Figure 5
<p>SFBARMA<sub>SDN</sub> filtering.</p>
Full article ">Figure 6
<p>Experimental topology.</p>
Full article ">Figure 7
<p>Response of the controlled output parameter to step changes in the tuning parameter.</p>
Full article ">Figure 8
<p>Comparison of the fake packets detected for the standard OpenFlow, SSBN<sub>SDN</sub>, and SFBARMA<sub>SDN</sub>.</p>
Full article ">Figure 9
<p>Comparison of the processing time among the standard OpenFlow, SSBN<sub>SDN</sub>, and SFBARMA<sub>SDN</sub>.</p>
Full article ">
14 pages, 4505 KiB  
Article
Dual Phase Lock-In Amplifier with Photovoltaic Modules and Quasi-Invariant Common-Mode Signal
by Pavel Baranov, Ivan Zatonov and Bien Bui Duc
Electronics 2022, 11(9), 1512; https://doi.org/10.3390/electronics11091512 - 9 May 2022
Cited by 2 | Viewed by 2707
Abstract
In measuring small voltage deviations of about 1 µV and lower, it is important to separate useful signals from noise. The measurement of small voltage deviations between the amplitudes of two AC signals in wide frequency and voltage ranges, is performed by using [...] Read more.
In measuring small voltage deviations of about 1 µV and lower, it is important to separate useful signals from noise. The measurement of small voltage deviations between the amplitudes of two AC signals in wide frequency and voltage ranges, is performed by using lock-in amplifiers with the differential input as a comparator (null-indicator). The resolution and measurement accuracy of lock-in amplifiers is largely determined by the common-mode rejection ratio in their measuring channel. This work presents a developed differential signal recovery circuit with embedded photovoltaic modules, which allows implementing the dual phase lock-in amplifier with the differential input and quasi-invariant common-mode signal. The obtained metrological parameters of the proposed dual phase analog lock-in amplifier prove its applicability in comparing two signal amplitudes of 10√2 µV to 10√2 V in the frequency range of 20 Hz to 100 kHz with a 10 nV resolution. The proposed dual phase analog lock-in amplifier was characterized by a 130 to 185 dB CMRR in the frequency range up to 100 kHz with 20 nV/√Hz white noise. Full article
(This article belongs to the Collection Instrumentation, Noise, Reliability)
Show Figures

Figure 1

Figure 1
<p>Flowchart of commercially available lock-in amplifier.</p>
Full article ">Figure 2
<p>Schematic circuit of differential signal recovery.</p>
Full article ">Figure 3
<p>Voltage follower output stage with embedded photovoltaic modules.</p>
Full article ">Figure 4
<p>Amplitude frequency (<b>a</b>) and phase-shift-frequency (<b>b</b>) responses of the IA voltage follower output stage with embedded photovoltaic modules.</p>
Full article ">Figure 4 Cont.
<p>Amplitude frequency (<b>a</b>) and phase-shift-frequency (<b>b</b>) responses of the IA voltage follower output stage with embedded photovoltaic modules.</p>
Full article ">Figure 5
<p>The PGA207 output voltage with the voltage follower output stage with embedded solar cells.</p>
Full article ">Figure 6
<p>Prototype of differential signal recovery.</p>
Full article ">Figure 7
<p>Schematic circuit of CMRR detection in the prototype of differential signal recovery.</p>
Full article ">Figure 8
<p>Flowchart of the dual phase analog lock-in amplifier.</p>
Full article ">Figure 9
<p>External view of the PCB of the developed lock-in amplifier model.</p>
Full article ">Figure 10
<p>Allan variance of the lock-in amplifier.</p>
Full article ">Figure 11
<p>Flowchart of resolution estimation of the lock-in amplifier.</p>
Full article ">Figure 12
<p>Flowchart of the maximum amplitude of the compared voltages.</p>
Full article ">
15 pages, 3775 KiB  
Article
Empowering the Internet of Things Using Light Communication and Distributed Edge Computing
by Abdelhamied A. Ateya, Mona Mahmoud, Adel Zaghloul, Naglaa. F. Soliman and Ammar Muthanna
Electronics 2022, 11(9), 1511; https://doi.org/10.3390/electronics11091511 - 9 May 2022
Cited by 6 | Viewed by 2554
Abstract
With the rapid growth of connected devices, new issues emerge, which will be addressed by boosting capacity, improving energy efficiency, spectrum usage, and cost, besides offering improved scalability to handle the growing number of linked devices. This can be achieved by introducing new [...] Read more.
With the rapid growth of connected devices, new issues emerge, which will be addressed by boosting capacity, improving energy efficiency, spectrum usage, and cost, besides offering improved scalability to handle the growing number of linked devices. This can be achieved by introducing new technologies to the traditional Internet of Things (IoT) networks. Visible light communication (VLC) is a promising technology that enables bidirectional transmission over the visible light spectrum achieving many benefits, including ultra-high data rate, ultra-low latency, high spectral efficiency, and ultra-high reliability. Light Fidelity (LiFi) is a form of VLC that represents an efficient solution for many IoT applications and use cases, including indoor and outdoor applications. Distributed edge computing is another technology that can assist communications in IoT networks and enable the dense deployment of IoT devices. To this end, this work considers designing a general framework for IoT networks using LiFi and a distributed edge computing scheme. It aims to enable dense deployment, increase reliability and availability, and reduce the communication latency of IoT networks. To meet the demands, the proposed architecture makes use of MEC and fog computing. For dense deployment situations, a proof-of-concept of the created model is presented. The LiFi-integrated fog-MEC model is tested in a variety of conditions, and the findings show that the model is efficient. Full article
Show Figures

Figure 1

Figure 1
<p>Layering system of the developed LiFi-based IoT system.</p>
Full article ">Figure 2
<p>Hierarchal structure of the considered LiFi system.</p>
Full article ">Figure 3
<p>Main levels of the developed fog-MEC scheme.</p>
Full article ">Figure 4
<p>Levels of offloading of the proposed fog-MEC scheme.</p>
Full article ">Figure 5
<p>Average offloading latency of the three systems with the distance from the access point (AP).</p>
Full article ">Figure 6
<p>Communication overhead of the three considered systems for different numbers of end devices.</p>
Full article ">Figure 7
<p>Average time of handling IoT tasks for the proposed fog-MEC model compared to other existing systems.</p>
Full article ">Figure 8
<p>Percentage of blocked tasks for the proposed fog-MEC model compared to other existing systems.</p>
Full article ">Figure 9
<p>Total utilization efficiency of the proposed fog-MEC model compared to other existing systems.</p>
Full article ">
13 pages, 970 KiB  
Article
Few-Shot Learning with Collateral Location Coding and Single-Key Global Spatial Attention for Medical Image Classification
by Wenjing Shuai and Jianzhao Li
Electronics 2022, 11(9), 1510; https://doi.org/10.3390/electronics11091510 - 9 May 2022
Cited by 8 | Viewed by 2533
Abstract
Humans are born with the ability to learn quickly by discerning objects from a few samples, to acquire new skills in a short period of time, and to make decisions based on limited prior experience and knowledge. The existing deep learning models for [...] Read more.
Humans are born with the ability to learn quickly by discerning objects from a few samples, to acquire new skills in a short period of time, and to make decisions based on limited prior experience and knowledge. The existing deep learning models for medical image classification often rely on a large number of labeled training samples, whereas the fast learning ability of deep neural networks has failed to develop. In addition, it requires a large amount of time and computing resource to retrain the model when the deep model encounters classes it has never seen before. However, for healthcare applications, enabling a model to generalize new clinical scenarios is of great importance. The existing image classification methods cannot explicitly use the location information of the pixel, making them insensitive to cues related only to the location. Besides, they also rely on local convolution and cannot properly utilize global information, which is essential for image classification. To alleviate these problems, we propose a collateral location coding to help the network explicitly exploit the location information of each pixel to make it easier for the network to recognize cues related to location only, and a single-key global spatial attention is designed to make the pixels at each location perceive the global spatial information in a low-cost way. Experimental results on three medical image benchmark datasets demonstrate that our proposed algorithm outperforms the state-of-the-art approaches in both effectiveness and generalization ability. Full article
(This article belongs to the Special Issue Applications of Computational Intelligence)
Show Figures

Figure 1

Figure 1
<p>Training stage of our method. Our method follows the classical routine of training a classifier during training.</p>
Full article ">Figure 2
<p>Testing stage of our method. We fix the feature extractor and use the nearest class mean method to classify the image during testing.</p>
Full article ">Figure 3
<p>Single-key global spatial attention. We utilize a similar idea to self-attention, but the difference is that the spatial dimension of the key collapses in our approach, and each spatially located feature has to interact with only one feature instead of interacting with all features as in self-attention. We use the idea of weighting similar to SE attention [<a href="#B41-electronics-11-01510" class="html-bibr">41</a>] to weight the important features, instead of the feature generation method in self-attention.</p>
Full article ">Figure 4
<p>Medical image classification datasets. (<b>a</b>) DermaMNIST. (<b>b</b>) PathMNIST. (<b>c</b>) OrganMNIST (Axial).</p>
Full article ">Figure 5
<p>Validation accuracy curve on the DermaMNIST dataset.</p>
Full article ">Figure 6
<p>Loss curve on the DermaMNIST dataset.</p>
Full article ">
17 pages, 1477 KiB  
Article
Business Process Outcome Prediction Based on Deep Latent Factor Model
by Ke Lu, Xinjian Fang and Xianwen Fang
Electronics 2022, 11(9), 1509; https://doi.org/10.3390/electronics11091509 - 8 May 2022
Viewed by 1989
Abstract
Business process outcome prediction plays an essential role in business process monitoring. It continuously analyzes completed process events to predict the executing cases’ outcome. Most of the current outcome prediction focuses only on the activity information in historical logs and less on the [...] Read more.
Business process outcome prediction plays an essential role in business process monitoring. It continuously analyzes completed process events to predict the executing cases’ outcome. Most of the current outcome prediction focuses only on the activity information in historical logs and less on the embedded and implicit knowledge that has not been explicitly represented. To address these issues, this paper proposes a Deep Latent Factor Model Predictor (DLFM Predictor) for uncovering the implicit factors affecting system operation and predicting the final results of continuous operation cases based on log behavior characteristics and resource information. First, the event logs are analyzed from the control flow and resource perspectives to construct composite data. Then, the stack autoencoder model is trained to extract the data’s main feature components for improving the training data’s reliability. Next, we capture the implicit factors at the control and data flow levels among events and construct a deep implicit factor model to optimize the parameter settings. After that, an expansive prefix sequence construction method is proposed to realize the outcome prediction of online event streams. Finally, the proposed algorithm is implemented based on the mainstream framework of neural networks and evaluated by real logs. The results show that the algorithm performs well under several evaluation metrics. Full article
(This article belongs to the Section Computer Science & Engineering)
Show Figures

Figure 1

Figure 1
<p>Business Process Outcome Prediction Framework.</p>
Full article ">Figure 2
<p>Decomposition diagram of the case–event matrix.</p>
Full article ">Figure 3
<p>Distribution of trace lengths in the logs.</p>
Full article ">Figure 4
<p>Loss values for Receipt log.</p>
Full article ">Figure 5
<p>Analysis of the relationship between various resources in the Receipt log.</p>
Full article ">
27 pages, 3460 KiB  
Article
Delivering Extended Cellular Coverage and Capacity Using High-Altitude Platforms
by Steve Chukwuebuka Arum, David Grace and Paul Daniel Mitchell
Electronics 2022, 11(9), 1508; https://doi.org/10.3390/electronics11091508 - 7 May 2022
Cited by 3 | Viewed by 2404
Abstract
Interest in delivering cellular communication using a high-altitude platform (HAP) is increasing partly due to its wide coverage capability. In this paper, we formulate analytical expressions for estimating the area of a HAP beam footprint, average per-user capacity per cell, average spectral efficiency [...] Read more.
Interest in delivering cellular communication using a high-altitude platform (HAP) is increasing partly due to its wide coverage capability. In this paper, we formulate analytical expressions for estimating the area of a HAP beam footprint, average per-user capacity per cell, average spectral efficiency (SE) and average area spectral efficiency (ASE), which are relevant for radio network planning, especially within the context of HAP extended contiguous cellular coverage and capacity. To understand the practical implications, we propose an enhanced and validated recursive HAP antenna beam-pointing algorithm, which forms HAP cells over an extended service area while considering beam broadening and the degree of overlap between neighbouring beams. The performance of the extended contiguous cellular structure resulting from the algorithm is compared with other alternative schemes using the carrier-to-noise ratio (CNR) and carrier-to-interference-plus-noise ratio (CINR). Results show that there is a steep reduction in average ASE at the edge of coverage. The achievable coverage is limited by the minimum acceptable average ASE at the edge, among other factors. In addition, the results highlight that efficient beam management can be achieved using the enhanced and validated algorithm, which significantly improves user CNR, CINR, and coverage area compared with other benchmark schemes. A simulated annealing comparison verifies that such an algorithm is close to optimal. Full article
(This article belongs to the Special Issue Feature Papers in "Networks" Section)
Show Figures

Figure 1

Figure 1
<p>HAP phased array antenna beamforming for cellular coverage.</p>
Full article ">Figure 2
<p>Antenna element excitation for an <math display="inline"><semantics> <mrow> <mi>M</mi> <mo>×</mo> <mi>N</mi> </mrow> </semantics></math> antenna array. Elements in rows and columns are referred to as <span class="html-italic">x</span>- and <span class="html-italic">y</span>-axes elements with <math display="inline"><semantics> <msub> <mi>d</mi> <mi>x</mi> </msub> </semantics></math> and <math display="inline"><semantics> <msub> <mi>d</mi> <mi>y</mi> </msub> </semantics></math> distances apart, respectively. There is proportionality between the excitation amplitudes of the elements in both <math display="inline"><semantics> <mrow> <mi>x</mi> <mo>,</mo> <mi>y</mi> </mrow> </semantics></math> axes. The <math display="inline"><semantics> <mrow> <mo>(</mo> <mi>m</mi> <mo>,</mo> <mi>n</mi> <mo>)</mo> </mrow> </semantics></math>th element excitation amplitude is expressed as <math display="inline"><semantics> <mrow> <msub> <mi>I</mi> <mrow> <mi>m</mi> <mi>n</mi> </mrow> </msub> <mo>=</mo> <msub> <mi>I</mi> <mrow> <mi>m</mi> <mi>j</mi> </mrow> </msub> <msub> <mi>I</mi> <mrow> <mi>j</mi> <mi>n</mi> </mrow> </msub> </mrow> </semantics></math> [<a href="#B24-electronics-11-01508" class="html-bibr">24</a>].</p>
Full article ">Figure 3
<p>HAP elliptical cell geometry (semi-major axis).</p>
Full article ">Figure 4
<p>HAP elliptical cell geometry (semi-minor axis).</p>
Full article ">Figure 5
<p>HAP elliptical cell geometry (polar coordinates).</p>
Full article ">Figure 6
<p>The HAP cell geometry. Dotted arrow lines show the initial cell boresight when neighbouring cells only touch each other. Solid arrow lines show the new boresight after adjusting the initial pointing angle to add overlap between neighbouring cells. The angle between the solid and dashed lines for each cell highlight the angle <math display="inline"><semantics> <mi>ρ</mi> </semantics></math> between the boresight and cell edge, which is constant.</p>
Full article ">Figure 7
<p>The HAP cell boresight geometry.</p>
Full article ">Figure 8
<p>The HAP cell overlap geometry.</p>
Full article ">Figure 9
<p>The cell tessellation processes. (<b>a</b>) The first step with cells pointing at increasing distance on the <span class="html-italic">x</span>-axis. (<b>b</b>) The rotation of the first cells from (<b>a</b>) yielding another set of cells in the second step. (<b>c</b>,<b>d</b>) The third and fourth steps where new cells are deployed between the structure in (<b>b</b>).</p>
Full article ">Figure 10
<p>The full cellular structure for the HAP extended coverage.</p>
Full article ">Figure 11
<p>Average area spectral efficiency against distance of cell centre.</p>
Full article ">Figure 12
<p>CNR contour within cells of the equidistant scheme.</p>
Full article ">Figure 13
<p>CNR contour within cells of the equiangular scheme.</p>
Full article ">Figure 14
<p>CNR contour within cells using the proposed scheme.</p>
Full article ">Figure 15
<p>CINR distribution of the proposed scheme with equidistant and equiangular schemes.</p>
Full article ">Figure 16
<p>CINR distribution of the proposed scheme and schemes in [<a href="#B14-electronics-11-01508" class="html-bibr">14</a>].</p>
Full article ">Figure 17
<p>Platform altitude vs. 50th percentile user CINR.</p>
Full article ">Figure 18
<p>Overlap ratio vs. user allocation probability and 95th percentile user throughput.</p>
Full article ">Figure 19
<p>User throughput distribution of different cell-pointing schemes.</p>
Full article ">Figure 20
<p>Average CINR vs. capacity per user of the different schemes.</p>
Full article ">
23 pages, 1301 KiB  
Article
Machine Learning Models for Early Prediction of Sepsis on Large Healthcare Datasets
by Javier Enrique Camacho-Cogollo, Isis Bonet, Bladimir Gil and Ernesto Iadanza
Electronics 2022, 11(9), 1507; https://doi.org/10.3390/electronics11091507 - 7 May 2022
Cited by 18 | Viewed by 6553
Abstract
Sepsis is a highly lethal syndrome with heterogeneous clinical manifestation that can be hard to identify and treat. Early diagnosis and appropriate treatment are critical to reduce mortality and promote survival in suspected cases and improve the outcomes. Several screening prediction systems have [...] Read more.
Sepsis is a highly lethal syndrome with heterogeneous clinical manifestation that can be hard to identify and treat. Early diagnosis and appropriate treatment are critical to reduce mortality and promote survival in suspected cases and improve the outcomes. Several screening prediction systems have been proposed for evaluating the early detection of patient deterioration, but the efficacy is still limited at individual level. The increasing amount and the versatility of healthcare data suggest implementing machine learning techniques to develop models for predicting sepsis. This work presents an experimental study of some machine-learning-based models for sepsis prediction considering vital signs, laboratory test results, and demographics using Medical Information Mart for Intensive Care III (MIMIC-III) (v1.4), a publicly available dataset. The experimental results demonstrate an overall higher performance of machine learning models over the commonly used Sequential Organ Failure Assessment (SOFA) and Quick SOFA (qSOFA) scoring systems at the time of sepsis onset. Full article
Show Figures

Figure 1

Figure 1
<p>Architecture of the proposed methodology.</p>
Full article ">Figure 2
<p>Representation of the labeling method.</p>
Full article ">Figure 3
<p>Inclusion and exclusion criteria for patients.</p>
Full article ">Figure 4
<p>Sepsis-3 suspected infection cohort.</p>
Full article ">Figure 5
<p>Sepsis-3 onset hourly computation.</p>
Full article ">Figure 6
<p>Feature selection process.</p>
Full article ">Figure 7
<p>Prediction methodology.</p>
Full article ">Figure 8
<p>Stacking learning model.</p>
Full article ">Figure 9
<p>Detailed information about the training and test sets data samples.</p>
Full article ">Figure 10
<p>Confusion matrices of four models.</p>
Full article ">Figure 11
<p>AUC-ROC curves comparison based on information gain sets.</p>
Full article ">Figure 12
<p>Feature importance for sepsis prediction using XGBoost algorithm.</p>
Full article ">Figure 13
<p>SHAP bee swarm plot for XGBoost model.</p>
Full article ">Figure 14
<p>Summary comparison of metrics results of four ensemble models vs. SOFA and qSOFA using t = 1 h and look back features = 24 h.</p>
Full article ">
13 pages, 1861 KiB  
Article
A Novel Anti-Risk Method for Portfolio Trading Using Deep Reinforcement Learning
by Han Yue, Jiapeng Liu, Dongmei Tian and Qin Zhang
Electronics 2022, 11(9), 1506; https://doi.org/10.3390/electronics11091506 - 7 May 2022
Cited by 6 | Viewed by 3114
Abstract
In the past decade, the application of deep reinforcement learning (DRL) in portfolio management has attracted extensive attention. However, most classical RL algorithms do not consider the exogenous and noise of financial time series data, which may lead to treacherous trading decisions. To [...] Read more.
In the past decade, the application of deep reinforcement learning (DRL) in portfolio management has attracted extensive attention. However, most classical RL algorithms do not consider the exogenous and noise of financial time series data, which may lead to treacherous trading decisions. To address this issue, we propose a novel anti-risk portfolio trading method based on deep reinforcement learning (DRL). It consists of a stacked sparse denoising autoencoder (SSDAE) network and an actor–critic based reinforcement learning (RL) agent. SSDAE will carry out off-line training first, while the decoder will used for on-line feature extraction in each state. The SSDAE network is used for the noise resistance training of financial data. The actor–critic algorithm we use is advantage actor–critic (A2C) and consists of two networks: the actor network learns and implements an investment policy, which is then evaluated by the critic network to determine the best action plan by continuously redistributing various portfolio assets, taking Sharp ratio as the optimization function. Through extensive experiments, the results show that our proposed method is effective and superior to the Dow Jones Industrial Average index (DJIA), several variants of our proposed method, and a state-of-the-art (SOTA) method. Full article
(This article belongs to the Topic Machine and Deep Learning)
Show Figures

Figure 1

Figure 1
<p>The structure of SSDAE autoencoder network.</p>
Full article ">Figure 2
<p>Overview of the proposed model.</p>
Full article ">Figure 3
<p>The cumulative wealth on DJIA.</p>
Full article ">
30 pages, 13065 KiB  
Article
Efficient Colour Image Encryption Algorithm Using a New Fractional-Order Memcapacitive Hyperchaotic System
by Zain-Aldeen S. A. Rahman, Basil H. Jasim, Yasir I. A. Al-Yasir and Raed A. Abd-Alhameed
Electronics 2022, 11(9), 1505; https://doi.org/10.3390/electronics11091505 - 7 May 2022
Cited by 12 | Viewed by 2268
Abstract
In comparison with integer-order chaotic systems, fractional-order chaotic systems exhibit more complex dynamics. In recent years, research into fractional chaotic systems for the utilization of image cryptosystems has become increasingly highlighted. This paper describes the development, testing, numerical analysis, and electronic realization of [...] Read more.
In comparison with integer-order chaotic systems, fractional-order chaotic systems exhibit more complex dynamics. In recent years, research into fractional chaotic systems for the utilization of image cryptosystems has become increasingly highlighted. This paper describes the development, testing, numerical analysis, and electronic realization of a fractional-order memcapacitor. Then, a new four-dimensional (4D) fractional-order memcapacitive hyperchaotic system is suggested based on this memcapacitor. Analytically and numerically, the nonlinear dynamic properties of the hyperchaotic system have been explored, where various methods, including equilibrium points, phase portraits of chaotic attractors, bifurcation diagrams, and the Lyapunov exponent, are considered to demonstrate the chaos behaviour of this new hyperchaotic system. Consequently, an encryption cryptosystem algorithm is used for colour image encryption based on the chaotic behaviour of the memcapacitive model, where every pixel value of the original image is incorporated in the secret key to strengthen the encryption algorithm pirate anti-attack robustness. For generating the keyspace of that employed cryptosystem, the initial condition values, parameters, and fractional-order derivative value(s) (q) of the memcapacitive chaotic system are utilized. The common cryptanalysis metrics are verified in detail by histogram, keyspace, key sensitivity, correlation coefficient values, entropy, time efficiency, and comparisons with other recent related fieldwork in order to demonstrate the security level of the proposed cryptosystem approach. Finally, images of various sizes were encrypted and recovered to ensure that the utilized cryptosystem approach is capable of encrypting/decrypting images of various sizes. The obtained experimental results and security metrics analyses illustrate the excellent accuracy, high security, and perfect time efficiency of the utilized cryptosystem, which is highly resistant to various forms of pirate attacks. Full article
(This article belongs to the Special Issue RF/Microwave Circuits for 5G and Beyond)
Show Figures

Figure 1

Figure 1
<p><math display="inline"><semantics> <mrow> <msub> <mi>q</mi> <mi>M</mi> </msub> <mo>−</mo> <msub> <mi>v</mi> <mi>M</mi> </msub> </mrow> </semantics></math> characteristic curve of memcapacitor (9): (<b>a</b>) <span class="html-italic">f</span> = 1 Hz and various amplitude values; (<b>b</b>) <span class="html-italic">A<sub>m</sub></span> = 10 C and various frequency values.</p>
Full article ">Figure 2
<p>The <math display="inline"><semantics> <mrow> <msub> <mi>q</mi> <mi>M</mi> </msub> <mo>−</mo> <msub> <mi>v</mi> <mi>M</mi> </msub> </mrow> </semantics></math> hysteresis loop characteristics of the fractional-order memcapacitor (11).</p>
Full article ">Figure 3
<p>Chain fractance (CF) arrangement of a fractional order (<span class="html-italic">q</span> = 0.99).</p>
Full article ">Figure 4
<p>The electronic circuit layout of the fractional-order memcapacitor.</p>
Full article ">Figure 5
<p>The <math display="inline"><semantics> <mrow> <msub> <mi>q</mi> <mi>M</mi> </msub> <mo>−</mo> <msub> <mi>v</mi> <mi>M</mi> </msub> </mrow> </semantics></math> characteristic curve of fractional-order memcapacitor realized circuit.</p>
Full article ">Figure 6
<p>The symbolic layout of the fractional-order memristor.</p>
Full article ">Figure 7
<p>Fractional-order memcapacitive chaotic circuit.</p>
Full article ">Figure 8
<p>Chaotic attractors of the memcapacitive chaotic system: (<b>a</b>) <span class="html-italic">x-y</span>; (<b>b</b>) <span class="html-italic">y-z</span>; (<b>c</b>) <span class="html-italic">x-z</span>; (<b>d</b>) <span class="html-italic">x-u</span>; (<b>e</b>) <span class="html-italic">y-u</span>; (<b>f</b>) 3-D layout (<span class="html-italic">x</span>-<span class="html-italic">y</span>-<span class="html-italic">z</span>).</p>
Full article ">Figure 9
<p>Bifurcation diagram, effect of the parameter <span class="html-italic">α</span> on the system state variable <span class="html-italic">x</span>(<span class="html-italic">t</span>).</p>
Full article ">Figure 10
<p>Bifurcation diagram, influence of the fractional-order derivative value (<span class="html-italic">q</span>) on the system state variable <span class="html-italic">x</span>(<span class="html-italic">t</span>).</p>
Full article ">Figure 11
<p>Lyapunov exponents in contradiction of time.</p>
Full article ">Figure 12
<p>The system (17) Lyapunov exponents in contradiction to varying the system fractional-order derivative value (<span class="html-italic">q</span>).</p>
Full article ">Figure 13
<p>Block illustration of the image encryption algorithm using the fractional-order memcapacitive hyper chaotic model (17).</p>
Full article ">Figure 14
<p>Block diagram of the decryption algorithm.</p>
Full article ">Figure 15
<p>Experimental results of a plain “Lena.png” 512 × 512 color image: (<b>a</b>) the original image, (<b>b</b>) R band, (<b>c</b>) G band, (<b>d</b>) B band, (<b>e</b>) encrypted Lena image, (<b>f</b>) encrypted R band, (<b>g</b>) encrypted G band, (<b>h</b>) encrypted B band, (<b>i</b>) recovered (decrypted) Lena image, (<b>j</b>) recovered R band, (<b>k</b>) recovered G band, (<b>l</b>) recovered B band.</p>
Full article ">Figure 16
<p>The histogram of a Lena image: (<b>a</b>) plain image, (<b>b</b>) R band, (<b>c</b>) G band, (<b>d</b>) B band, (<b>e</b>) encrypted image, (<b>f</b>) encrypted R band, (<b>g</b>) encrypted G band, (<b>h</b>) encrypted B band, (<b>i</b>) recovered image, (<b>j</b>) recovered R band, (<b>k</b>) recovered G band, (<b>l</b>) recovered B band.</p>
Full article ">Figure 16 Cont.
<p>The histogram of a Lena image: (<b>a</b>) plain image, (<b>b</b>) R band, (<b>c</b>) G band, (<b>d</b>) B band, (<b>e</b>) encrypted image, (<b>f</b>) encrypted R band, (<b>g</b>) encrypted G band, (<b>h</b>) encrypted B band, (<b>i</b>) recovered image, (<b>j</b>) recovered R band, (<b>k</b>) recovered G band, (<b>l</b>) recovered B band.</p>
Full article ">Figure 17
<p>Sensitivity of key test: (<b>a</b>) original image, (<b>b</b>) encrypted image, (<b>c</b>) recovered image with variation (10<sup>−15</sup> is added to <span class="html-italic">q</span>) of the decryption keys, (<b>d</b>) difference between (<b>a</b>,<b>c</b>).</p>
Full article ">Figure 18
<p>Correlation analysis of a plain Lena image and its consistent encrypted image: (<b>a</b>,<b>d</b>) horizontal correlation, (<b>b</b>,<b>e</b>) vertical correlation, (<b>c</b>,<b>f</b>) diagonal correlation.</p>
Full article ">Figure 19
<p>Correlation analysis of the R band of a plain Lena image and its consistent encrypted image: (<b>a</b>,<b>d</b>) horizontal correlation, (<b>b</b>,<b>e</b>) vertical correlation, (<b>c</b>,<b>f</b>) diagonal correlation.</p>
Full article ">Figure 20
<p>Correlation analysis of the G band of a plain Lena image and its consistent encrypted image: (<b>a</b>,<b>d</b>) horizontal correlation, (<b>b</b>,<b>e</b>) vertical correlation, (<b>c</b>,<b>f</b>) diagonal correlation.</p>
Full article ">Figure 21
<p>Correlation analysis of the B band of a plain Lena image and its consistent encrypted image: (<b>a</b>,<b>d</b>) horizontal c arrangement, (<b>b</b>,<b>e</b>) vertical arrangement, (<b>c</b>,<b>f</b>) diagonal arrangement.</p>
Full article ">Figure 22
<p>Test of “macaws.jpg”, 300 × 309: (<b>a</b>) original image, (<b>b</b>) encrypted image, (<b>c</b>) recovered image, (<b>d</b>–<b>f</b>) histograms consistent with (<b>a</b>–<b>c</b>), respectively.</p>
Full article ">Figure 23
<p>Test of “fruits.jpg”, 236 × 235: (<b>a</b>) plain image, (<b>b</b>) encrypted image, (<b>c</b>) recovered image, (<b>d</b>–<b>f</b>) histograms consistent with (<b>a</b>–<b>c</b>), respectively.</p>
Full article ">
21 pages, 15097 KiB  
Article
SSA-SL Transformer for Bearing Fault Diagnosis under Noisy Factory Environments
by Seoyeong Lee and Jongpil Jeong
Electronics 2022, 11(9), 1504; https://doi.org/10.3390/electronics11091504 - 7 May 2022
Cited by 4 | Viewed by 2949
Abstract
Among the smart factory studies, we describe defect detection research conducted on bearings, which are elements of mechanical facilities. Bearing research has been consistently conducted in the past; however, most of the research has been limited to using existing artificial intelligence models. In [...] Read more.
Among the smart factory studies, we describe defect detection research conducted on bearings, which are elements of mechanical facilities. Bearing research has been consistently conducted in the past; however, most of the research has been limited to using existing artificial intelligence models. In addition, previous studies assumed the factories situated in the bearing defect research were insufficient. Therefore, a recent research was conducted that applied an artificial intelligence model and the factory environment. The transformer model was selected as state-of-the-art (SOTA) and was also applied to bearing research. Then, an experiment was conducted with Gaussian noise applied to assume a factory situation. The swish-LSTM transformer (Sl transformer) framework was constructed by redesigning the internal structure of the transformer using the swish activation function and long short-term memory (LSTM). Then, the data in noise were removed and reconstructed using the singular spectrum analysis (SSA) preprocessing method. Based on the SSA-Sl transformer framework, an experiment was performed by adding Gaussian noise to the Case Western Reserve University (CWRU) dataset. In the case of no noise, the Sl transformer showed more than 95% performance, and when noise was inserted, the SSA-Sl transformer showed better performance than the comparative artificial intelligence models. Full article
(This article belongs to the Special Issue Advances in Fault Detection/Diagnosis of Electrical Power Devices)
Show Figures

Figure 1

Figure 1
<p>The structure of SSA algorithm, (<b>a</b>) shows embedding process block, (<b>b</b>) shows SVD process block, (<b>c</b>) shows grouping process block and (<b>d</b>) shows averaging process block.</p>
Full article ">Figure 2
<p>The architecture of the vanilla transformer.</p>
Full article ">Figure 3
<p>The framework of SSA-Sl transformer.</p>
Full article ">Figure 4
<p>The architecture of the Sl transformer.</p>
Full article ">Figure 5
<p>The block diagram of the proposed research process.</p>
Full article ">Figure 6
<p>(<b>a</b>) The bearing simulator of CWRU and (<b>b</b>) its cross-section view.</p>
Full article ">Figure 7
<p>Components of rolling bearing.</p>
Full article ">Figure 8
<p>Schematic of an electric discharge machining (EDM) machine tool.</p>
Full article ">Figure 9
<p>Data comparison diagram, original data (inner race), noise, noise with original data, and reconstructed data from above.</p>
Full article ">Figure 10
<p>Figure shows how SSA decomposes noise-mixed data. The top right, it shows the first 11 elements, the bottom left—maintaining 339 components, and the bottom right—the original time series (TS), and the top left shows mixture of components.</p>
Full article ">Figure 11
<p>The result of SSA when L is 20, the middle shows first 11 components, the right shows remaining 339 components, the left shows mixture of components.</p>
Full article ">Figure 12
<p>The result of SSA when L is 100, the middle shows first 11 components, the right shows remaining 339 components, the left shows mixture of components.</p>
Full article ">Figure 13
<p>Comparison of GeLU, ReLU, and swish.</p>
Full article ">Figure 14
<p>The loss graph of Sl transformer. The <span class="html-italic">x</span>-axis describes epochs and the <span class="html-italic">y</span>-axis describes loss.</p>
Full article ">Figure 15
<p>This accuracy graph of Sl transformer. The <span class="html-italic">x</span>-axis describes epochs and the <span class="html-italic">y</span>-axis describes accuracy.</p>
Full article ">Figure 16
<p>The accuracy graph of non-applying SSA algorithm for noise data.</p>
Full article ">Figure 17
<p>The accuracy graph of applying SSA algorithm for noise data.</p>
Full article ">
15 pages, 6872 KiB  
Article
Current Collapse Conduction Losses Minimization in GaN Based PMSM Drive
by Pavel Skarolek, Ondrej Lipcak and Jiri Lettl
Electronics 2022, 11(9), 1503; https://doi.org/10.3390/electronics11091503 - 7 May 2022
Cited by 2 | Viewed by 1908
Abstract
The ever-increasing demands on the efficiency and power density of power electronics converters lead to the replacement of traditional silicon-based components with new structures. One of the promising technologies represents devices based on Gallium-Nitride (GaN). Compared to silicon transistors, GaN semiconductor switches offer [...] Read more.
The ever-increasing demands on the efficiency and power density of power electronics converters lead to the replacement of traditional silicon-based components with new structures. One of the promising technologies represents devices based on Gallium-Nitride (GaN). Compared to silicon transistors, GaN semiconductor switches offer superior performance in high-frequency converters, since their fast switching process significantly decreases the switching losses. However, when used in hard-switched converters such as voltage-source inverters (VSI) for motor control applications, GaN transistors increase the power dissipated due to the current conduction. The loss increase is caused by the current-collapse phenomenon, which increases the dynamic drain-source resistance of the device shortly after the turn-on. This disadvantage makes it hard for GaN converters to compete with other technologies in electric drives. Therefore, this paper offers a purely software-based solution to mitigate the negative consequences of the current-collapse phenomenon. The proposed method is based on the minimum pulse length optimization of the classical 7-segment space-vector modulation (SVM) and is verified within a field-oriented control (FOC) of a three-phase permanent magnet synchronous motor (PMSM) supplied by a two-level GaN VSI. The compensation in the control algorithm utilizes an offline measured look-up table dependent on the machine input power. Full article
(This article belongs to the Section Power Electronics)
Show Figures

Figure 1

Figure 1
<p>Three-phase two-level GaN voltage-source inverter.</p>
Full article ">Figure 2
<p>Current-collapse effect on on-state resistance.</p>
Full article ">Figure 3
<p>Problems connected with current-collapse, their causes, and possible solutions.</p>
Full article ">Figure 4
<p>SVPWM modulation (<b>a</b>) linear mode and (<b>b</b>) deformed modulation with 5% pulse limitation.</p>
Full article ">Figure 5
<p>Deformation of the voltage vector caused by the pulse length limiting set to 5% of the duty cycle at a load angle of 15°; the reference voltage components are denoted by an asterisk. (<b>a</b>) stationary <math display="inline"><semantics> <mrow> <mi>α</mi> <mi>β</mi> </mrow> </semantics></math> reference frame; (<b>b</b>) synchronous <math display="inline"><semantics> <mrow> <mi>d</mi> <mi>q</mi> </mrow> </semantics></math> reference frame.</p>
Full article ">Figure 6
<p>Electric drive loss (<math display="inline"><semantics> <mrow> <msub> <mi>P</mi> <mrow> <mi>loss</mi> </mrow> </msub> </mrow> </semantics></math>) distribution based on the limiting duty cycle <math display="inline"><semantics> <mrow> <msub> <mi>d</mi> <mrow> <mi>limit</mi> </mrow> </msub> </mrow> </semantics></math>.</p>
Full article ">Figure 7
<p>PMSM control scheme with added LUT-based pulse length limitation block. The values modified by the pulse-limiting algorithm are primed.</p>
Full article ">Figure 8
<p>Experimental workplace.</p>
Full article ">Figure 9
<p>Simplified schematic diagram of the experimental setup.</p>
Full article ">Figure 10
<p>Modulation with: (<b>a</b>) no pulse length limited, (<b>b</b>) with limiting value 5%, (<b>c</b>) with limiting value 10%, and (<b>d</b>) with limiting value 20% of the nominal duty cycle. Measured during 2000 RPM speed, 100 V DC-link voltage, and load resistance <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mrow> <mi>load</mi> </mrow> </msub> <mo>=</mo> <mn>100</mn> <mrow> <mo> </mo> <mi mathvariant="sans-serif">Ω</mi> </mrow> </mrow> </semantics></math>; <math display="inline"><semantics> <mrow> <msub> <mi>I</mi> <mi mathvariant="normal">a</mi> </msub> </mrow> </semantics></math> and <math display="inline"><semantics> <mrow> <msub> <mi>I</mi> <mi mathvariant="normal">b</mi> </msub> </mrow> </semantics></math> denote the current in phases “a” and “b”, respectively.</p>
Full article ">Figure 11
<p>Input DC-link current measured at different load at (<b>a</b>) 1500 RPM, (<b>b</b>) 2000 RPM, and (<b>c</b>) 2500 RPM; <math display="inline"><semantics> <mrow> <msub> <mi>I</mi> <mrow> <mi>dc</mi> </mrow> </msub> </mrow> </semantics></math> denotes the input DC-link current.</p>
Full article ">Figure 12
<p>Relative input DC-link current measured at different load at (<b>a</b>) 1500 RPM, (<b>b</b>) 2000 RPM, and (<b>c</b>) 2500 RPM; <math display="inline"><semantics> <mrow> <msub> <mi>I</mi> <mrow> <mi>dc</mi> </mrow> </msub> </mrow> </semantics></math> denotes the input DC-link current.</p>
Full article ">Figure 13
<p>Third-order polynomial function fitted to the measured data (limiting duty-cycle <math display="inline"><semantics> <mrow> <msub> <mi>d</mi> <mrow> <mi>limit</mi> </mrow> </msub> </mrow> </semantics></math> as a function of the inverter output power <math display="inline"><semantics> <mi>P</mi> </semantics></math> in per-units).</p>
Full article ">Figure 14
<p>Infrared camera measurement at 2000 RPM and <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mrow> <mi>load</mi> </mrow> </msub> <mo>=</mo> <mn>25</mn> <mrow> <mo> </mo> <mi mathvariant="sans-serif">Ω</mi> </mrow> </mrow> </semantics></math> (<b>a</b>) without pulse length limitation and (<b>b</b>) with pulse length limitation.</p>
Full article ">Figure 15
<p>Detail of the inverter board.</p>
Full article ">
20 pages, 966 KiB  
Review
Learning-Based Methods for Cyber Attacks Detection in IoT Systems: A Survey on Methods, Analysis, and Future Prospects
by Usman Inayat, Muhammad Fahad Zia, Sajid Mahmood, Haris M. Khalid and Mohamed Benbouzid
Electronics 2022, 11(9), 1502; https://doi.org/10.3390/electronics11091502 - 7 May 2022
Cited by 105 | Viewed by 10795
Abstract
Internet of Things (IoT) is a developing technology that provides the simplicity and benefits of exchanging data with other devices using the cloud or wireless networks. However, the changes and developments in the IoT environment are making IoT systems susceptible to cyber attacks [...] Read more.
Internet of Things (IoT) is a developing technology that provides the simplicity and benefits of exchanging data with other devices using the cloud or wireless networks. However, the changes and developments in the IoT environment are making IoT systems susceptible to cyber attacks which could possibly lead to malicious intrusions. The impacts of these intrusions could lead to physical and economical damages. This article primarily focuses on the IoT system/framework, the IoT, learning-based methods, and the difficulties faced by the IoT devices or systems after the occurrence of an attack. Learning-based methods are reviewed using different types of cyber attacks, such as denial-of-service (DoS), distributed denial-of-service (DDoS), probing, user-to-root (U2R), remote-to-local (R2L), botnet attack, spoofing, and man-in-the-middle (MITM) attacks. For learning-based methods, both machine and deep learning methods are presented and analyzed in relation to the detection of cyber attacks in IoT systems. A comprehensive list of publications to date in the literature is integrated to present a complete picture of various developments in this area. Finally, future research directions are also provided in the paper. Full article
(This article belongs to the Special Issue Resilience-Oriented Smart Grid Systems)
Show Figures

Figure 1

Figure 1
<p>General representation of an IoT system.</p>
Full article ">Figure 2
<p>Paper selection procedure.</p>
Full article ">Figure 3
<p>General structure of a decision tree algorithm.</p>
Full article ">Figure 4
<p>Working structure of a k-nearest neighbors algorithm.</p>
Full article ">Figure 5
<p>Working structure of a random forest algorithm.</p>
Full article ">Figure 6
<p>The general structure of a deep belief network.</p>
Full article ">Figure 7
<p>Working of adaptive boost algorithm.</p>
Full article ">
16 pages, 614 KiB  
Article
100 Gbps Dynamic Extensible Protocol Parser Based on an FPGA
by Ke Wang, Zhichuan Guo, Mangu Song and Meng Sha
Electronics 2022, 11(9), 1501; https://doi.org/10.3390/electronics11091501 - 7 May 2022
Cited by 1 | Viewed by 2208
Abstract
In order to facilitate the transition between networks and the integration of heterogeneous networks, the underlying link design of the current mainstream Information-Centric Networking (ICN) still considers the characteristics of the general network and extends the customized ICN protocol on this basis. This [...] Read more.
In order to facilitate the transition between networks and the integration of heterogeneous networks, the underlying link design of the current mainstream Information-Centric Networking (ICN) still considers the characteristics of the general network and extends the customized ICN protocol on this basis. This requires that the network transmission equipment can not only distinguish general network packets but also support the identification of ICN-specific protocols. However, traditional network protocol parsers are designed for specific network application scenarios, and it is difficult to flexibly expand new protocol parsing rules for different ICN network architectures. For this reason, we propose a general dynamic extensible protocol parser deployed on FPGA, which supports the real-time update of network protocol parsing rules by configuring extended protocol descriptors. At the same time, the multi-queue protocol management mechanism is adopted to realize the grouping management and rapid parsing of the extended protocol. The results demonstrate that the method can effectively support the protocol parsing of 100 Gbps high-speed network data packets and can dynamically update the protocol parsing rules under ultra-low latency. Compared with the current commercial programmable network equipment, this solution improves the protocol update efficiency by several orders of magnitude and better supports the online updating of network equipment. Full article
(This article belongs to the Section Networks)
Show Figures

Figure 1

Figure 1
<p>The abstract module of DEPP.</p>
Full article ">Figure 2
<p>Protocol extension diagram.</p>
Full article ">Figure 3
<p>Diagram of the queue.</p>
Full article ">Figure 4
<p>Bus protocol transformation diagram.</p>
Full article ">Figure 5
<p>Protocol update microstructure.</p>
Full article ">Figure 6
<p>The extended protocol detection process.</p>
Full article ">Figure 7
<p>Equipment connection diagram.</p>
Full article ">Figure 8
<p>The average update delay.</p>
Full article ">Figure 9
<p>Statistics of sending and receiving.</p>
Full article ">Figure 10
<p>Parsing latency variation.</p>
Full article ">
23 pages, 6905 KiB  
Article
Recognizing Students and Detecting Student Engagement with Real-Time Image Processing
by Mustafa Uğur Uçar and Ersin Özdemir
Electronics 2022, 11(9), 1500; https://doi.org/10.3390/electronics11091500 - 7 May 2022
Cited by 15 | Viewed by 4161
Abstract
With COVID-19, formal education was interrupted in all countries and the importance of distance learning has increased. It is possible to teach any lesson with various communication tools but it is difficult to know how far this lesson reaches to the students. In [...] Read more.
With COVID-19, formal education was interrupted in all countries and the importance of distance learning has increased. It is possible to teach any lesson with various communication tools but it is difficult to know how far this lesson reaches to the students. In this study, it is aimed to monitor the students in a classroom or in front of the computer with a camera in real time, recognizing their faces, their head poses, and scoring their distraction to detect student engagement based on their head poses and Eye Aspect Ratios. Distraction was determined by associating the students’ attention with looking at the teacher or the camera in the right direction. The success of the face recognition and head pose estimation was tested by using the UPNA Head Pose Database and, as a result of the conducted tests, the most successful result in face recognition was obtained with the Local Binary Patterns method with a 98.95% recognition rate. In the classification of student engagement as Engaged and Not Engaged, support vector machine gave results with 72.4% accuracy. The developed system will be used to recognize and monitor students in the classroom or in front of the computer, and to determine the course flow autonomously. Full article
(This article belongs to the Section Computer Science & Engineering)
Show Figures

Figure 1

Figure 1
<p>Masks used to obtain the gradient images on the <span class="html-italic">x</span>- and <span class="html-italic">y</span>-axis.</p>
Full article ">Figure 2
<p>Original image (<b>left</b>), gradient image on <span class="html-italic">x</span>-axis (<b>center</b>), gradient image on <span class="html-italic">y</span>-axis (<b>right</b>).</p>
Full article ">Figure 3
<p>Application of the Local Binary Patterns method.</p>
Full article ">Figure 4
<p>Circular LBP examples. (<b>a</b>) r = 1, <span class="html-italic">n</span> = 8 (<b>b</b>) r = 2, <span class="html-italic">n</span> = 16 (<b>c</b>) r = 2, <span class="html-italic">n</span> = 8.</p>
Full article ">Figure 5
<p>Forming the LBP histogram, the image was taken with the permission of Kelvin Salton do Prado from his article titled “Face Recognition: Understanding LBPH Algorithm” [<a href="#B60-electronics-11-01500" class="html-bibr">60</a>].</p>
Full article ">Figure 6
<p>Decision boundary and support vectors in SVM.</p>
Full article ">Figure 7
<p>Euler angles.</p>
Full article ">Figure 8
<p>The states of the landmarks according to the open and closed position of eyes.</p>
Full article ">Figure 9
<p>Sample images taken from the UPNA Head Pose Database.</p>
Full article ">Figure 10
<p>Flowchart of face recognition model preparation module of SECS.</p>
Full article ">Figure 11
<p>Flowchart of main module of SECS.</p>
Full article ">Figure 12
<p>SECS main module—instant total distraction results.</p>
Full article ">Figure 13
<p>SECS main module—various screenshots.</p>
Full article ">Figure 14
<p>Sample photos from the dataset formed to test the overall success of SECS (first line: “Engaged”; second line: “Not Engaged”).</p>
Full article ">
17 pages, 17522 KiB  
Article
Video Super-Resolution Using Multi-Scale and Non-Local Feature Fusion
by Yanghui Li, Hong Zhu, Qian Hou, Jing Wang and Wenhuan Wu
Electronics 2022, 11(9), 1499; https://doi.org/10.3390/electronics11091499 - 7 May 2022
Cited by 10 | Viewed by 2245
Abstract
Video super-resolution can generate corresponding to high-resolution video frames from a plurality of low-resolution video frames which have rich details and temporally consistency. Most current methods use two-level structure to reconstruct video frames by combining optical flow network and super-resolution network, but this [...] Read more.
Video super-resolution can generate corresponding to high-resolution video frames from a plurality of low-resolution video frames which have rich details and temporally consistency. Most current methods use two-level structure to reconstruct video frames by combining optical flow network and super-resolution network, but this process does not deeply mine the effective information contained in video frames. Therefore, we propose a video super-resolution method that combines non-local features and multi-scale features to extract more in-depth effective information contained in video frames. Our method obtains long-distance effective information by calculating the similarity between any two pixels in the video frame through the non-local module, extracts the local information covered by different scale convolution cores through the multi-scale feature fusion module, and fully fuses feature information using different connection modes of convolution cores. Experiments on different data sets show that the proposed method is superior to the existing methods in quality and quantity. Full article
(This article belongs to the Section Electronic Multimedia)
Show Figures

Figure 1

Figure 1
<p>Results of different reconstruction methods by ×4 scale on calendar video sequence.</p>
Full article ">Figure 2
<p>The framework of the proposed video super-resolution network.</p>
Full article ">Figure 3
<p>Non-local calculation in module. (<b>a</b>) The output matrix of feature map <span class="html-italic">X</span> through non-local module; (<b>b</b>) The calculation process of non-local module.</p>
Full article ">Figure 4
<p>The structure of the proposed MSFFB.</p>
Full article ">Figure 5
<p>Reconstruction of different video sequences on DAVIS-10 dataset with ×4 scale, the reconstruction results of different methods are compared. The vertical axis represents the PSNR value and the horizontal axis represents the video frame sequence.</p>
Full article ">Figure 6
<p>Texture of reconstruction results between proposed method and SOF-VSR method on city video sequence of Vid4 dataset. Figure (<b>a</b>) is the label image, and Figure (<b>b</b>) is the texture image processed by the Sobel operator of the label image; Figure (<b>c</b>) shows the dominant pixel distribution of the reconstruction result of SOF-VSR method in the texture region, and Figure (<b>d</b>) shows the dominant pixel distribution of the reconstruction result of proposed method in the texture region; Figures (<b>e1</b>,<b>f1</b>,<b>g1</b>) show the enlargement of the local texture area of the label image, Figures (<b>e2</b>,<b>f2</b>,<b>g2</b>) show the enlargement of the result local area of the SOF-VSR method, and Figures (<b>e3</b>,<b>f3</b>,<b>g3</b>) show the enlargement of the result local area of the proposed method.</p>
Full article ">Figure 7
<p>Visual comparisons of reconstruction results of ×4 scale on calendar and city video frames. The left side of the dashed line is the reconstruction result of the BI degradation model. The magnified area is the HR in sequence, and the reconstruction results of methods Bicubic, CARN, SOF-VSR, Proposed*, Proposed. The right side of the dashed line is the reconstruction result of the BD degradation model, and the zoomed-in area is HR, the reconstruction results of methods SOF-VSR, Proposed* and Proposed based on BD degradation model. The method marked with * only used non-local module.</p>
Full article ">Figure 8
<p>Visual comparisons of reconstruction results of ×4 scale on boxing and demolition video frames. The left side of the dashed line is the reconstruction result of the BI degradation model. The magnified area is the HR in sequence, and the reconstruction results of methods Bicubic, CARN, SOF-VSR, Proposed*, Proposed. The right side of the dashed line is the reconstruction result of the BD degradation model, and the zoomed-in area is HR, the reconstruction results of methods SOF-VSR, Proposed* and Proposed based on BD degradation model. The method marked with * only used non-local module.</p>
Full article ">
18 pages, 1094 KiB  
Article
Modeling Distributed MQTT Systems Using Multicommodity Flow Analysis
by Pietro Manzoni, Vittorio Maniezzo and Marco A. Boschetti
Electronics 2022, 11(9), 1498; https://doi.org/10.3390/electronics11091498 - 7 May 2022
Cited by 2 | Viewed by 1953
Abstract
The development of technologies that exploit the Internet of Things (IoT) paradigm has led to the increasingly widespread use of networks formed by different devices scattered throughout the territory. The Publish/Subscribe paradigm is one of the most used communication paradigms for applications of [...] Read more.
The development of technologies that exploit the Internet of Things (IoT) paradigm has led to the increasingly widespread use of networks formed by different devices scattered throughout the territory. The Publish/Subscribe paradigm is one of the most used communication paradigms for applications of this type. However, adopting these systems due to their centralized structure also leads to the emergence of various problems and limitations. For example, the broker is typically the single point of failure of the system: no communication is possible if the broker is unavailable. Moreover, they may not scale well considering the massive numbers of IoT devices forecasted in the future. Finally, a network architecture with a single central broker is partially at odds with the edge-oriented approach. This work focuses on the development of an adaptive topology control approach, able to find the most efficient network configuration maximizing the number of connections and reduce the waste of resources within it, starting from the definition of the devices and the connections between them present in the system. To reach the goal, we leverage an integer linear programming mathematical formulation, providing the basis to solve and optimize the problem of network configuration in contexts where the resources available to the devices are limited. Full article
(This article belongs to the Special Issue Feature Papers in "Networks" Section)
Show Figures

Figure 1

Figure 1
<p>A simple example showing the relations among clients and brokers.</p>
Full article ">Figure 2
<p>Flow network with multiple commodities.</p>
Full article ">Figure 3
<p>Min cost equivalent to max flow min cost network of <a href="#electronics-11-01498-f002" class="html-fig">Figure 2</a>.</p>
Full article ">Figure 4
<p>The full graph of instance TinyInstance.</p>
Full article ">Figure 5
<p>TinyInstance results.</p>
Full article ">Figure 6
<p>Instance case3 Lagrangian result.</p>
Full article ">Figure 7
<p>Instance case3 MIP result.</p>
Full article ">Figure 8
<p>Instance res0 result.</p>
Full article ">Figure 9
<p>Result on use case res7 using the Lagrangian formulation.</p>
Full article ">
15 pages, 34302 KiB  
Article
Remote Prototyping of FPGA-Based Devices in the IoT Concept during the COVID-19 Pandemic
by Michał Melosik, Mariusz Naumowicz, Marek Kropidłowski and Wieslaw Marszalek
Electronics 2022, 11(9), 1497; https://doi.org/10.3390/electronics11091497 - 7 May 2022
Cited by 2 | Viewed by 3071
Abstract
This paper presents a system for the remote design and testing of electronic circuits and devices with FPGAs during COVID-19 and similar lockdown periods when physical access to laboratories is not permitted. The system is based on the application of the IoT concept, [...] Read more.
This paper presents a system for the remote design and testing of electronic circuits and devices with FPGAs during COVID-19 and similar lockdown periods when physical access to laboratories is not permitted. The system is based on the application of the IoT concept, in which the final device is a test board with an FPGA chip. The system allows for remote visual inspection of the board and the devices linked to it in the laboratory. The system was developed for remote learning taking place during the lockdown periods at Poznan University of Technology (PUT) in Poland. The functionality of the system is confirmed by two demonstration tasks (the use of the temperature and humidity DHT11 sensor and the design of a generator of sinusoidal waveforms) for students in the fundamentals of digital design and synthesis courses. The proposed solution allows, in part, to bypass the time-consuming simulations, and accelerate the process of prototyping digital circuits by remotely accessing the infrastructure of the microelectronics laboratory. Full article
Show Figures

Figure 1

Figure 1
<p>System architecture for designing FPGAs operating under the IoT concept. The right side of the firewall shows the camera view that a student sees and a screen of the software used for remote programming. The left side of the firewall shows the conceptual connection between the PC and a prototype board with an FPGA chip and the connection of the webcam transmitting the image from the preview in the lab.</p>
Full article ">Figure 2
<p>One lab stand behind the firewall in <a href="#electronics-11-01497-f001" class="html-fig">Figure 1</a>. Computer A is connected to the FPGA test board C and camera B; oscilloscope D; and camera E is connected to computer F.</p>
Full article ">Figure 3
<p>Screen of the home page of the Fundamentals of Digital Design and Synthesis course. The interface of the ’ekursy’ platform is in the Polish language. 1—Course name, 2—announcement section, 3—help forum, 4 and 5—links to the audio and video communicators, 6—remote connection procedure instruction, 7—FPGA programming test files, 8—section with embedded twich.tv channels, 9—basic information about the course, 10—student course evaluation, 11—section related to remote access, 12—section with lectures on VHDL (VHSIC Hardware Description Language), 13—section with VHDL tasks, 14—section with lectures on programmable logic devices, 15—section with ISE tasks, 16—final project exam, 17 and 18—additional materials and tutorials, and 19—user name.</p>
Full article ">Figure 4
<p>The steps required for remote access (see the text below).</p>
Full article ">Figure 5
<p>Remote programming and simulation. All steps are performed remotely from the students’ home computers. The detection of devices on the test board is confirmed by creating a tree with DEV0-DEV2 positions in the JTAG Chain section. The FPGA to be programmed is the XC3S500E. To simulate a test project, the following steps are needed: 1—open the trigger window, 2—open the time waveform window and 3—force the measurement trigger signal.</p>
Full article ">Figure 6
<p>(<b>Top</b>) The measurement results of a counter based on a frequency divider frequency for a Spartan-3E Board. On the DataPort[0] channel, the frequency of 50 MHz is present. Subsequent channels show frequency waveforms with half the frequency of the previous channel. (<b>Bottom</b>) The Spartan 3e board configuration with a counter (based on a frequency divider).</p>
Full article ">Figure 7
<p>The unit designed for temperature and humidity sensing.</p>
Full article ">Figure 8
<p>(<b>Top</b>) The chipscope temperature (in °C) and humidity (in %) readings. (<b>Bottom</b>) Monitoring the reading of the physical sensor data via the OBS environment in the lab room.</p>
Full article ">Figure 9
<p>The DAC_TOP (designed sinusoidal waveform generator) along with its simulation results in different time intervals. The control signals are: <math display="inline"><semantics> <mrow> <mi>d</mi> <mi>a</mi> <mi>c</mi> <mo>_</mo> <mi>m</mi> <mi>o</mi> <mi>s</mi> <mi>i</mi> </mrow> </semantics></math>—sends commands and data to the DAC, <math display="inline"><semantics> <mrow> <mi>d</mi> <mi>a</mi> <mi>c</mi> <mo>_</mo> <mi>c</mi> <mi>l</mi> <mi>r</mi> </mrow> </semantics></math>—resets the DAC, <math display="inline"><semantics> <mrow> <mi>d</mi> <mi>a</mi> <mi>c</mi> <mo>_</mo> <mi>s</mi> <mi>c</mi> <mi>k</mi> </mrow> </semantics></math>—SPI clock to the DAC, <math display="inline"><semantics> <mrow> <mi>d</mi> <mi>a</mi> <mi>c</mi> <mo>_</mo> <mi>c</mi> <mi>s</mi> </mrow> </semantics></math>—selects the SPI device to communicate with in this case, it is the DAC.</p>
Full article ">Figure 10
<p>A preview of two channels. The design of the sine wave generator does not require the use of an LCD display. The first video channel is only for visual inspection of the oscilloscope probe connection.</p>
Full article ">
24 pages, 4851 KiB  
Article
Integration and Deployment of Cloud-Based Assistance System in Pharaon Large Scale Pilots—Experiences and Lessons Learned
by Andrej Grguric, Miran Mosmondor and Darko Huljenic
Electronics 2022, 11(9), 1496; https://doi.org/10.3390/electronics11091496 - 6 May 2022
Cited by 2 | Viewed by 2615
Abstract
The EU project Pharaon aims to support older European adults by integrating digital services, tools, interoperable open platforms, and devices. One of the objectives is to validate the integrated solutions in large-scale pilots. The integration of mature solutions and existing systems is one [...] Read more.
The EU project Pharaon aims to support older European adults by integrating digital services, tools, interoperable open platforms, and devices. One of the objectives is to validate the integrated solutions in large-scale pilots. The integration of mature solutions and existing systems is one of the preconditions for the successful realization of the different aims of the pilots. One such solution is an intelligent, privacy-aware home-care assistance system, SmartHabits. After briefly introducing the Pharaon and SmartHabits, the authors propose different Pharaon models in the Ambient/Active Assisted Living (AAL) domain, namely the Pharaon conceptual model, Pharaon reference logical architecture view, AAL ecosystem model, meta AAL ecosystem model, and Pharaon ecosystem and governance models. Building on the proposed models, the authors provide details of the holistic integration and deployment process of the SmartHabits system into the Pharaon ecosystem. Both technical and supporting integration challenges and activities are discussed. Technical activities, including syntactic and semantic integration and securing the transfer of the Pharaon sensitive data, are among the priorities. Supporting activities include achieving legal and regulatory compliance, device procurement, and use-case co-designing in COVID-19 conditions. Full article
Show Figures

Figure 1

Figure 1
<p>SmartHabits system overview. Adapted from Ref. [<a href="#B13-electronics-11-01496" class="html-bibr">13</a>].</p>
Full article ">Figure 2
<p>Pharaon and AAL conceptual model.</p>
Full article ">Figure 3
<p>Pharaon reference logical architecture view mapped to CREATE-IoT 3D RAM.</p>
Full article ">Figure 4
<p>Socio-technical AAL ecosystem model.</p>
Full article ">Figure 5
<p>Meta AAL ecosystem.</p>
Full article ">Figure 6
<p>Pharaon–AIoTES layers mapping.</p>
Full article ">Figure 7
<p>Pharaon ecosystem governance model.</p>
Full article ">Figure 8
<p>Pharaon ecosystem model.</p>
Full article ">Figure 9
<p>SmartHabits Platform mapped to Pharaon architecture.</p>
Full article ">Figure 10
<p>SmartHabits Platform-related data flow in Pharaon.</p>
Full article ">Figure 11
<p>Basic input data model mapping in SHP Data Ingestion.</p>
Full article ">Figure 12
<p>Activities in the AAL system integration and deployment process.</p>
Full article ">Figure 13
<p>Challenges and choices in the AAL domain.</p>
Full article ">
44 pages, 17214 KiB  
Article
Coronary Artery Disease Detection Model Based on Class Balancing Methods and LightGBM Algorithm
by Shasha Zhang, Yuyu Yuan, Zhonghua Yao, Jincui Yang, Xinyan Wang and Jianwei Tian
Electronics 2022, 11(9), 1495; https://doi.org/10.3390/electronics11091495 - 6 May 2022
Cited by 13 | Viewed by 3409
Abstract
Coronary artery disease (CAD) is a disease with high mortality and disability. By 2019, there were 197 million CAD patients in the world. Additionally, the number of disability-adjusted life years (DALYs) owing to CAD reached 182 million. It is widely known that the [...] Read more.
Coronary artery disease (CAD) is a disease with high mortality and disability. By 2019, there were 197 million CAD patients in the world. Additionally, the number of disability-adjusted life years (DALYs) owing to CAD reached 182 million. It is widely known that the early and accurate diagnosis of CAD is the most efficient method to reduce the damage of CAD. In medical practice, coronary angiography is considered to be the most reliable basis for CAD diagnosis. However, unfortunately, due to the limitation of inspection equipment and expert resources, many low- and middle-income countries do not have the ability to perform coronary angiography. This has led to a large loss of life and medical burden. Therefore, many researchers expect to realize the accurate diagnosis of CAD based on conventional medical examination data with the help of machine learning and data mining technology. The goal of this study is to propose a model for early, accurate and rapid detection of CAD based on common medical test data. This model took the classical logistic regression algorithm, which is the most commonly used in medical model research as the classifier. The advantages of feature selection and feature combination of tree models were used to solve the problem of manual feature engineering in logical regression. At the same time, in order to solve the class imbalance problem in Z-Alizadeh Sani dataset, five different class balancing methods were applied to balance the dataset. In addition, according to the characteristics of the dataset, we also adopted appropriate preprocessing methods. These methods significantly improved the classification performance of logistic regression classifier in terms of accuracy, recall, precision, F1 score, specificity and AUC when used for CAD detection. The best accuracy, recall, F1 score, precision, specificity and AUC were 94.7%, 94.8%, 94.8%, 95.3%, 94.5% and 0.98, respectively. Experiments and results have confirmed that, according to common medical examination data, our proposed model can accurately identify CAD patients in the early stage of CAD. Our proposed model can be used to help clinicians make diagnostic decisions in clinical practice. Full article
(This article belongs to the Special Issue Electronic Solutions for Artificial Intelligence Healthcare Volume II)
Show Figures

Figure 1

Figure 1
<p>The frame diagram of our proposed machine learning model.</p>
Full article ">Figure 2
<p>An example of an ECG waveform and an echocardiac image (pictures from the Internet).</p>
Full article ">Figure 3
<p>Standard for Borderline_SMOTE algorithm to assign minority samples to SAFE set, DANGER set and NOISE set.</p>
Full article ">Figure 4
<p>The decision mechanism of SVM_SMOTE algorithm for synthesizing new minority samples. A, B and C are minority class support vector samples.</p>
Full article ">Figure 5
<p>The process of realizing automatic feature combination based on lightGBM model.</p>
Full article ">Figure 6
<p>The histogram of <a href="#electronics-11-01495-t006" class="html-table">Table 6</a>.</p>
Full article ">Figure 7
<p>The ROC curve and AUC value of each fold in the 10-fold cross-validation obtained by classification models on the original dataset. (<b>a</b>–<b>d</b>) are the ROC curves of lightGBM, lr, lightGBM + lr and lightGBM + LR classifiers, respectively.</p>
Full article ">Figure 8
<p>The histogram of <a href="#electronics-11-01495-t007" class="html-table">Table 7</a>.</p>
Full article ">Figure 9
<p>The ROC curve and AUC value of each fold in the 10-fold cross-validation obtained by classification models on the dataset processed by SMOTE. (<b>a</b>–<b>d</b>) are the ROC curves of lightGBM, lr, lightGBM + lr and lightGBM + LR classifiers, respectively.</p>
Full article ">Figure 10
<p>The histogram of <a href="#electronics-11-01495-t008" class="html-table">Table 8</a>.</p>
Full article ">Figure 11
<p>The ROC curve and AUC value of each fold in the 10-fold cross-validation obtained by classification models on the dataset processed by Borderline_SMOTE. (<b>a</b>–<b>d</b>) are the ROC curves of lightGBM, lr, lightGBM + lr and lightGBM + LR classifiers, respectively.</p>
Full article ">Figure 12
<p>The histogram of <a href="#electronics-11-01495-t009" class="html-table">Table 9</a>.</p>
Full article ">Figure 13
<p>The ROC curve and AUC value of each fold in the 10-fold cross-validation obtained by classification models on the dataset processed by SMOTE_SVM. (<b>a</b>–<b>d</b>) are the ROC curves of lightGBM, lr, lightGBM + lr and lightGBM + LR classifiers, respectively.</p>
Full article ">Figure 14
<p>The histogram of <a href="#electronics-11-01495-t010" class="html-table">Table 10</a>.</p>
Full article ">Figure 15
<p>The ROC curve and AUC value of each fold in the 10-fold cross-validation obtained by classification models on the dataset processed by SMOTE_Tomek. (<b>a</b>–<b>d</b>) are the ROC curves of lightGBM, lr, lightGBM + lr and lightGBM + LR classifiers, respectively.</p>
Full article ">Figure 16
<p>The histogram of <a href="#electronics-11-01495-t011" class="html-table">Table 11</a>.</p>
Full article ">Figure 17
<p>The ROC curve and AUC value of each fold in the 10-fold cross-validation obtained by classification models on the dataset processed by SMOTENC. (<b>a</b>–<b>d</b>) are the ROC curves of lightGBM, lr, lightGBM + lr and lightGBM + LR classifiers, respectively.</p>
Full article ">Figure 18
<p>The histogram of <a href="#electronics-11-01495-t012" class="html-table">Table 12</a>.</p>
Full article ">Figure 19
<p>The ROC curve and AUC value of each fold in the 10-fold cross-validation obtained by classification models on the dataset processed by data standardization. (<b>a</b>–<b>d</b>) are the ROC curves of lightGBM, lr, lightGBM + lr and lightGBM + LR classifiers, respectively.</p>
Full article ">Figure 20
<p>The histogram of <a href="#electronics-11-01495-t013" class="html-table">Table 13</a>.</p>
Full article ">Figure 21
<p>The ROC curve and AUC value of each fold in the 10-fold cross-validation obtained by classification models on the dataset processed by data standardization and SMOTE. (<b>a</b>–<b>d</b>) are the ROC curves of lightGBM, lr, lightGBM + lr and lightGBM + LR classifiers, respectively.</p>
Full article ">Figure 22
<p>The histogram of <a href="#electronics-11-01495-t014" class="html-table">Table 14</a>.</p>
Full article ">Figure 23
<p>The ROC curve and AUC value of each fold in the 10-fold cross-validation obtained by classification models on the dataset processed by data standardization and Borderline_SMOTE. (<b>a</b>–<b>d</b>) are the ROC curves of lightGBM, lr, lightGBM + lr and lightGBM + LR classifiers, respectively.</p>
Full article ">Figure 24
<p>The histogram of <a href="#electronics-11-01495-t015" class="html-table">Table 15</a>.</p>
Full article ">Figure 25
<p>The ROC curve and AUC value of each fold in the 10-fold cross-validation obtained by classification models on the dataset processed by data standardization and SMOTE_SVM. (<b>a</b>–<b>d</b>) are the ROC curves of lightGBM, lr, lightGBM + lr and lightGBM + LR classifiers, respectively.</p>
Full article ">Figure 26
<p>The histogram of <a href="#electronics-11-01495-t016" class="html-table">Table 16</a>.</p>
Full article ">Figure 27
<p>The ROC curve and AUC value of each fold in the 10-fold cross-validation obtained by classification models on the dataset processed by data standardization and SMOTE_Tomek. (<b>a</b>–<b>d</b>) are the ROC curves of lightGBM, lr, lightGBM + lr and lightGBM + LR classifiers, respectively.</p>
Full article ">Figure 28
<p>The histogram of <a href="#electronics-11-01495-t017" class="html-table">Table 17</a>.</p>
Full article ">Figure 29
<p>The ROC curve and AUC value of each fold in the 10-fold cross-validation obtained by classification models on the dataset processed by data standardization and SMOTENC. (<b>a</b>–<b>d</b>) are the ROC curves of lightGBM, lr, lightGBM + lr and lightGBM + LR classifiers, respectively.</p>
Full article ">Figure 30
<p>The trend charts of the performance evaluation indexes with class balancing methods on original dataset and standardized dataset.</p>
Full article ">Figure 31
<p>The performance comparison of lightGBM + lr model on six groups of datasets. (<b>a</b>–<b>f</b>) correspond to the datasets without balancing and the datasets processed by SMOTE, BorderLine_SMOTE, SMOTE_SVM, SMOTE_Tomek and SMOTENC, respectively.</p>
Full article ">Figure 32
<p>The performance comparison of lightGBM + LR model on six groups of datasets. (<b>a</b>–<b>f</b>) correspond to the datasets without balancing and the datasets processed by SMOTE, BorderLine_SMOTE, SMOTE_SVM, SMOTE_Tomek and SMOTENC, respectively.</p>
Full article ">Figure 33
<p>Comparison of improvement effects of different modules on performance results.</p>
Full article ">Figure 34
<p>The change trend of the loss of the two models on the training set.</p>
Full article ">Figure 35
<p>The change trend of the loss of the two models on the test set.</p>
Full article ">
16 pages, 8253 KiB  
Article
Isolation and Grading of Faults in Battery Packs Based on Machine Learning Methods
by Sen Yang, Boran Xu and Hanlin Peng
Electronics 2022, 11(9), 1494; https://doi.org/10.3390/electronics11091494 - 6 May 2022
Cited by 4 | Viewed by 2400
Abstract
As the installed energy storage stations increase year by year, the safety of energy storage batteries has attracted the attention of industry and academia. In this work, an intelligent fault diagnosis scheme for series-connected battery packs based on wavelet characteristics of battery voltage [...] Read more.
As the installed energy storage stations increase year by year, the safety of energy storage batteries has attracted the attention of industry and academia. In this work, an intelligent fault diagnosis scheme for series-connected battery packs based on wavelet characteristics of battery voltage correlations is designed. First, the cross-cell voltages of multiple cells are preprocessed using an improved recursive Pearson correlation coefficient to capture the abnormal electrical signals. Secondly, the wavelet packet decomposition is applied to the coefficient series to obtain fault-related features from wavelet sub-bands, and the most representative characteristic principal components are extracted. Finally, the artificial neural network (ANN) and multi-classification relevance vector machine (mRVM) are employed to classify and evaluate fault mode and fault degree, respectively. Physical injection of external and internal short circuits, thermal damage, and loose connection failure is carried out to collect real fault data for model training and method validation. Experimental results show that the proposed method can effectively detect and locate different faults using the extracted fault features; mRVM is better than ANN in thermal fault diagnosis, while the overall diagnosis performance of ANN is better than mRVM. The success rates of fault isolation are 82% and 81%, and the success rates of fault grading are 98% and 90%, by ANN and mRVM, respectively. Full article
Show Figures

Figure 1

Figure 1
<p>Connection of battery pack and sensors.</p>
Full article ">Figure 2
<p>Schematic diagram of the proposed diagnosis framework.</p>
Full article ">Figure 3
<p>Physical view of experimental setup.</p>
Full article ">Figure 4
<p>Fault isolation performance based on ANN: (<b>a</b>) PCC; (<b>b</b>) ESC; (<b>c</b>) ISC; (<b>d</b>) THD; (<b>e</b>) No fault.</p>
Full article ">Figure 5
<p>Fault grading performance based on ANN: (<b>a</b>) Overview; (<b>b</b>) Critical fault; (<b>c</b>) Moderate fault; (<b>d</b>) Minor fault; (<b>e</b>) Healthy.</p>
Full article ">Figure 5 Cont.
<p>Fault grading performance based on ANN: (<b>a</b>) Overview; (<b>b</b>) Critical fault; (<b>c</b>) Moderate fault; (<b>d</b>) Minor fault; (<b>e</b>) Healthy.</p>
Full article ">Figure 6
<p>Fault isolation performance based on mRVM: (<b>a</b>) PCC; (<b>b</b>) ESC; (<b>c</b>) ISC; (<b>d</b>) THD; (<b>e</b>) No fault.</p>
Full article ">Figure 6 Cont.
<p>Fault isolation performance based on mRVM: (<b>a</b>) PCC; (<b>b</b>) ESC; (<b>c</b>) ISC; (<b>d</b>) THD; (<b>e</b>) No fault.</p>
Full article ">Figure 7
<p>Fault grading performance based on mRVM: (<b>a</b>) Overview; (<b>b</b>) Critical fault; (<b>c</b>) Moderate fault; (<b>d</b>) Minor fault; (<b>e</b>) Healthy.</p>
Full article ">
16 pages, 1758 KiB  
Article
A Design Methodology for Wideband Current-Reuse Receiver Front-Ends Aimed at Low-Power Applications
by Arash Abbasi and Frederic Nabki
Electronics 2022, 11(9), 1493; https://doi.org/10.3390/electronics11091493 - 6 May 2022
Cited by 3 | Viewed by 2122
Abstract
This work gives a design perspective on low-power and wideband RF-to-Baseband current-reuse receivers (CRR). The proposed CRR architecture design shares a single supply and biasing current among both LNTA and baseband circuits to reduce power consumption. The work discusses topology selection and a [...] Read more.
This work gives a design perspective on low-power and wideband RF-to-Baseband current-reuse receivers (CRR). The proposed CRR architecture design shares a single supply and biasing current among both LNTA and baseband circuits to reduce power consumption. The work discusses topology selection and a suitable design procedure of the low noise transconductance amplifier (LNTA), down-conversion passive-mixer, active-inductor (AI) and TIA circuits. Layout considerations are also discussed. The receiver was simulated in 130 nm CMOS technology and occupies an active area of 0.025 mm2. It achieves a wideband input matching of less than 10 dB from 0.8 GHz to 3.4 GHz. A conversion-gain of 39.5 dB, IIP3 of 28 dBm and a double-sideband (DSB) NF of 5.6 dB is simulated at a local-oscillator (LO) frequency of 2.4 GHz and an intermediate frequency (IF) of 10 MHz, while consuming 1.92 mA from a 1.2 V supply. Full article
(This article belongs to the Special Issue Design of Mixed Analog/Digital Circuits)
Show Figures

Figure 1

Figure 1
<p>Block diagram of the proposed current-reuse receiver. Adapted with permission from Ref. [<a href="#B12-electronics-11-01493" class="html-bibr">12</a>]. Copyright 2022 Frederic Nabki.</p>
Full article ">Figure 2
<p>(<b>a</b>) Test bench to optimize and characterize <math display="inline"><semantics> <msub> <mi mathvariant="normal">g</mi> <mi mathvariant="normal">m</mi> </msub> </semantics></math> and <math display="inline"><semantics> <msub> <mi mathvariant="normal">f</mi> <mi mathvariant="normal">T</mi> </msub> </semantics></math>, (<b>b</b>) resulting <math display="inline"><semantics> <msub> <mi mathvariant="normal">g</mi> <mi mathvariant="normal">m</mi> </msub> </semantics></math> plots, and (<b>c</b>) resulting <math display="inline"><semantics> <msub> <mi mathvariant="normal">f</mi> <mi mathvariant="normal">T</mi> </msub> </semantics></math> plots.</p>
Full article ">Figure 3
<p>Test bench to design and optimize the down-conversion mixer switches.</p>
Full article ">Figure 4
<p>(<b>a</b>) The noise figure and (<b>b</b>) the 1 dB compression point versus the W/L ratio for different gate voltages of the input transistors.</p>
Full article ">Figure 5
<p>(<b>a</b>) Test bench to design the active inductor and (<b>b</b>) its impedance magnitude over frequency with and without <math display="inline"><semantics> <msub> <mi>R</mi> <mi>S</mi> </msub> </semantics></math> considered.</p>
Full article ">Figure 6
<p>Magnitude of <math display="inline"><semantics> <msub> <mi mathvariant="normal">Z</mi> <mi>AI</mi> </msub> </semantics></math> for different values of <math display="inline"><semantics> <msub> <mi mathvariant="normal">C</mi> <mi>AI</mi> </msub> </semantics></math>.</p>
Full article ">Figure 7
<p>Magnitude of <math display="inline"><semantics> <msub> <mi mathvariant="normal">Z</mi> <mi>AI</mi> </msub> </semantics></math> for different W/L ratio multipliers of transistor <math display="inline"><semantics> <msub> <mi mathvariant="normal">M</mi> <mi>AI</mi> </msub> </semantics></math>. The unit W/L ratio is 10 <math display="inline"><semantics> <mi mathvariant="sans-serif">μ</mi> </semantics></math><math display="inline"><semantics> <mi mathvariant="normal">m</mi> </semantics></math>/ 260<math display="inline"><semantics> <mi mathvariant="normal">n</mi> </semantics></math><math display="inline"><semantics> <mi mathvariant="normal">m</mi> </semantics></math>.</p>
Full article ">Figure 8
<p>The NF and conversion gain versus the gate bias voltage of the LNTA.</p>
Full article ">Figure 9
<p>The input matching (<math display="inline"><semantics> <msub> <mi mathvariant="normal">S</mi> <mn>11</mn> </msub> </semantics></math>) versus frequency for several gate bias voltages of LNTA.</p>
Full article ">Figure 10
<p>The NF and conversion gain versus the gate bias voltage of the mixer.</p>
Full article ">Figure 11
<p>The NF and conversion gain versus supply variations.</p>
Full article ">Figure 12
<p>The NF and conversion gain versus the LO frequency.</p>
Full article ">Figure 13
<p>The layout of CRR front-end. Reprinted with permission from Ref. [<a href="#B12-electronics-11-01493" class="html-bibr">12</a>]. Copyright 2022, Frederic Nabki.</p>
Full article ">Figure 14
<p>The post-layout simulated NF versus the IF.</p>
Full article ">Figure 15
<p>The post-layout simulated fundamental and third-order intermodulation products versus the input power.</p>
Full article ">Figure 16
<p>Receiver post-layout simulated performance versus the RF signal.</p>
Full article ">Figure 17
<p>Comparison of the FoM to prior works versus the power consumption. Tedeschi, M. 2010 [<a href="#B2-electronics-11-01493" class="html-bibr">2</a>]. Lin, Z. 2014 [<a href="#B4-electronics-11-01493" class="html-bibr">4</a>]. Lin, Z. 2014* [<a href="#B5-electronics-11-01493" class="html-bibr">5</a>]. Kim, S. 2019 [<a href="#B8-electronics-11-01493" class="html-bibr">8</a>]. Ramella, M. 2017 [<a href="#B11-electronics-11-01493" class="html-bibr">11</a>]. Park, B. 2021 [<a href="#B10-electronics-11-01493" class="html-bibr">10</a>].</p>
Full article ">
15 pages, 364 KiB  
Article
On the Potential of MP-QUIC as Transport Layer Aggregator for Multiple Cellular Networks
by Zsolt Krämer, Felicián Németh, Attila Mihály, Sándor Molnár, István Pelle, Gergely Pongrácz and Donát Scharnitzky
Electronics 2022, 11(9), 1492; https://doi.org/10.3390/electronics11091492 - 6 May 2022
Cited by 3 | Viewed by 2627
Abstract
Multipath transport protocols have the ability to simultaneously utilize the different paths and thus outperform single-path solutions in terms of achievable goodput, latency, or reliability. In this paper our goal is to examine the potential of connecting a mobile terminal to multiple mobile [...] Read more.
Multipath transport protocols have the ability to simultaneously utilize the different paths and thus outperform single-path solutions in terms of achievable goodput, latency, or reliability. In this paper our goal is to examine the potential of connecting a mobile terminal to multiple mobile networks simultaneously in a dynamically changing environment. To achieve this, first we analyze a dataset obtained from an LTE drive test involving two operators. Then we study the performance of MP-QUIC, the multipath extension of QUIC, in a dynamic emulated environment generated from the collected traces. Our results show that MP-QUIC may leverage multiple available channels to provide uninterrupted connectivity, and a better overall goodput even when compared to using only the best available channel for communication. We also compare the MP-QUIC performance with MPTCP, identify challenges with the current protocol implementations to fill in the available aggregate capacity, and give insights on how the achievable throughput could be increased. Full article
(This article belongs to the Special Issue Telecommunication Networks)
Show Figures

Figure 1

Figure 1
<p>Evolution of the available downlink channel capacities of the primary and secondary cells over time: averaged to 1 s (<b>top</b>) and 5 s (<b>bottom</b>) time windows.</p>
Full article ">Figure 2
<p>Cumulative distribution of available cells.</p>
Full article ">Figure 3
<p>Propagation delay as a function of link capacity. The curve is fitted on two typical 5G values: <math display="inline"><semantics> <mrow> <mi>d</mi> <mo>(</mo> <mn>50</mn> <mo>)</mo> <mo>=</mo> <mn>20</mn> </mrow> </semantics></math>, <math display="inline"><semantics> <mrow> <mi>d</mi> <mo>(</mo> <mn>200</mn> <mo>)</mo> <mo>=</mo> <mn>100</mn> </mrow> </semantics></math>.</p>
Full article ">Figure 4
<p>Network setup.</p>
Full article ">Figure 5
<p>High-level internal operation of the emulation environment as a data-flow diagram.</p>
Full article ">Figure 6
<p>QUIC in an idealized environment.</p>
Full article ">Figure 7
<p>A measurement for Case 2, BBR CC. algorithm (line 9 of <a href="#electronics-11-01492-t005" class="html-table">Table 5</a>).</p>
Full article ">
12 pages, 501 KiB  
Article
Quality Enhancement of MPEG-H 3DA Binaural Rendering Using a Spectral Compensation Technique
by Hyeongi Moon and Young-cheol Park
Electronics 2022, 11(9), 1491; https://doi.org/10.3390/electronics11091491 - 6 May 2022
Viewed by 2107
Abstract
The latest MPEG standard, MPEG-H 3D Audio, employs the virtual loudspeaker rendering (VLR) technique to support virtual reality (VR) and augmented reality (AR). During the rendering, the binaural downmixing of channel signals often induces the so-called comb filter effect, an undesirable spectral artifact, [...] Read more.
The latest MPEG standard, MPEG-H 3D Audio, employs the virtual loudspeaker rendering (VLR) technique to support virtual reality (VR) and augmented reality (AR). During the rendering, the binaural downmixing of channel signals often induces the so-called comb filter effect, an undesirable spectral artifact, due to the phase difference between the binaural filters. In this paper, we propose an efficient algorithm that can mitigate such spectral artifacts. The proposed algorithm performs spectral compensation in both the panning gain and downmix signal domains depending on the frequency range. In the low-frequency bands where a band has a wider bandwidth than the critical-frequency scale, panning gains are directly compensated. In the high-frequency bands, where a band has a narrower bandwidth than the critical-frequency scale, a signal compensation similar to the active downmix is performed. As a result, the proposed algorithm optimizes the performance and the complexity within MPEG-H 3DA framework. By implementing the algorithm on MPEG-H 3DA BR, we verify that the additional computation complexity is minor. We also show that the proposed algorithm improves the subjective quality of MPEG-H 3DA BR significantly. Full article
(This article belongs to the Special Issue Applications of Audio and Acoustic Signal)
Show Figures

Figure 1

Figure 1
<p>A block diagram of the MPEG-H 3DA 3DoF rendering of audio objects and channel signals.</p>
Full article ">Figure 2
<p>Block diagram of the VLR-based BR in MPEG-H 3DA.</p>
Full article ">Figure 3
<p><math display="inline"><semantics> <mrow> <mrow> <mo>|</mo> </mrow> <msubsup> <mi>H</mi> <mrow> <mi>B</mi> <mi>R</mi> </mrow> <mi>L</mi> </msubsup> <mrow> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>|</mo> </mrow> </mrow> </semantics></math> (dotted line) for a virtual source at <math display="inline"><semantics> <mrow> <mo>+</mo> <msup> <mn>10</mn> <mo>∘</mo> </msup> </mrow> </semantics></math> to the right from the frontal direction, rendered using a pair of loudspeakers at <math display="inline"><semantics> <msup> <mn>30</mn> <mo>∘</mo> </msup> </semantics></math> and <math display="inline"><semantics> <msup> <mn>0</mn> <mo>∘</mo> </msup> </semantics></math> (<math display="inline"><semantics> <mrow> <mi>p</mi> <mo>=</mo> <mn>2</mn> </mrow> </semantics></math>). Left ear is located at the azimuth of <math display="inline"><semantics> <msup> <mn>90</mn> <mo>∘</mo> </msup> </semantics></math>.</p>
Full article ">Figure 4
<p>Approximation results of the HRTF magnitude in Equation (4) calculated at the 5th subband of the 64-band QMF operating on a 22.2-channel virtual loudspeaker layout. Loudspeakers are located at azimuth angles <math display="inline"><semantics> <msup> <mn>0</mn> <mo>∘</mo> </msup> </semantics></math>, <math display="inline"><semantics> <mrow> <mo>±</mo> <msup> <mn>30</mn> <mo>∘</mo> </msup> </mrow> </semantics></math>, <math display="inline"><semantics> <mrow> <mo>±</mo> <msup> <mn>60</mn> <mo>∘</mo> </msup> </mrow> </semantics></math>, <math display="inline"><semantics> <mrow> <mo>±</mo> <msup> <mn>90</mn> <mo>∘</mo> </msup> </mrow> </semantics></math>, <math display="inline"><semantics> <mrow> <mo>±</mo> <msup> <mn>135</mn> <mo>∘</mo> </msup> </mrow> </semantics></math>, and <math display="inline"><semantics> <msup> <mn>180</mn> <mo>∘</mo> </msup> </semantics></math> in the horizontal plane.</p>
Full article ">Figure 5
<p>Block diagram of the MPEG-H 3DA BR comprising the PGC and BSC.</p>
Full article ">Figure 6
<p>Measurement results of (<b>a</b>) SD and (<b>b</b>) ILD error with a <math display="inline"><semantics> <mrow> <mn>95</mn> <mo>%</mo> </mrow> </semantics></math> confidence interval, measured in octave bands.</p>
Full article ">Figure 7
<p>The MUSHRA test results.</p>
Full article ">
9 pages, 3299 KiB  
Article
High-Power Electromagnetic Pulse Effect Prediction for Vehicles Based on Convolutional Neural Network
by Le Cao, Shuai Hao, Yuan Zhao and Cheng Wang
Electronics 2022, 11(9), 1490; https://doi.org/10.3390/electronics11091490 - 6 May 2022
Cited by 2 | Viewed by 2313
Abstract
This study presents a prediction model for high-power electromagnetic pulse (HPEMP) effects on aboveground vehicles based on convolutional neural networks (CNNs). Since a vehicle is often located aboveground and is close to the air-groundhalf-space interface, the electromagnetic energy coupled into the [...] Read more.
This study presents a prediction model for high-power electromagnetic pulse (HPEMP) effects on aboveground vehicles based on convolutional neural networks (CNNs). Since a vehicle is often located aboveground and is close to the air-groundhalf-space interface, the electromagnetic energy coupled into the vehicle by the ground reflected waves cannot be ignored. Consequently, the analysis of the vehicle’s HPEMP effect is a composite electromagnetic scattering problem of the half-space and the vehicles above it, which is often analyzed using different half-space numerical methods. However, traditional numerical methods are often limited by the complexity of the actual half-space models and the high computational demands of complex targets. In this study, a prediction method is proposed based on a CNN, which can analyze the electric field and energy density under different incident conditions and half-space environments. Compared with the half-space finite-difference time-domain (FDTD) method, the accuracy of the prediction results was above 98% after completing the training of the CNN network, which proves the correctness and effectiveness of the method. In summary, the CNN prediction model in this study can provide a reference for evaluating the HPEMP effect on the target over a complex half-space medium. Full article
Show Figures

Figure 1

Figure 1
<p>Scattering problem of the target over half-space.</p>
Full article ">Figure 2
<p>Geometry of 3-D vehicle cab.</p>
Full article ">Figure 3
<p>Overall framework of the proposed CNN model.</p>
Full article ">Figure 4
<p>CNN structure of proposed method.</p>
Full article ">Figure 5
<p>Comparison of electric field predicted by CNN model to half-space FDTD method.</p>
Full article ">Figure 6
<p>Distribution pattern of the electric field in the <span class="html-italic">yoz</span> plane at different time steps. (<b>a</b>) Time step = 50; (<b>b</b>) time step = 100; (<b>c</b>) time step = 200; (<b>d</b>) time step = 300.</p>
Full article ">Figure 7
<p>LeNet-5, VGG-16, and GoogleNet models. (<b>a</b>) Accuracy; (<b>b</b>) loss.</p>
Full article ">Figure 8
<p>Comparison of coupling electric field in free space and half-space cases.</p>
Full article ">Figure 9
<p>Prediction of electric field with polarization angle.</p>
Full article ">
21 pages, 2524 KiB  
Article
BSTProv: Blockchain-Based Secure and Trustworthy Data Provenance Sharing
by Lian-Shan Sun, Xue Bai, Chao Zhang, Yang Li, Yong-Bin Zhang and Wen-Qiang Guo
Electronics 2022, 11(9), 1489; https://doi.org/10.3390/electronics11091489 - 6 May 2022
Cited by 6 | Viewed by 2529
Abstract
In the Big Data era, data provenance has become an important concern for enhancing the trustworthiness of key data that are rapidly generated and shared across organizations. Prevailing solutions employ authoritative centers to efficiently manage and share massive data. They are not suitable [...] Read more.
In the Big Data era, data provenance has become an important concern for enhancing the trustworthiness of key data that are rapidly generated and shared across organizations. Prevailing solutions employ authoritative centers to efficiently manage and share massive data. They are not suitable for secure and trustworthy decentralized data provenance sharing due to the inevitable dishonesty or failure of trusted centers. With the advent of the blockchain technology, embedding data provenance in immutable blocks is believed to be a promising solution. However, a provenance file, usually a directed acyclic graph, cannot be embedded in blocks as a whole because its size may exceed the limit of a block, and may include various sensitive information that can be legally accessed by different users. To this end, this paper proposed the BSTProv, a blockchain-based system for secure and trustworthy decentralized data provenance sharing. It enables secure and trustworthy provenance sharing by partitioning a large provenance graph into multiple small subgraphs and embedding the encrypted subgraphs instead of raw subgraphs or their hash values into immutable blocks of a consortium blockchain; it enables decentralized and flexible authorization by allowing each peer to define appropriate permissions for selectively sharing some sets of subgraphs to specific requesters; and it enables efficient cross-domain provenance composition and tracing by maintaining a high-level dependency structure among provenance graphs from different domains in smart contracts, and by locally storing, decrypting, and composing subgraphs obtained from the blockchain. Finally, a prototype is implemented on top of an Ethereum-based consortium blockchain and experiment results show the advantages of our approach. Full article
(This article belongs to the Section Computer Science & Engineering)
Show Figures

Figure 1

Figure 1
<p>Core structure of a provenance data model. Reprinted with permission from Ref. [<a href="#B14-electronics-11-01489" class="html-bibr">14</a>]. 2014, Wood, G.</p>
Full article ">Figure 2
<p>A provenance graph of email.</p>
Full article ">Figure 3
<p>Expansion of provenance subgraphs of <a href="#electronics-11-01489-f002" class="html-fig">Figure 2</a>.</p>
Full article ">Figure 4
<p>An architecture of the BSTProv system.</p>
Full article ">Figure 5
<p>Relations in the local database of a provenance owner.</p>
Full article ">Figure 6
<p>Snapshots of the storage of the smart contracts PGIC and PGAC.</p>
Full article ">Figure 7
<p>Snapshots of the storage of the smart contract ACC.</p>
Full article ">Figure 8
<p>Sequence diagram of provenance retrieval.</p>
Full article ">Figure 9
<p>Traceability analysis.</p>
Full article ">Figure 10
<p>Correlations between the storage cost and the number of subgraphs.</p>
Full article ">
15 pages, 5830 KiB  
Article
Analysis and Design of an S/PS−Compensated WPT System with Constant Current and Constant Voltage Charging
by Lin Yang, Zhi Geng, Shuai Jiang and Can Wang
Electronics 2022, 11(9), 1488; https://doi.org/10.3390/electronics11091488 - 6 May 2022
Cited by 6 | Viewed by 2641
Abstract
In recent years, more and more scholars have paid attention to the research of wireless power transfer (WPT) technology, and have achieved a lot of results. In practical charging application, ensuring that the WPT system can achieve constant current and constant voltage output [...] Read more.
In recent years, more and more scholars have paid attention to the research of wireless power transfer (WPT) technology, and have achieved a lot of results. In practical charging application, ensuring that the WPT system can achieve constant current and constant voltage output with zero phase angle (ZPA) operation is very important to prolong battery life and improve power transfer efficiency. This paper proposes an series/parallel series(S/PS)-compensated WPT system that can charge the battery load in constant current and constant voltage modes at two different frequency points through frequency switching. The proposed S/PS structure contains only three compensation capacitors, few compensation elements, simple structure, low economic cost, in addition, the secondary-side does not contain compensation inductor, ensuring the compactness of the secondary-side. An experimental prototype with an input voltage of 40 V is established, and the experiment proves that the model can obtain output voltage of 48 V and current of 2 A. Maximum system transmission efficiency of up to 92.48% The experimental results are consistent with the theoretical analysis results, which verifies the feasibility of the method. Full article
Show Figures

Figure 1

Figure 1
<p>Typical charging profile of the Li-ion battery.</p>
Full article ">Figure 2
<p>Circuit diagram of the S/PS−compensated WPT System.</p>
Full article ">Figure 3
<p>Equivalent circuit of the S/PS−compensated WPT System.</p>
Full article ">Figure 4
<p>Design approaches of the S/PS−compensated WPT system.</p>
Full article ">Figure 5
<p>Experimental model of the loosely coupled transformer.</p>
Full article ">Figure 6
<p>Voltage gain and phase of input impedance of the S/PS−compensated WPT system.</p>
Full article ">Figure 7
<p>Transconductance gain and phase of input impedance of the S/PS−compensated WPT system.</p>
Full article ">Figure 8
<p>Switch strategy of the S/PS−compensated WPT system.</p>
Full article ">Figure 9
<p>Experimental model of the S/PS−compensated WPT system.</p>
Full article ">Figure 10
<p>Experimental waveforms of <b><span class="html-italic">U</span></b><sub>P</sub>, <b><span class="html-italic">I</span></b><sub>P</sub> and <span class="html-italic">I</span><sub>RL</sub> in CC charging mode. (<b>a</b>) <span class="html-italic">R</span><sub>L</sub> = 5Ω (<b>b</b>) <span class="html-italic">R</span><sub>L</sub> = 15 Ω.</p>
Full article ">Figure 11
<p>Experimental waveforms of <b><span class="html-italic">U</span></b><sub>P</sub>, <b><span class="html-italic">I</span></b><sub>P</sub> and <span class="html-italic">U</span><sub>RL</sub> in CV charging mode. (<b>a</b>) <span class="html-italic">R</span><sub>L</sub> = 40 Ω (<b>b</b>) <span class="html-italic">R</span><sub>L</sub> = 60 Ω.</p>
Full article ">Figure 12
<p>The power transfer efficiency profile of the S/PS−compensated WPT system.</p>
Full article ">
17 pages, 2275 KiB  
Article
Designing an Intelligent Virtual Educational System to Improve the Efficiency of Primary Education in Developing Countries
by Vidal Alonso-Secades, Alfonso-José López-Rivero, Manuel Martín-Merino-Acera, Manuel-José Ruiz-García and Olga Arranz-García
Electronics 2022, 11(9), 1487; https://doi.org/10.3390/electronics11091487 - 6 May 2022
Cited by 7 | Viewed by 4557
Abstract
Incorporating technology into virtual education encourages educational institutions to demand a migration from the current learning management system towards an intelligent virtual educational system, seeking greater benefit by exploiting the data generated by students in their day-to-day activities. Therefore, the design of these [...] Read more.
Incorporating technology into virtual education encourages educational institutions to demand a migration from the current learning management system towards an intelligent virtual educational system, seeking greater benefit by exploiting the data generated by students in their day-to-day activities. Therefore, the design of these intelligent systems must be performed from a new perspective, which will take advantage of the new analytical functions provided by technologies such as artificial intelligence, big data, educational data mining techniques, and web analytics. This paper focuses on primary education in developing countries, showing the design of an intelligent virtual educational system to improve the efficiency of primary education through recommendations based on reliable data. The intelligent system is formed of four subsystems: data warehousing, analytical data processing, monitoring process and recommender system for educational agents. To illustrate this, the paper contains two dashboards that analyze, respectively, the digital resources usage time and an aggregate profile of teachers’ digital skills, in order to infer new activities that improve efficiency. These intelligent virtual educational systems focus the teaching–learning process on new forms of interaction on an educational future oriented to personalized teaching for the students, and new evaluation and teaching processes for each professor. Full article
(This article belongs to the Special Issue Recent Trends in Intelligent Systems)
Show Figures

Figure 1

Figure 1
<p>ProFuturo map.</p>
Full article ">Figure 2
<p>Data Cleaning Process.</p>
Full article ">Figure 3
<p>Intelligent virtual educational system design.</p>
Full article ">Figure 4
<p>Data warehousing subsystem.</p>
Full article ">Figure 5
<p>Analytical data processing subsystem.</p>
Full article ">Figure 6
<p>Monitoring process subsystem.</p>
Full article ">Figure 7
<p>Recommender systems.</p>
Full article ">Figure 8
<p>Digital resources usage time per month.</p>
Full article ">Figure 9
<p>Digital resources usage time per country.</p>
Full article ">Figure 10
<p>Digital resources usage time per educational institution.</p>
Full article ">Figure 11
<p>Distribution of teachers in each skill in accordance with the level of appropriation.</p>
Full article ">
24 pages, 7338 KiB  
Article
Unsupervised and Self-Supervised Tensor Train for Change Detection in Multitemporal Hyperspectral Images
by Muhammad Sohail, Haonan Wu, Zhao Chen and Guohua Liu
Electronics 2022, 11(9), 1486; https://doi.org/10.3390/electronics11091486 - 6 May 2022
Cited by 4 | Viewed by 3046
Abstract
Remote sensing change detection (CD) using multitemporal hyperspectral images (HSIs) provides detailed information on spectral–spatial changes and is useful in a variety of applications such as environmental monitoring, urban planning, and disaster detection. However, the high dimensionality and low spatial resolution of HSIs [...] Read more.
Remote sensing change detection (CD) using multitemporal hyperspectral images (HSIs) provides detailed information on spectral–spatial changes and is useful in a variety of applications such as environmental monitoring, urban planning, and disaster detection. However, the high dimensionality and low spatial resolution of HSIs do not only lead to expensive computation but also bring about inter-class homogeneity and inner-class heterogeneity. Meanwhile, labeled samples are difficult to obtain in reality as field investigation is expensive, which limits the application of supervised CD methods. In this paper, two algorithms for CD based on the tensor train (TT) decomposition are proposed and are called the unsupervised tensor train (UTT) and self-supervised tensor train (STT). TT uses a well-balanced matricization strategy to capture global correlations from tensors and can therefore effectively extract low-rank discriminative features, so the curse of the dimensionality and spectral variability of HSIs can be overcome. In addition, the two proposed methods are based on unsupervised and self-supervised learning, where no manual annotations are needed. Meanwhile, the ket-augmentation (KA) scheme is used to transform the low-order tensor into a high-order tensor while keeping the total number of entries the same. Therefore, high-order features with richer texture can be extracted without increasing computational complexity. Experimental results on four benchmark datasets show that the proposed methods outperformed their tensor counterpart, the tucker decomposition (TD), the higher-order singular value decomposition (HOSVD), and some other state-of-the-art approaches. For the Yancheng dataset, OA and KAPPA of UTT reached as high as 98.11% and 0.9536, respectively, while OA and KAPPA of STT were at 98.20% and 0.9561, respectively. Full article
(This article belongs to the Special Issue Deep Learning for Big Data Processing)
Show Figures

Figure 1

Figure 1
<p>Graphical representation of tensor manipulations. (<b>a</b>) Basic building blocks for tensor network diagrams; (<b>b</b>) mode-2 matricization (left), mode-(1,2) matricization (middle), and vectorization (right) of a 3-order tensor <math display="inline"><semantics> <mrow> <munder accentunder="true"> <mstyle mathvariant="bold" mathsize="normal"> <mi>G</mi> </mstyle> <mo stretchy="true">¯</mo> </munder> <mo>∈</mo> <msup> <mo>ℝ</mo> <mrow> <msub> <mi>I</mi> <mn>1</mn> </msub> <mo>×</mo> <msub> <mi>I</mi> <mn>2</mn> </msub> <mo>×</mo> <msub> <mi>I</mi> <mn>3</mn> </msub> </mrow> </msup> </mrow> </semantics></math>.</p>
Full article ">Figure 2
<p>Visualization of a 5-order tensor train decomposition in graphical notations.</p>
Full article ">Figure 3
<p>Framework of the proposed UTT for multitemporal HSI change detection.</p>
Full article ">Figure 4
<p>Framework of the proposed STT for multitemporal HSI change detection. (<b>a</b>) Overall framework of STT, (<b>b</b>) tensor train decomposition (TTD) layer, (<b>c</b>) tensor train output (TTO) layer.</p>
Full article ">Figure 5
<p>False-color composition of (<b>a</b>,<b>e</b>) Yancheng, (<b>b</b>,<b>f</b>) Bay Area, (<b>c</b>,<b>g</b>) River, and (<b>d</b>,<b>h</b>) Hermiston city.</p>
Full article ">Figure 6
<p>Binary detection results in the Yancheng dataset: (<b>a</b>) ground truth, (<b>b</b>) LSCD, (<b>c</b>) ASCD, (<b>d</b>) HOSVD, I TDRD, (<b>f</b>) PCANet, (<b>g</b>) DSFANet, (<b>h</b>) HI-DRL, (<b>i</b>) SSTN, (<b>j</b>) UTT-SVD, (<b>k</b>) UTT-noSVD, and (<b>l</b>) STT.</p>
Full article ">Figure 7
<p>Binary detection results in the River dataset: (<b>a</b>) ground truth, (<b>b</b>) LSCD, (<b>c</b>) ASCD, (<b>d</b>) HOSVD, (<b>e</b>) TDRD, (<b>f</b>) PCANet, (<b>g</b>) DSFANet, (<b>h</b>) HI-DRL, (<b>i</b>) SSTN, (<b>j</b>) UTT-SVD, (<b>k</b>) UTT-noSVD, and (<b>l</b>) STT.</p>
Full article ">Figure 8
<p>Binary detection results in the Bay Area dataset: (<b>a</b>) ground truth, (<b>b</b>) LSCD, (<b>c</b>) ASCD, (<b>d</b>) HOSVD, (<b>e</b>) TDRD, (<b>f</b>) PCANet, (<b>g</b>) DSFANet, (<b>h</b>) HI-DRL, (<b>i</b>) SSTN, (<b>j</b>) UTT-SVD, (<b>k</b>) UTT-noSVD, and (<b>l</b>) STT.</p>
Full article ">Figure 9
<p>Binary detection results in the Hermiston dataset: (<b>a</b>) ground truth, (<b>b</b>) LSCD, (<b>c</b>) ASCD, (<b>d</b>) HOSVD, (<b>e</b>) TDRD, (<b>f</b>) PCANet, (<b>g</b>) DSFANet, (<b>h</b>) HI-DRL, (<b>i</b>) SSTN, (<b>j</b>) UTT-SVD, (<b>k</b>) UTT-noSVD, and (<b>l</b>) STT.</p>
Full article ">Figure 10
<p>ROCs of different methods for the (<b>a</b>) Yancheng dataset (<b>b</b>) Hermiston dataset (<b>c</b>) River dataset (<b>d</b>) Bay Area dataset.</p>
Full article ">Figure 11
<p>Scatter plots produced by t-SNE of all samples from the Yancheng dataset (<b>a</b>,<b>b</b>) and Hermiston dataset (<b>c</b>,<b>d</b>). Red circles represent unchanged samples, and blue circles represent changed samples. Yancheng (<b>a</b>) original difference tensor, (<b>b</b>) features learnt by STT, Hermiston (<b>c</b>) original difference tensor, (<b>d</b>) features learnt by STT.</p>
Full article ">Figure 12
<p>Ablation study of STT. River dataset (<b>a</b>) ground truth, (<b>b</b>) with 5-dimensional input patches, (<b>c</b>) with 3-dimensional input patches; Bay Area dataset (<b>d</b>) ground truth, (<b>e</b>) with 5-dimensional input patches, and (<b>f</b>) with 3-dimensional input patches.</p>
Full article ">Figure 13
<p>Convergence curve of STT for different datasets. (<b>a</b>) Yancheng dataset, (<b>b</b>) Hermiston dataset, (<b>c</b>) River dataset, (<b>d</b>) Bay Area dataset.</p>
Full article ">
16 pages, 1732 KiB  
Article
Mathematical Modelling of the Influence of Parasitic Capacitances of the Components of the Logarithmic Analogue-to-Digital Converter (LADC) with a Successive Approximation on Switched Capacitors for Increasing Accuracy of Conversion
by Zynoviy Mychuda, Igor Zhuravel, Lesia Mychuda, Adam Szcześniak, Zbigniew Szcześniak and Hanna Yelisieieva
Electronics 2022, 11(9), 1485; https://doi.org/10.3390/electronics11091485 - 6 May 2022
Cited by 7 | Viewed by 1634
Abstract
This paper presents an analysis of the influence of parasitic inter-electrode capacitances of the components of logarithmic analogue-to-digital converters with successive approximation with a variable logarithm base. Mathematical models of converter errors were developed and analyzed taking into account the parameters of modern [...] Read more.
This paper presents an analysis of the influence of parasitic inter-electrode capacitances of the components of logarithmic analogue-to-digital converters with successive approximation with a variable logarithm base. Mathematical models of converter errors were developed and analyzed taking into account the parameters of modern components. It has been shown that to achieve satisfactory accuracy for the 16 bit LADC, the capacitance of the capacitor cell must not be less than 10 nF; for the 12 bit LADC, 1 nF is sufficient. Full article
(This article belongs to the Special Issue Advances on Analog-to-Digital and Digital-to-Analog Converters)
Show Figures

Figure 1

Figure 1
<p>Simplified functional diagram of the LADC with successive approximation on switchable PK capacitors with a variable logarithm base.</p>
Full article ">Figure 2
<p>A model of the LADC with successive approximation on SC with a variable logarithm base that takes into account parasitic capacitances.</p>
Full article ">Figure 3
<p>A model of the LADC with successive approximation on switched capacitors which takes into account parasitic charge transfer.</p>
Full article ">Figure 4
<p>A model of the LADC with successive approximation on SC which takes into account control voltage transmission.</p>
Full article ">Figure 5
<p>The absolute summative error <math display="inline"><semantics> <mrow> <mo>Δ</mo> <msub> <mi>U</mi> <mrow> <mi>m</mi> <mi>s</mi> </mrow> </msub> </mrow> </semantics></math>, caused by the influence of parasitic capacitances of components of the LADC with successive approximation on SC with a variable logarithm base (<span class="html-italic">C</span>1 = <span class="html-italic">C</span>2 = 10 nF). LADC bits: 16—square, 12—circle, 10—star; C<sub>p</sub>: 4 pF—red, 2 pF—green, 1 pF—blue.</p>
Full article ">Figure 6
<p>The absolute summative error <math display="inline"><semantics> <mrow> <mo>Δ</mo> <msub> <mi>U</mi> <mrow> <mi>m</mi> <mi>s</mi> </mrow> </msub> </mrow> </semantics></math>, caused by the influence of parasitic capacitances of components of the LADC with successive approximation on SC with a variable logarithm base (<span class="html-italic">C</span>1 = <span class="html-italic">C</span>2 = 1 nF). LADC bits: 16—square, 12—circle, 10—star; C<sub>p</sub>: 4 pF—red, 2 pF—green, 1 pF—blue.</p>
Full article ">
Previous Issue
Next Issue
Back to TopTop