Nothing Special   »   [go: up one dir, main page]
Next Issue
Volume 10, October-2
Previous Issue
Volume 10, September-2
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
5.3
2.6
Journals
Electronics
Volume 10
Issue 19
share announcement

Electronics, Volume 10, Issue 19 (October-1 2021) – 129 articles

Cover Story (view full-size image): In this paper, we introduce TETRAPAC (TElematic data of TRucks for Advanced Predictive Analysis of their Component), a methodology able to analyze data collected from heavy trucks during their use, offering a generalizable approach to estimating vehicle health conditions based on monitored features enriched by innovative key performance indicators. The methodology has been evaluated using two different use cases: (1) identifying vehicles with potential DTCs (diagnostic trouble codes) and (2) the estimation of the battery life of the trucks. In both use cases, TETRAPAC has been proven to bring significant benefits to the company, in terms of cost savings and increasing customer satisfaction. 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
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
12 pages, 3791 KiB  
Article
Single-Sensor EMI Source Localization Using Time Reversal: An Experimental Validation
by Hamidreza Karami, Mohammad Azadifar, Zhaoyang Wang, Marcos Rubinstein and Farhad Rachidi
Electronics 2021, 10(19), 2448; https://doi.org/10.3390/electronics10192448 - 8 Oct 2021
Cited by 6 | Viewed by 2064
Abstract
The localization of electromagnetic interference (EMI) sources is of high importance in electromagnetic compatibility applications. Recently, a novel localization technique based on the time-reversal cavity (TRC) concept was proposed using only one sensor, and its application to localize EMI sources was validated numerically. [...] Read more.
The localization of electromagnetic interference (EMI) sources is of high importance in electromagnetic compatibility applications. Recently, a novel localization technique based on the time-reversal cavity (TRC) concept was proposed using only one sensor, and its application to localize EMI sources was validated numerically. In this paper, we present a validation of the proposed time-reversal process in which the forward step of the time-reversal process is performed experimentally and the backward step is carried out via numerical simulations, a realistic scenario which is applicable to practical source localization problems. To the best of the authors’ knowledge, this is the first implementation of a three-dimensional electromagnetic time-reversal process in which the forward signal is provided experimentally while the backward propagation step is carried out numerically. The considered experimental setup is formed by a partially open cavity and two monopole antennas to emulate the EMI source and the sensor (receiving antenna), respectively. Assuming that the location of the source is the feed point of the monopole antenna, the resulting three-dimensional location error in the experimental validation was only 1.49 cm, which is about one-third the length of the monopole antenna, corresponding to about λmin/2 (diffraction limit). Full article
(This article belongs to the Section Circuit and Signal Processing)
►▼ Show Figures

Figure 1

Figure 1
<p>Open cavity. (<b>a</b>) Geometry of the 3D problem, including an open metallic cavity and two monopole antennas. The dimensions of the cavity are <span class="html-italic">w</span> = 25 cm, <span class="html-italic">h</span> = 13 cm, <span class="html-italic">l</span> = 17.4 cm, and <span class="html-italic">t</span> = 1.5 mm. (<b>b</b>) An expanded view of the two monopoles. The left and right monopoles are considered as the sensor and the source, respectively, in the EMTR method. The two monopole antennas are identical, with a length of 5.3 cm.</p> Full article ">Figure 2
<p>Gaussian pulse with a bandwidth of 0 to 10 GHz that was used.</p> Full article ">Figure 3
<p>Test setup including the open cavity, absorber, connectors, cables, and two monopole antennas. (<b>a</b>) The open cavity placed in front of the absorbers to reduce the effects of reflections from the environment on the experimental results. <b>(b</b>) The cables and connectors used to connect the two monopole antennas to the VNA. (<b>c</b>) A front view of the open cavity, including the two monopole antennas inside it, and (<b>d</b>) an expanded view of the two monopole antennas.</p> Full article ">Figure 4
<p>Open cavity without transmitter antenna used in the back-propagation phase. Geometry of the 3D problem.</p> Full article ">Figure 5
<p>Distribution of the normalized maximum power at the <span class="html-italic">z</span> = 1 cm cut plane in absence of the transmitter monopole antenna in the backward propagation step. The black cross shows the global maximum power, which occurs on the <span class="html-italic">z</span> = 1 cm cut plane. The red circle is the intersection between the <span class="html-italic">z</span> = 1 cm cut plane and the monopole antenna. The estimated location error, defined as the two-dimensional distance between the red circle and the black cross, is 0.18 cm. If the error is measured as the three-dimensional distance from the black cross to the antenna feeding point, it equals 1.02 cm. The blue square with side length of 16 cells (3.4 cm) at the location of the sensor shows the mask filter that was used in this simulation.</p> Full article ">Figure 6
<p>The scattering parameter obtained using the VNA and the full-wave CST MWS software. (<b>a</b>) The scattering parameter in the 0–6 GHz range. (<b>b</b>) The scattering parameter in the 6–10 GHz range.</p> Full article ">Figure 6 Cont.
<p>The scattering parameter obtained using the VNA and the full-wave CST MWS software. (<b>a</b>) The scattering parameter in the 0–6 GHz range. (<b>b</b>) The scattering parameter in the 6–10 GHz range.</p> Full article ">Figure 7
<p>Time-domain sensor signal obtained by way of the frequency domain measurement of the scattering parameter S<sub>21</sub> using the VNA (black solid curve) and the full-wave CST MWS software (red, dashed line) (<b>a</b>) in the 030 ns time interval and (<b>b</b>) in the 1020 ns interval.</p> Full article ">Figure 8
<p>Distribution of the normalized maximum power over the calculation domain in presence of the source monopole antenna (matched media) at cut plane <span class="html-italic">x</span> = 14.78 cm. The solid white line represents the location of the source monopole antenna.</p> Full article ">Figure 9
<p>Distribution of the normalized maximum power over the calculation domain in absence of the source monopole antenna (mismatched media): (<b>a</b>) At cut plane <span class="html-italic">x</span> = 14.78 cm. The solid white line represents the location of the removed source monopole antenna. (<b>b</b>) At cut plane <span class="html-italic">x</span> = 16.78 cm. The white dashed line represents the projection of the removed source monopole antenna on the cut plane. The location of maximum power is shown using a green “+” sign.</p> Full article ">Figure 10
<p>Distribution of the normalized maximum power over the calculation domain in absence of the source monopole antenna (mismatched media): (<b>a</b>) At cut plane <span class="html-italic">y</span> = 6.66 cm. The solid white line represents the location of the removed source monopole antenna. (<b>b</b>) At cut plane <span class="html-italic">y</span> = 8.66 cm. The white dashed line represents the projection of the removed source monopole antenna on the cut plane. The location of maximum power is shown using a green “+” sign.</p> Full article ">
18 pages, 8582 KiB  
Article
Hardware-in-the-Loop Simulation of Self-Driving Electric Vehicles by Dynamic Path Planning and Model Predictive Control
by Yi Chung and Yee-Pien Yang
Electronics 2021, 10(19), 2447; https://doi.org/10.3390/electronics10192447 - 8 Oct 2021
Cited by 6 | Viewed by 3900
Abstract
This paper applies a dynamic path planning and model predictive control (MPC) to simulate self-driving and parking for an electric van on a hardware-in-the-loop (HiL) platform. The hardware platform is a simulator which consists of an electric power steering system, accelerator and brake [...] Read more.
This paper applies a dynamic path planning and model predictive control (MPC) to simulate self-driving and parking for an electric van on a hardware-in-the-loop (HiL) platform. The hardware platform is a simulator which consists of an electric power steering system, accelerator and brake pedals, and an Nvidia drive PX2 with a robot operating system (ROS). The vehicle dynamics model, sensors, controller, and test field map are virtually built with the PreScan simulation platform. Both manual and autonomous driving modes can be simulated, and a graphic user interface allows a test driver to select a target parking space on a display screen. Three scenarios are demonstrated: forward parking, reverse parking, and obstacle avoidance. When the vehicle perceives an obstacle, the map is updated and the route is adaptively planned. The effectiveness of the proposed MPC is verified in experiments and proved to be superior to a traditional proportional–integral–derivative controller with regards to safety, energy-saving, comfort, and agility. Full article
(This article belongs to the Special Issue Unmanned Vehicles and Intelligent Robotic Alike Systems)
►▼ Show Figures

Figure 1

Figure 1
<p>Bicycle model of the vehicle.</p> Full article ">Figure 2
<p>Vehicle dynamics models in the (<b>a</b>) longitudinal and (<b>b</b>) lateral directions.</p> Full article ">Figure 3
<p>Traditional A* searching path (<b>left</b>) vs. hybrid A* searching path (<b>right</b>).</p> Full article ">Figure 4
<p>Forward part in the path planning process on PreScan platform.</p> Full article ">Figure 5
<p>A visualization process of path planning on the parking lot created by RViz.</p> Full article ">Figure 6
<p>Structure of model predictive control.</p> Full article ">Figure 7
<p>Definition of cross tracking and heading errors.</p> Full article ">Figure 8
<p>Architecture of HiL simulation platform and experimental rig.</p> Full article ">Figure 9
<p>HiL experimental setup.</p> Full article ">Figure 10
<p>Self-driving route (red star line) by PreScan and parking lot map on the test field.</p> Full article ">Figure 11
<p>Tracking performance on the prescribed route with different emphasis on safety and comfort. Setting 1 has better tracking performance than Setting 2.</p> Full article ">Figure 12
<p>Vehicle speed performance on the prescribed route with different emphasis on safety and comfort. Setting 1 follows the reference speed better than Setting 2.</p> Full article ">Figure 13
<p>Vehicle lateral acceleration curve on the prescribed route with different emphasis on safety and comfort. Setting 2 provides more comfort driving than Setting 1.</p> Full article ">Figure 14
<p>Vehicle tracking error on the DLC test by MPC and PID controllers.</p> Full article ">Figure 15
<p>Vehicle yaw rate on the DLC test by MPC and PID controllers.</p> Full article ">Figure 16
<p>Accelerator pedal angle (%) history of (<b>a</b>) PID controller, and MPC with (<b>b</b>) Case 1, (<b>c</b>) Case 3, and (<b>d</b>) Case 10 for energy saving simulation. (The area under the curve represents energy consumption).</p> Full article ">Figure 17
<p>Energy efficiency improvement versus the resultant weighing factors <span class="html-italic">(</span><math display="inline"><semantics> <mrow> <msub> <mi>w</mi> <mi>p</mi> </msub> <mo>+</mo> <msub> <mi>w</mi> <mover accent="true"> <mi>p</mi> <mo>˙</mo> </mover> </msub> </mrow> </semantics></math>).</p> Full article ">Figure 18
<p>2D map of path planning and auto-parking scenario.</p> Full article ">Figure 19
<p>Vehicle steering angle curve during the path planning and auto-parking scenario.</p> Full article ">Figure 20
<p>Vehicle yaw rate curve during the path planning and auto-parking scenario.</p> Full article ">Figure 21
<p>Vehicle brake pedal curve during the path planning and auto-parking scenario.</p> Full article ">
7 pages, 4197 KiB  
Article
RF Pogo-Pin Probe Card Design Aimed at Automated Millimeter-Wave Multi-Port Integrated-Circuit Testing
by K. M. Lee, J. H. Oh, M. S. Kim, T. S. Kim and M. Kim
Electronics 2021, 10(19), 2446; https://doi.org/10.3390/electronics10192446 - 8 Oct 2021
Cited by 5 | Viewed by 5168
Abstract
A prototype RF probe card is assembled to test the feasibility of Pogo-pins as robust probe tips for the automized testing of multiple-port millimeter-wave circuits. A custom-made ceramic housing machined from a low-loss dielectric holds an array of 157 Pogo-pins, each with 2.9 [...] Read more.
A prototype RF probe card is assembled to test the feasibility of Pogo-pins as robust probe tips for the automized testing of multiple-port millimeter-wave circuits. A custom-made ceramic housing machined from a low-loss dielectric holds an array of 157 Pogo-pins, each with 2.9 mm-length in fixed positions. The ceramic housing is then mounted onto a probe-card PCB for power-loss measurements on two signal-ground Pogo-pin connections arbitrarily selected from the array. The probing results on a test circuit with a simple thru-line indicate a successful power transfer with a small insertion loss of less than 0.5 dB per single Pogo-pin connection up to 25 GHz. A new probe card design using shorter Pogo-pins is being prepared to extend the operation frequency to beyond 40 GHz. Full article
(This article belongs to the Section Microwave and Wireless Communications)
►▼ Show Figures

Figure 1

Figure 1
<p>Ball pad map of the SoC under test with the chip size of 3.3 × 5.6 mm<sup>2</sup>.</p> Full article ">Figure 2
<p>Sketches of (<b>a</b>) uncompressed Pogo-pin and (<b>b</b>) ceramic housing containing Pogo-pin array.</p> Full article ">Figure 3
<p>(<b>a</b>) Loss-estimation PCB and (<b>b</b>) the measured and simulated losses.</p> Full article ">Figure 4
<p>Sketch of probe-card PCB.</p> Full article ">Figure 5
<p>Photographs of (<b>a</b>) the probe card and (<b>b</b>) the test PCB, and (<b>c</b>) the total insertion and (<b>d</b>) reflection losses.</p> Full article ">Figure 6
<p>(<b>a</b>) TDR waveform converted from S-parameters and (<b>b</b>) measured eye diagram for 20 Gbps OOK modulation.</p> Full article ">Figure 7
<p>(<b>a</b>) Simplified Pogo-pin structure for HFSS analysis, (<b>b</b>) HFSS E-field plot at the resonance, and (<b>c</b>) simulated resonant frequency shifts for varying pin lengths.</p> Full article ">
14 pages, 5971 KiB  
Article
Formulation and Analysis of Single Switch High Gain Hybrid DC to DC Converter for High Power Applications
by Sathiya Ranganathan and Arun Noyal Doss Mohan
Electronics 2021, 10(19), 2445; https://doi.org/10.3390/electronics10192445 - 8 Oct 2021
Cited by 10 | Viewed by 2954
Abstract
The necessity for DC−DC converters has been rapidly increasing due to the emergence of RES-based electrification. However, the converter designed so far exhibits the drawbacks of lower efficiency and non-compactness in size. Hence, to rectify this problem, the new topology of a flyback [...] Read more.
The necessity for DC−DC converters has been rapidly increasing due to the emergence of RES-based electrification. However, the converter designed so far exhibits the drawbacks of lower efficiency and non-compactness in size. Hence, to rectify this problem, the new topology of a flyback converter for PV application is proposed in this work. The proposed converter exhibits reduced ripple in input current and enhances the conversion efficiency. Finally, the efficiency of this proposed converter is verified using MATLAB. The results indicate that this projected topology can be suitable for high voltage DC applications. Full article
(This article belongs to the Special Issue 10th Anniversary of Electronics: Related Researches in Electrical and Autonomous Vehicles)
►▼ Show Figures

Figure 1

Figure 1
<p>(<b>a</b>) Schematic diagram of a proposed module. (<b>b</b>) Diode model of a PV cell.</p> Full article ">Figure 2
<p>(<b>a</b>) Schematic diagram of proposed converter (<b>b</b>) Steady state waveforms (<b>c</b>) Mode 2 operation. (<b>d</b>) Mode 3 operation. (<b>e</b>) Mode 4 operation. (<b>f</b>) Mode 5 operation.</p> Full article ">Figure 2 Cont.
<p>(<b>a</b>) Schematic diagram of proposed converter (<b>b</b>) Steady state waveforms (<b>c</b>) Mode 2 operation. (<b>d</b>) Mode 3 operation. (<b>e</b>) Mode 4 operation. (<b>f</b>) Mode 5 operation.</p> Full article ">Figure 3
<p>Block diagram of FOPID controller.</p> Full article ">Figure 4
<p>Performance of system (<b>a</b>) irradiationand (<b>b</b>) temperature.</p> Full article ">Figure 5
<p>Output PV power and converter voltage with respect to irradiance.</p> Full article ">Figure 6
<p>Output voltage/current of the converter.</p> Full article ">Figure 7
<p>Converter output voltage waveform.</p> Full article ">Figure 8
<p>Hardware layout.</p> Full article ">Figure 9
<p>Experimental setup of proposed converter.</p> Full article ">Figure 10
<p>(<b>a</b>) Input voltage waveform of the proposed converter, (<b>b</b>) output voltage waveform of the proposed converter, (<b>c</b>) switching pulse to the switch S<sub>1.</sub></p> Full article ">Figure 10 Cont.
<p>(<b>a</b>) Input voltage waveform of the proposed converter, (<b>b</b>) output voltage waveform of the proposed converter, (<b>c</b>) switching pulse to the switch S<sub>1.</sub></p> Full article ">Figure 11
<p>Voltage conversion ratio versus duty ratio.</p> Full article ">Figure 12
<p>Efficiency versus input voltage.</p> Full article ">
19 pages, 6321 KiB  
Article
Image-Based Malware Classification Using VGG19 Network and Spatial Convolutional Attention
by Mazhar Javed Awan, Osama Ahmed Masood, Mazin Abed Mohammed, Awais Yasin, Azlan Mohd Zain, Robertas Damaševičius and Karrar Hameed Abdulkareem
Electronics 2021, 10(19), 2444; https://doi.org/10.3390/electronics10192444 - 8 Oct 2021
Cited by 100 | Viewed by 8008
Abstract
In recent years the amount of malware spreading through the internet and infecting computers and other communication devices has tremendously increased. To date, countless techniques and methodologies have been proposed to detect and neutralize these malicious agents. However, as new and automated malware [...] Read more.
In recent years the amount of malware spreading through the internet and infecting computers and other communication devices has tremendously increased. To date, countless techniques and methodologies have been proposed to detect and neutralize these malicious agents. However, as new and automated malware generation techniques emerge, a lot of malware continues to be produced, which can bypass some state-of-the-art malware detection methods. Therefore, there is a need for the classification and detection of these adversarial agents that can compromise the security of people, organizations, and countless other forms of digital assets. In this paper, we propose a spatial attention and convolutional neural network (SACNN) based on deep learning framework for image-based classification of 25 well-known malware families with and without class balancing. Performance was evaluated on the Malimg benchmark dataset using precision, recall, specificity, precision, and F1 score on which our proposed model with class balancing reached 97.42%, 97.95%, 97.33%, 97.11%, and 97.32%. We also conducted experiments on SACNN with class balancing on benign class, also produced above 97%. The results indicate that our proposed model can be used for image-based malware detection with high performance, despite being simpler as compared to other available solutions. Full article
(This article belongs to the Special Issue Security and Privacy for IoT and Multimedia Services)
►▼ Show Figures

Figure 1

Figure 1
<p>Process of creating a grayscale image from binary of a malware program.</p> Full article ">Figure 2
<p>Grayscale images of malware: (<b>a</b>) unrelated classes of malware, and (<b>b</b>) related classes from the same malware family [<a href="#B33-electronics-10-02444" class="html-bibr">33</a>].</p> Full article ">Figure 3
<p>Sections of the image generated from binary that visually seem to have some correlation.</p> Full article ">Figure 4
<p>Visualization of the frequency of different variants of malware in the Malimg dataset.</p> Full article ">Figure 5
<p>Down sampling inside a convolutional neural network with pooling layer.</p> Full article ">Figure 6
<p>Long short-term memory generates attention from forwarding and backward hidden states.</p> Full article ">Figure 7
<p>Process of spatial attention is added to enhance features before they are fed into CNN.</p> Full article ">Figure 8
<p>Proposed model architecture along with the attention generation mechanism.</p> Full article ">Figure 9
<p>Training and testing accuracies for spatial attention and convolutional neural network with VGG19 model: (<b>a</b>) with class balancing on 25 Malware classes; (<b>b</b>) with class balancing on benign class.</p> Full article ">Figure 10
<p>Training and testing loss for SACNN with VGG19 model: (<b>a</b>) with class balancing on 25 Malware classes; (<b>b</b>) with class balancing on benign class.</p> Full article ">Figure 11
<p>Normalized confusion matrix for classification of 25 malware classes.</p> Full article ">Figure 12
<p>Normalized confusion matrix for classification with class balancing and benign class.</p> Full article ">Figure 13
<p>ROC (receiver operating characteristic) curves with area under curve values for classification into 25 malware families.</p> Full article ">Figure 14
<p>ROC curves with area under curve values for classification with class balancing and benign class.</p> Full article ">
16 pages, 1396 KiB  
Article
DWSA: An Intelligent Document Structural Analysis Model for Information Extraction and Data Mining
by Tan Yue, Yong Li and Zonghai Hu
Electronics 2021, 10(19), 2443; https://doi.org/10.3390/electronics10192443 - 8 Oct 2021
Cited by 10 | Viewed by 2640
Abstract
The structure of a document contains rich information such as logical relations in context, hierarchy, affiliation, dependence, and applicability. It will greatly affect the accuracy of document information processing, particularly of legal documents and business contracts. Therefore, intelligent document structural analysis is important [...] Read more.
The structure of a document contains rich information such as logical relations in context, hierarchy, affiliation, dependence, and applicability. It will greatly affect the accuracy of document information processing, particularly of legal documents and business contracts. Therefore, intelligent document structural analysis is important to information extraction and data mining. However, unlike the well-studied field of text semantic analysis, current work in document structural analysis is still scarce. In this paper, we propose an intelligent document structural analysis framework through data pre-processing, feature engineering, and structural classification with a dynamic sample weighting algorithm. As a typical application, we collect more than 11,000 insurance document content samples and carry out the machine learning experiments to check the efficiency of our framework. Meanwhile, to address the sample imbalance problem in the hierarchy classification task, a dynamic sample weighting algorithm is incorporated into our Dynamic Weighting Structural Analysis (DWSA) framework, in which the weights of different category tags according to the structural levels are iterated dynamically in training. Our results show that the DWSA has significantly improved the comprehensive accuracy and the classification F1-score of each category. The comprehensive accuracy is as high as 94.68% (3.36% absolute improvement) and the Macro F1-score is 88.29% (5.1% absolute improvement). Full article
(This article belongs to the Special Issue Advances in Swarm Intelligence, Data Science and Their Applications)
►▼ Show Figures

Figure 1

Figure 1
<p>The framework of Dynamic Weighting Structural analysis (DWSA). The DWSA framework mainly includes three parts: data pre-processing, feature engineering, and structural classification.</p> Full article ">Figure 2
<p>The file format in the data pre-processing step. We modify the Pdfplumber algorithm framework for PDF document information conversion, which can get more features in document content. The features mainly include size, font family, top, company, etc. The features named top, left, width, and height mean distance from page top, distance from page left, total width, and the total height of the sentence, respectively.</p> Full article ">Figure 3
<p>The context feature fusion algorithm. Through feature engineering, more context feature information is extracted. The context feature information is mapped into vectors for fusion.</p> Full article ">Figure 4
<p>Structured documents output. Document information is extracted for structured comparison. (The language of insurance documents is Chinese. We translate the structured output into English for better understanding. See the appendix file <a href="#electronics-10-02443-f0A3" class="html-fig">Figure A3</a> for the Chinese output).</p> Full article ">Figure 5
<p>The result of the algorithm comparison experiment on the insurance data set. Accuracy and Macro F1-score are used to evaluate different algorithms. The proposed DWSA has the accuracy of 94.68% (3.36% absolute improvement over Adaboost) and the F1-score of 88.29% (5.1% absolute improvement over Adaboost).</p> Full article ">Figure 6
<p>The result of the algorithm comparison experiment on the public data set Baike2018qa and Iris.</p> Full article ">Figure 7
<p>The comparison of the DWSA and the Adaboost algorithm (Insurance data set). The improved model increases the F1-score of the level-1 heading(1) from 71.7% to 82.42%. For the level-3 heading(3), the F1-score is increased by 6.27%. The accuracy of the level-2 heading(2) is improved by 5.49% and reaches 94.99%. The weights of useless content(-1) and body content(0) are reduced relatively, but the accuracy is still increased by 0.32% and 4%, and the final accuracy achieves 92.8% and 96.5%.</p> Full article ">Figure 8
<p>The Macro-average ROC curve of the Adaboost algorithm and The DWSA algorithm. (Insurance data set) (<b>a</b>) Adaboost; (<b>b</b>) DWSA.</p> Full article ">Figure 9
<p>The feature importance of data. F0-F16 respectively represent the features, including size, count, content, font-family, top, left, width, page, height, company, pre-left, pre-size, part, pre-font, total-part, pre-top, and pre-behind. The feature importance of ’page’, ’part’, and ’total-part’ is too low to show in the figure; ’size’ means the font size, ’count’ means the word and punctuation count, ’content’ means the content in documents, ’font-family’ means the font, ’company’ means the insurance company, ’width’ means the total width of the sentence, ’page’ means the page number, ’height’ means the total height of the sentence, ’part’ means the order of the sentence in a paragraph, ’total-part’ means the number of sentences in a paragraph, ’top’ and ’left’ mean the coordinate position of each sentence in the page, ’pre-left’, ’pre-size’, ’pre-font’, and ’pre-top’ are the corresponding features of the previous sentence. (‘pre-left’, ’pre-top’ and so on are just the names we call the features). ’pre-behind’ means the distance from page bottom of the previous sentence. The previous sentence features are incorporated into context feature fusion.</p> Full article ">Figure A1
<p>A typical original Chinese insurance document in PDF format.</p> Full article ">Figure A2
<p>The file format in the data pre-processing step (in Chinese). We modify the Pdfplumber algorithm framework for PDF document information conversion, which can get more features in documents content. The features mainly include size, font family, top, and company etc. The features named top, left, width, and height mean the context relative location information.</p> Full article ">Figure A3
<p>Structured document output. Document information is extracted for structural comparison. (Chinese output).</p> Full article ">
19 pages, 2207 KiB  
Article
CA-CRE: Classification Algorithm-Based Controller Area Network Payload Format Reverse-Engineering Method
by Cheongmin Ji, Taehyoung Ko and Manpyo Hong
Electronics 2021, 10(19), 2442; https://doi.org/10.3390/electronics10192442 - 8 Oct 2021
Cited by 1 | Viewed by 2109
Abstract
In vehicles, dozens of electronic control units are connected to one or more controller area network (CAN) buses to exchange information and send commands related to the physical system of the vehicles. Furthermore, modern vehicles are connected to the Internet via telematics control [...] Read more.
In vehicles, dozens of electronic control units are connected to one or more controller area network (CAN) buses to exchange information and send commands related to the physical system of the vehicles. Furthermore, modern vehicles are connected to the Internet via telematics control units (TCUs). This leads to an attack vector in which attackers can control vehicles remotely once they gain access to in-vehicle networks (IVNs) and can discover the formats of important messages. Although the format information is kept secret by car manufacturers, CAN is vulnerable, since payloads are transmitted in plain text. In contrast, the secrecy of message formats inhibits IVN security research by third-party researchers. It also hinders effective security tests for in-vehicle networks as performed by evaluation authorities. To mitigate this problem, a method of reverse-engineering CAN payload formats is proposed. The method utilizes classification algorithms to predict signal boundaries from CAN payloads. Several features were uniquely chosen and devised to quantify the type-specific characteristics of signals. The method is evaluated on real-world and synthetic CAN traces, and the results show that our method can predict at least 10% more signal boundaries than the existing methods. Full article
(This article belongs to the Special Issue Data-Driven Security)
►▼ Show Figures

Figure 1

Figure 1
<p>Overall process of CA-CRE.</p> Full article ">Figure 2
<p>Distribution of FR and entropy/size values by signal type.</p> Full article ">Figure 3
<p>Distribution of sizes of all signals defined in the DBCs of OpenDBC [<a href="#B13-electronics-10-02442" class="html-bibr">13</a>].</p> Full article ">Figure 4
<p>Examples of synthetic SENSOR Type-1 (<b>a</b>) and Type-2 (<b>b</b>).</p> Full article ">Figure 5
<p>Accuracy of a classifier by the size of training data.</p> Full article ">
16 pages, 6868 KiB  
Article
A Low-Cost Hardware-Friendly Spiking Neural Network Based on Binary MRAM Synapses, Accelerated Using In-Memory Computing
by Yihao Wang, Danqing Wu, Yu Wang, Xianwu Hu, Zizhao Ma, Jiayun Feng and Yufeng Xie
Electronics 2021, 10(19), 2441; https://doi.org/10.3390/electronics10192441 - 8 Oct 2021
Cited by 5 | Viewed by 2916
Abstract
In recent years, the scaling down that Moore’s Law relies on has been gradually slowing down, and the traditional von Neumann architecture has been limiting the improvement of computing power. Thus, neuromorphic in-memory computing hardware has been proposed and is becoming a promising [...] Read more.
In recent years, the scaling down that Moore’s Law relies on has been gradually slowing down, and the traditional von Neumann architecture has been limiting the improvement of computing power. Thus, neuromorphic in-memory computing hardware has been proposed and is becoming a promising alternative. However, there is still a long way to make it possible, and one of the problems is to provide an efficient, reliable, and achievable neural network for hardware implementation. In this paper, we proposed a two-layer fully connected spiking neural network based on binary MRAM (Magneto-resistive Random Access Memory) synapses with low hardware cost. First, the network used an array of multiple binary MRAM cells to store multi-bit fixed-point weight values. This helps to simplify the read/write circuit. Second, we used different kinds of spike encoders that ensure the sparsity of input spikes, to reduce the complexity of peripheral circuits, such as sense amplifiers. Third, we designed a single-step learning rule, which fit well with the fixed-point binary weights. Fourth, we replaced the traditional exponential Leak-Integrate-Fire (LIF) neuron model to avoid the massive cost of exponential circuits. The simulation results showed that, compared to other similar works, our SNN with 1184 neurons and 313,600 synapses achieved an accuracy of up to 90.6% in the MNIST recognition task with full-resolution (28 × 28) and full-bit-depth (8-bit) images. In the case of low-resolution (16 × 16) and black-white (1-bit) images, the smaller version of our network with 384 neurons and 32,768 synapses still maintained an accuracy of about 77%, extending its application to ultra-low-cost situations. Both versions need less than 30,000 samples to reach convergence, which is a >50% reduction compared to other similar networks. As for robustness, it is immune to the fluctuation of MRAM cell resistance. Full article
(This article belongs to the Special Issue Neuromorphic Sensing and Computing Systems)
►▼ Show Figures

Figure 1

Figure 1
<p>Block diagram.</p> Full article ">Figure 2
<p>The output spike trains from the Poisson encoders in the first half (0–127) in a time window.</p> Full article ">Figure 3
<p>Probability distribution of the number of input neurons that fire at the same time under different input resolutions using 1-bit fixed frequency coding. (<b>a</b>) Under full resolution (input neurons = 28 × 28), (<b>b</b>) under low resolution (input neurons = 16 × 16).</p> Full article ">Figure 4
<p>SNN workflow.</p> Full article ">Figure 5
<p>(<b>a</b>) Accuracy vs. number of input layer neurons; (<b>b</b>) accuracy vs. number of feature layer neurons.</p> Full article ">Figure 6
<p>Weight map after training; each image is for a feature layer neuron.</p> Full article ">Figure 7
<p>Changes in weight image of a single feature layer during training (samples = 0, 1000, 2000, 3000, 4000, 5000, 6000).</p> Full article ">Figure 8
<p>Relationship between accuracy and the number of training samples when the number of feature neurons is 100 and 400.</p> Full article ">Figure 9
<p>Impact of the number of pre-classified samples on accuracy when the number of feature neurons is 100 or 400.</p> Full article ">Figure 10
<p>Impact of learning rules on accuracy.</p> Full article ">Figure 11
<p>Impact of resistance fluctuation on accuracy.</p> Full article ">
18 pages, 2816 KiB  
Article
k-NDDP: An Efficient Anonymization Model for Social Network Data Release
by Shafaq Shakeel, Adeel Anjum, Alia Asheralieva and Masoom Alam
Electronics 2021, 10(19), 2440; https://doi.org/10.3390/electronics10192440 - 8 Oct 2021
Cited by 8 | Viewed by 2393
Abstract
With the evolution of Internet technology, social networking sites have gained a lot of popularity. People make new friends, share their interests, experiences in life, etc. With these activities on social sites, people generate a vast amount of data that is analyzed by [...] Read more.
With the evolution of Internet technology, social networking sites have gained a lot of popularity. People make new friends, share their interests, experiences in life, etc. With these activities on social sites, people generate a vast amount of data that is analyzed by third parties for various purposes. As such, publishing social data without protecting an individual’s private or confidential information can be dangerous. To provide privacy protection, this paper proposes a new degree anonymization approach k-NDDP, which extends the concept of k-anonymity and differential privacy based on Node DP for vertex degrees. In particular, this paper considers identity disclosures on social data. If the adversary efficiently obtains background knowledge about the victim’s degree and neighbor connections, it can re-identify its victim from the social data even if the user’s identity is removed. The contribution of this paper is twofold. First, a simple and, at the same time, effective method k–NDDP is proposed. The method is the extension of k-NMF, i.e., the state-of-the-art method to protect against mutual friend attack, to defend against identity disclosures by adding noise to the social data. Second, the achieved privacy using the concept of differential privacy is evaluated. An extensive empirical study shows that for different values of k, the divergence produced by k-NDDP for CC, BW and APL is not more than 0.8%, also added dummy links are 60% less, as compared to k-NMF approach, thereby it validates that the proposed k-NDDP approach provides strong privacy while maintaining the usefulness of data. Full article
(This article belongs to the Special Issue Big Data Privacy-Preservation)
►▼ Show Figures

Figure 1

Figure 1
<p>Social networks data publishing scenario.</p> Full article ">Figure 2
<p>Social graphs: (<b>a</b>) Original graph, (<b>b</b>) naïve anonymized graph.</p> Full article ">Figure 3
<p>Neighborhood graphs in a social network: (<b>a</b>) Graph for Bob, (<b>b</b>) graph for Alice.</p> Full article ">Figure 4
<p>Anonymization techniques for social networks data publishing.</p> Full article ">Figure 5
<p><span class="html-italic">k</span>-NMF anonymized graph.</p> Full article ">Figure 6
<p>Overview of proposed <span class="html-italic">k</span>-NDDP approach.</p> Full article ">Figure 7
<p>Comparison of <span class="html-italic">k</span>-NDDP with <span class="html-italic">k</span>-NMF(FEATHER-DEEZER-SOCIAL). (<b>a</b>) CC (<b>b</b>) APL (<b>c</b>) BW.</p> Full article ">Figure 8
<p>Comparison of <span class="html-italic">k</span>-NDDP with <span class="html-italic">k</span>-NMF(SOCFB-USFCA72). (<b>a</b>) CC; (<b>b</b>) APL; (<b>c</b>) BW.</p> Full article ">Figure 9
<p>Comparison of <span class="html-italic">k</span>-NDDP with <span class="html-italic">k</span>-NMF(SOCFB-USFCA72). (<b>a</b>) Time; (<b>b</b>) edge change.</p> Full article ">
10 pages, 1870 KiB  
Communication
Radiation Beam Pattern Control of UHF RFID Tag Antenna Design for Automotive License Plates
by Youchung Chung and Teklebrhan H. Berhe
Electronics 2021, 10(19), 2439; https://doi.org/10.3390/electronics10192439 - 8 Oct 2021
Cited by 1 | Viewed by 2489
Abstract
This paper presents a design of a radio frequency identification (RFID) tag antenna in the ultra-high-frequency (UHF) range, which is applicable to a vehicular license plate attached to a vehicle bumper. The main goals are to first improve the identification ratio by controlling [...] Read more.
This paper presents a design of a radio frequency identification (RFID) tag antenna in the ultra-high-frequency (UHF) range, which is applicable to a vehicular license plate attached to a vehicle bumper. The main goals are to first improve the identification ratio by controlling the radiation beam pattern and, second, to control the beam direction. Since every vehicle has a license plate, the available plate structure is used to design the antenna. The shape of the tag is rectangular and has a dimension of 525 mm × 116 mm, which is smaller than the typical size of standard plates, 540 mm × 120 mm, used in Europe and Korea. The fabricated tag antenna, the license plate, and the vehicular bumper are fixed by volt and nut. For vehicle tracking and identification, RFID readers are deployed on the road side. For efficient identification, a long distance passive UHF RFID license plate with a patch antenna is proposed to provide not only line-of-sight identification but also left and right beams. Unlike the general UHF tag antennas, in this paper, the patch antenna is designed to attach to the metal part of the car, the license plate holder. The beam patterns of the RFID tag antenna can be controlled by the patch antenna parameter values. The simulation result demonstrates that the proposed UHF RFID tag antenna has a beam radiation pattern as required at 920 MHz. In addition, the estimated read range of the proposed plate meets the requirement of RFID systems. Full article
(This article belongs to the Collection Smart Sensing RFID Tags)
►▼ Show Figures

Figure 1

Figure 1
<p>Size of a typical license plate holder. [<a href="#B26-electronics-10-02439" class="html-bibr">26</a>].</p> Full article ">Figure 2
<p>Parameters and design structure of the RFID tag antenna without bumper.</p> Full article ">Figure 3
<p>Front and back side of license plate RFID tag antenna with bumper.</p> Full article ">Figure 4
<p>Radiation pattern of tag antenna.</p> Full article ">Figure 4 Cont.
<p>Radiation pattern of tag antenna.</p> Full article ">Figure 5
<p>The three-dimensional radiation pattern of the tag antenna at 920 MHz.</p> Full article ">Figure 6
<p>(<b>a</b>) Simulated return loss of tag antenna at 920 MHz. (<b>b</b>) Current density of the tag antenna at 920 MHz.</p> Full article ">Figure 7
<p>Optimization using parameter sweep S11 (dB) vs. frequency (GHz).</p> Full article ">
12 pages, 723 KiB  
Article
Reliability of Recurrence Quantification Analysis Measures for Sit-to-Stand and Stand-to-Sit Activities in Healthy Older Adults Using Wearable Sensors
by Amnah Nasim, David C. Nchekwube and Yoon Sang Kim
Electronics 2021, 10(19), 2438; https://doi.org/10.3390/electronics10192438 - 8 Oct 2021
Cited by 2 | Viewed by 2214
Abstract
Standing up and sitting down are prerequisite motions in most activities of daily living scenarios. The ability to sit down in and stand up from a chair or a bed depreciates and becomes a complex task with increasing age. Hence, research on the [...] Read more.
Standing up and sitting down are prerequisite motions in most activities of daily living scenarios. The ability to sit down in and stand up from a chair or a bed depreciates and becomes a complex task with increasing age. Hence, research on the analysis and recognition of these two activities can help in the design of algorithms for assistive devices. In this work, we propose a reliability analysis for testing the internal consistency of nonlinear recurrence features for sit-to-stand (Si2St) and stand-to-sit (St2Si) activities for motion acceleration data collected by a wearable sensing device for 14 healthy older subjects in the age range of 78 ± 4.9 years. Four recurrence features—%recurrence rate, %determinism, entropy, and average diagonal length—were calculated by using recurrence plots for both activities. A detailed relative and absolute reliability statistical analysis based on Cronbach’s correlation coefficient (α) and standard error of measurement was performed for all recurrence measures. Correlation values as high as α = 0.68 (%determinism) and α = 0.72 (entropy) in the case of Si2St and α = 0.64 (%determinism) and α = 0.69 (entropy) in the case of St2Si—with low standard error in the measurements—show the reliability of %determinism and entropy for repeated acceleration measurements for the characterization of both the St2Si and Si2St activities in the case of healthy older adults. Full article
(This article belongs to the Special Issue Wearable Electronics for Assessing Human Motor (dis)Abilities)
►▼ Show Figures

Figure 1

Figure 1
<p>Wearable sensor settings during data acquisition: (<b>a</b>) positioning of the W<math display="inline"><semantics> <msup> <mrow/> <mn>2</mn> </msup> </semantics></math>ISP sensor on the subject’s body, (<b>b</b>) the W<math display="inline"><semantics> <msup> <mrow/> <mn>2</mn> </msup> </semantics></math>ISP sensor attached to the subject’s clothing with isolating silver fabric [<a href="#B47-electronics-10-02438" class="html-bibr">47</a>].</p> Full article ">Figure 2
<p>Experimental setting for data collection using W<math display="inline"><semantics> <msup> <mrow/> <mn>2</mn> </msup> </semantics></math>ISP during activities: (<b>a</b>) sit-to-stand (Si2St); (i) sitting on the bed/chair, (ii) transition, (iii) standing; (<b>b</b>) stand-to-sit (St2Si); (i) standing, (ii) transition, (iii) sitting on the bed/chair.</p> Full article ">Figure 3
<p>Raw, interpolated, and filtered vertical acceleration signal <math display="inline"><semantics> <msub> <mi mathvariant="bold">a</mi> <mi>v</mi> </msub> </semantics></math> acquired from W<math display="inline"><semantics> <msup> <mrow/> <mn>2</mn> </msup> </semantics></math>ISP while performing (<b>a</b>) Si2St and (<b>b</b>) St2Si.</p> Full article ">Figure 4
<p>Recurrence matrices for the activities: (<b>a</b>) Si2St (m = 5, <math display="inline"><semantics> <mi>τ</mi> </semantics></math> = 15, <math display="inline"><semantics> <mi>ρ</mi> </semantics></math> = 0.01); (<b>b</b>) St2Si (m = 5, <math display="inline"><semantics> <mi>τ</mi> </semantics></math> = 18, <math display="inline"><semantics> <mi>ρ</mi> </semantics></math> = 0.04) before the step function; (<b>c</b>) Si2St; (<b>d</b>) St2Si after the step function.</p> Full article ">
14 pages, 4153 KiB  
Article
Transient Stability Enhancement of a Grid-Connected Large-Scale PV System Using Fuzzy Logic Controller
by Md. Rifat Hazari, Effat Jahan, Mohammad Abdul Mannan and Narottam Das
Electronics 2021, 10(19), 2437; https://doi.org/10.3390/electronics10192437 - 8 Oct 2021
Cited by 16 | Viewed by 3595
Abstract
This paper presents a new intelligent control strategy to augment the low-voltage ride-through (LVRT) potential of photovoltaic (PV) plants, and the transient stability of a complete grid system. Modern grid codes demand that a PV plant should be connected to the main power [...] Read more.
This paper presents a new intelligent control strategy to augment the low-voltage ride-through (LVRT) potential of photovoltaic (PV) plants, and the transient stability of a complete grid system. Modern grid codes demand that a PV plant should be connected to the main power system during network disturbance, providing voltage support. Therefore, in this paper, a novel fuzzy logic controller (FLC) using the controlled cascaded strategy is proposed for the grid side converter (GSC) of a PV plant to guarantee voltage recovery. The proposed FLC offers variable gains based upon the system requirements, which can inject a useful amount of reactive power after a severe network disturbance. Therefore, the terminal voltage dip will be low, restoring its pre-fault value and resuming its operation quickly. To make it realistic, the PV system is linked to the well-known IEEE nine bus system. Comparative analysis is shown—using power system computer-aided design/electromagnetic transients including DC (PSCAD/EMTDC) software—between the conventional proportional–integral (PI) controller-based cascaded strategy and the proposed control strategy to authenticate the usefulness of the proposed strategy. The comparative simulation results indicate that the transient stability and the LVRT capability of a grid-tied PV system can be augmented against severe fault using the proposed FLC-based cascaded GSC controller. Full article
(This article belongs to the Special Issue Power Quality Conditioning and Stability Enhancement of More-Electronics Power Systems)
►▼ Show Figures

Figure 1

Figure 1
<p>Control mechanism of a grid-tied PV power station/plant.</p> Full article ">Figure 2
<p>Equivalent circuit diagram of a single diode PV module.</p> Full article ">Figure 3
<p>Characteristic curves of a 50 MW PV station.</p> Full article ">Figure 4
<p>Boost converter controller of a PV plant.</p> Full article ">Figure 5
<p>Proposed FLC-based cascaded GSC controller.</p> Full article ">Figure 6
<p>Structure of the proposed FLC.</p> Full article ">Figure 7
<p>Membership functions of an FLC.</p> Full article ">Figure 8
<p>System model of a modified IEEE nine bus power plant model.</p> Full article ">Figure 9
<p>Terminal voltage profile of the conventional and proposed PV plants at bus 10.</p> Full article ">Figure 10
<p>Reactive power profile of the conventional and proposed PV plants at bus 10.</p> Full article ">Figure 11
<p>Active power profile of the conventional and proposed PV plants.</p> Full article ">Figure 12
<p>Rotor speed profile of the SGs.</p> Full article ">Figure 13
<p>Active power profile of the SGs.</p> Full article ">Figure 14
<p>Power angle profile of the SGs.</p> Full article ">Figure 15
<p>Frequency profile of the power systems.</p> Full article ">Figure 16
<p>Profile of the transient stability index for SGs.</p> Full article ">
2 pages, 133 KiB  
Editorial
High-Power Vacuum Electronic Devices from Microwave to THz Band: Way Forward
by Glyavin Mikhail
Electronics 2021, 10(19), 2436; https://doi.org/10.3390/electronics10192436 - 8 Oct 2021
Cited by 6 | Viewed by 1909
Abstract
It is generally accepted that the 20th century was the age of electronics [...] Full article
(This article belongs to the Special Issue Design, Technologies and Applications of High Power Vacuum Electronic Devices from Microwave to THz Band)
15 pages, 3893 KiB  
Article
Social Distance Monitoring Approach Using Wearable Smart Tags
by Tareq Alhmiedat and Majed Aborokbah
Electronics 2021, 10(19), 2435; https://doi.org/10.3390/electronics10192435 - 8 Oct 2021
Cited by 23 | Viewed by 12257
Abstract
Coronavirus has affected millions of people worldwide, with the rate of infected people still increasing. The virus is transmitted between people through direct, indirect, or close contact with infected people. To help prevent the social transmission of COVID-19, this paper presents a new [...] Read more.
Coronavirus has affected millions of people worldwide, with the rate of infected people still increasing. The virus is transmitted between people through direct, indirect, or close contact with infected people. To help prevent the social transmission of COVID-19, this paper presents a new smart social distance system that allows individuals to keep social distances between others in indoor and outdoor environments, avoiding exposure to COVID-19 and slowing its spread locally and across the country. The proposed smart monitoring system consists of a new smart wearable prototype of a compact and low-cost electronic device, based on human detection and proximity distance functions, to estimate the social distance between people and issue a notification when the social distance is less than a predefined threshold value. The developed social system has been validated through several experiments, and achieved a high acceptance rate (96.1%) and low localization error (<6 m). Full article
(This article belongs to the Special Issue Information and Communications Technologies (ICT) to Deal with COVID-19)
►▼ Show Figures

Figure 1

Figure 1
<p>Categorization of social distance monitoring systems.</p> Full article ">Figure 2
<p>The main concept of the proposed social distance solution.</p> Full article ">Figure 3
<p>Main modules for the proposed social distance monitoring system.</p> Full article ">Figure 4
<p>The main concept of the human detection system.</p> Full article ">Figure 5
<p>The flowchart for the social distance monitoring system.</p> Full article ">Figure 6
<p>SD-Tag architecture.</p> Full article ">Figure 7
<p>The developed SD-Tag.</p> Full article ">Figure 8
<p>Tabuk Park Mall side-view.</p> Full article ">Figure 9
<p>Acceptability percentage for 33 users.</p> Full article ">Figure 10
<p>Ease of use percentage for 33 users.</p> Full article ">Figure 11
<p>Comfortability percentage for 33 users.</p> Full article ">Figure 12
<p>Measuring the estimated distance of the SD-Tag to the heading person(s) for four different users.</p> Full article ">Figure 13
<p>The localization error (in meters) for 33 users.</p> Full article ">Figure 14
<p>The remaining energy (in volts) for each SD-Tag.</p> Full article ">
12 pages, 6244 KiB  
Article
An Effective Multi-Task Two-Stage Network with the Cross-Scale Training Strategy for Multi-Scale Image Super Resolution
by Jucheng Yang, Feng Wei, Yaxin Bai, Meiran Zuo, Xiao Sun and Yarui Chen
Electronics 2021, 10(19), 2434; https://doi.org/10.3390/electronics10192434 - 7 Oct 2021
Cited by 3 | Viewed by 2194
Abstract
Convolutional neural networks and the per-pixel loss function have shown their potential to be the best combination for super-resolving severely degraded images. However, there are still challenges, such as the massive number of parameters requiring prohibitive memory and vast computing and storage resources [...] Read more.
Convolutional neural networks and the per-pixel loss function have shown their potential to be the best combination for super-resolving severely degraded images. However, there are still challenges, such as the massive number of parameters requiring prohibitive memory and vast computing and storage resources as well as time-consuming training and testing. What is more, the per-pixel loss measured by L2 and the Peak Signal-to-Noise Ratio do not correlate well with human perception of image quality, since L2 simply does not capture the intricate characteristics of human visual systems. To address these issues, we propose an effective two-stage hourglass network with multi-task co-optimization, which enables the entire network to focus on training and testing time and inherent image patterns such as local luminance, contrast, structure and data distribution. Moreover, to avoid overwhelming memory overheads, our model is capable of performing real-time single image multi-scale super-resolution, so it is memory-friendly, meaning that memory space is utilized efficiently. In addition, in order to best use the underlying structure and perception of image quality and the intermediate estimates during the inference process, we introduce a cross-scale training strategy with 2×, 3× and 4× image super-resolution. This effective multi-task two-stage network with the cross-scale strategy for multi-scale image super-resolution is named EMTCM. Quantitative and qualitative experiment results show that the proposed EMTCM network outperforms state-of-the-art methods in recovering high-quality images. Full article
(This article belongs to the Special Issue New Techniques for Image and Video Coding)
►▼ Show Figures

Figure 1

Figure 1
<p>Overall framework of proposed two-stage EMTCM methods. The architecture of EMTCM is composed of two networks, Memory-Friendly Encoder and Recovery Decoder.</p> Full article ">Figure 2
<p>Qualitative and quantitative comparisons between our EMTCM and state-of-the-art SR models with the scale factor 3. Best viewed zoomed in.</p> Full article ">Figure 3
<p>Qualitative comparison between our EMTCM and state-of-the-art SR models with the scale factor 3. (<b>a</b>) HR (<b>b</b>) Bicubic (<b>c</b>) FSRCNN (<b>d</b>) LapSRN (<b>e</b>) VDSR (<b>f</b>) RDN (<b>g</b>) SRFRN (<b>h</b>) Ours. Best viewed zoomed in.</p> Full article ">Figure 4
<p>Qualitative comparison between our EMTCM and state-of-the-art SR models with the scale factor of 3. (<b>a</b>) HR (<b>b</b>) Bicubic (<b>c</b>) FSRCNN (<b>d</b>) LapSRN (<b>e</b>) VDSR (<b>f</b>) RDN (<b>g</b>) SRFRN (<b>h</b>) Ours. Best viewed zoomed in.</p> Full article ">Figure 5
<p>Qualitative comparison between our EMTCM and state-of-the-art SR models with the scale factor 3. (<b>a</b>) HR (<b>b</b>) Bicubic (<b>c</b>) SRCNN (<b>d</b>) IRCNN (<b>e</b>) SRMD (<b>f</b>) RDN (<b>g</b>) SRFRN (<b>h</b>) Ours. Best viewed zoomed in.</p> Full article ">
17 pages, 4224 KiB  
Article
QiBAM: Approximate Sub-String Index Search on Quantum Accelerators Applied to DNA Read Alignment
by Aritra Sarkar, Zaid Al-Ars, Carmen G. Almudever and Koen L. M. Bertels
Electronics 2021, 10(19), 2433; https://doi.org/10.3390/electronics10192433 - 7 Oct 2021
Cited by 10 | Viewed by 3493
Abstract
With small-scale quantum processors transitioning from experimental physics labs to industrial products, these processors in a few years are expected to scale up and be more robust for efficiently computing important algorithms in various fields. In this paper, we propose a quantum algorithm [...] Read more.
With small-scale quantum processors transitioning from experimental physics labs to industrial products, these processors in a few years are expected to scale up and be more robust for efficiently computing important algorithms in various fields. In this paper, we propose a quantum algorithm to address the challenging field of data processing for genome sequence reconstruction. This research describes an architecture-aware implementation of a quantum algorithm for sub-sequence alignment. A new algorithm named QiBAM (quantum indexed bidirectional associative memory) is proposed, which uses approximate pattern-matching based on Hamming distances. QiBAM extends the Grover’s search algorithm in two ways, allowing: (1) approximate matches needed for read errors in genomics, and (2) a distributed search for multiple solutions over the quantum encoding of DNA sequences. This approach gives a quadratic speedup over the classical algorithm. A full implementation of the algorithm is provided and verified using the OpenQL compiler and QX Simulator framework. Our implementation represents a first exploration towards a full-stack quantum accelerated genome sequencing pipeline design. Full article
(This article belongs to the Special Issue Quantum Computing System Design and Architecture)
►▼ Show Figures

Figure 1

Figure 1
<p>Simple example of associative search.</p> Full article ">Figure 2
<p>DNA sub-sequence alignment problem.</p> Full article ">Figure 3
<p>Grover’s search steps.</p> Full article ">Figure 4
<p>Quantum associative search algorithm with distributed query.</p> Full article ">Figure 5
<p>Quantum phone directory algorithm.</p> Full article ">Figure 6
<p>Quantum circuit blocks of the proposed QiBAM algorithm.</p> Full article ">Figure 7
<p>Quantum architectural stack.</p> Full article ">Figure 8
<p>Quantum Algorithm (block with solid outline) and interfacing software architecture.</p> Full article ">Figure 9
<p>Quantum database for search pattern <span class="html-italic">CA</span> and reference string <span class="html-italic">AATTGTCTAGGCGACC</span>.</p> Full article ">Figure 10
<p>Estimate of the solution probability trend as a numerical estimate of expected results of a sample run.</p> Full article ">Figure 11
<p>Results of a sample execution of QiBAM on QX Simulator, which matches the results derived analytically in <a href="#electronics-10-02433-f010" class="html-fig">Figure 10</a>.</p> Full article ">
23 pages, 9017 KiB  
Article
A Three-Stage Data-Driven Approach for Determining Reaction Wheels’ Remaining Useful Life Using Long Short-Term Memory
by Md Sirajul Islam and Afshin Rahimi
Electronics 2021, 10(19), 2432; https://doi.org/10.3390/electronics10192432 - 7 Oct 2021
Cited by 13 | Viewed by 2867
Abstract
Reaction wheels are widely used in the attitude control system of small satellites. Unfortunately, reaction wheels failure restricts the efficacy of a satellite, and it is one of the many reasons leading to premature abandonment of the satellites. This study observes the measurable [...] Read more.
Reaction wheels are widely used in the attitude control system of small satellites. Unfortunately, reaction wheels failure restricts the efficacy of a satellite, and it is one of the many reasons leading to premature abandonment of the satellites. This study observes the measurable system parameter of a faulty reaction wheel induced with incipient fault to estimate the remaining useful life of the reaction wheels. We achieve this goal in three stages, as none of the observable system parameters are directly related to the health of a reaction wheel. In the first stage, we identify the necessary observable system parameter and predict the future of these parameters using sensor acquired data and a long short-term memory recurrent neural network. In the second stage, we estimate the health index parameter using a multivariate long short-term memory network. In the third stage, we predict the remaining useful life of reaction wheels based on historical data of the health index parameter. Normalized root mean squared error is used to evaluate the performance of the various models in each stage. Additionally, three different timespans (short, moderate, and extended in the scale of small satellite orbit times) are simulated and tested for the performance of the proposed methodology regarding the malfunction of reaction wheels. Furthermore, the robustness of the proposed method to missing values, input frequency, and noise is studied. The results show promising performance for the proposed scheme with accuracy in predicting health index parameter around 0.01–0.02 normalized root mean squared error, the accuracy in prediction of RUL of 1%–2.5%, and robustness to various uncertainty factors, as discussed above. Full article
(This article belongs to the Special Issue Advances in Machine Condition Monitoring and Fault Diagnosis)
►▼ Show Figures

Figure 1

Figure 1
<p>Steps towards prognosis of satellite RW.</p> Full article ">Figure 2
<p>Bialke’s ITHACO Type-A reaction wheel model [<a href="#B35-electronics-10-02432" class="html-bibr">35</a>].</p> Full article ">Figure 3
<p>A block inside an LSTM network. Adapted from [<a href="#B39-electronics-10-02432" class="html-bibr">39</a>].</p> Full article ">Figure 4
<p>Fast Fourier transform of a sample data to obtain <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mrow> <mi>m</mi> <mi>a</mi> <mi>x</mi> </mrow> </msub> </mrow> </semantics></math>.</p> Full article ">Figure 5
<p>RW simulation data sample from ITHACO model.</p> Full article ">Figure 6
<p>Comparison between fitted model and original data for state reconstruction.</p> Full article ">Figure 7
<p>Prediction of RW speed data.</p> Full article ">Figure 8
<p>Prediction of motor current data.</p> Full article ">Figure 9
<p>Prediction of voltage input data.</p> Full article ">Figure 10
<p>State prediction model loss trends for training and validation sets (<math display="inline"><semantics> <mrow> <msub> <mi>ω</mi> <mi>m</mi> </msub> </mrow> </semantics></math>).</p> Full article ">Figure 11
<p>State prediction model loss trends for training and validation sets (<math display="inline"><semantics> <mrow> <msub> <mi>I</mi> <mi>m</mi> </msub> </mrow> </semantics></math>).</p> Full article ">Figure 12
<p>State prediction model loss trends for training and validation sets (<math display="inline"><semantics> <mrow> <msub> <mi>V</mi> <mrow> <mi>c</mi> <mi>o</mi> <mi>m</mi> <mi>m</mi> </mrow> </msub> </mrow> </semantics></math>).</p> Full article ">Figure 13
<p>Prediction of HI parameter (<math display="inline"><semantics> <mrow> <msub> <mi>k</mi> <mi>t</mi> </msub> </mrow> </semantics></math>).</p> Full article ">Figure 14
<p>Parameter prediction model loss trends for training and validation sets (<math display="inline"><semantics> <mrow> <msub> <mi>k</mi> <mi>t</mi> </msub> </mrow> </semantics></math>).</p> Full article ">Figure 15
<p>Confidence boundary of predicted HI parameter and RUL for hours span.</p> Full article ">Figure 16
<p>Predicted HI parameter and RUL for days span.</p> Full article ">Figure 17
<p>Predicted HI parameter and RUL for months span.</p> Full article ">
15 pages, 4524 KiB  
Article
Development of C-Shaped Parasitic MIMO Antennas for Mutual Coupling Reduction
by Hamizan Yon, Nurul Huda Abd Rahman, Mohd Aziz Aris, Mohd Haizal Jamaluddin, Irene Kong Cheh Lin, Hadi Jumaat, Fatimah Nur Mohd Redzwan and Yoshihide Yamada
Electronics 2021, 10(19), 2431; https://doi.org/10.3390/electronics10192431 - 7 Oct 2021
Cited by 26 | Viewed by 2974
Abstract
In the 5G system, multiple-input multiple-output (MIMO) antennas for both transmitting and receiving ends are required. However, the design of MIMO antennas at the 5G upper band is challenging due to the mutual coupling issues. Many techniques have been proposed to improve antenna [...] Read more.
In the 5G system, multiple-input multiple-output (MIMO) antennas for both transmitting and receiving ends are required. However, the design of MIMO antennas at the 5G upper band is challenging due to the mutual coupling issues. Many techniques have been proposed to improve antenna isolation; however, some of the designs have impacts on the antenna performance, especially on the gain and bandwidth reduction, or an increase in the overall size. Thus, a design with a detailed trade-off study must be implemented. This article proposes a new C-shaped parasitic structure around a main circular radiating patch of a MIMO antenna at 16 GHz with enhanced isolation features. The proposed antenna comprises two elements with a separation of 0.32λ edge to edge between radiation parts placed in a linear configuration with an overall dimension of 15 mm × 26 mm. The C-shaped parasitic element was introduced around the main radiating antenna for better isolation. Based on the measurement results, the proposed structure significantly improved the isolation from −23.86 dB to −32.32 dB and increased the bandwidth from 1150 MHz to 1400 MHz. For validation, the envelope correlation coefficient (ECC) and the diversity gain (DG) were also measuredas 0.148 dB and 9.89 dB, respectively. Other parameters, such as the radiation pattern, the total average reflection coefficient and the mean effective gain, were also calculated to ensure the validity of the proposed structure. Based on the design work and analysis, the proposed structure was proven to improve the antenna isolation and increase the bandwidth, while maintaining the small overall dimension. Full article
(This article belongs to the Special Issue Antennas in the 5G System)
►▼ Show Figures

Figure 1

Figure 1
<p>Single-element antenna (<b>a</b>) without and (<b>b</b>) with parasitic elements.</p> Full article ">Figure 2
<p>S11 performance of both antennas.</p> Full article ">Figure 3
<p>Antenna bandwidth for various widths, Wp, and gap distances, Wg.</p> Full article ">Figure 4
<p>Two-element MIMO antenna (<b>a</b>) without parasitic elements (Antenna 1) and (<b>b</b>) with parasitic elements (Antenna 2).</p> Full article ">Figure 5
<p>Isolation for various spacing distances, <span class="html-italic">d</span>, (<b>a</b>) without (Antenna 1) and (<b>b</b>) with parasitic elements (Antenna 2).</p> Full article ">Figure 6
<p>Comparison of S-parameters for Antenna 1 and Antenna 2.</p> Full article ">Figure 7
<p>Proposed antenna with C-shaped parasitic structure.</p> Full article ">Figure 8
<p>E-field distribution for (<b>a</b>) Antenna 1, (<b>b</b>) Antenna 2, (<b>c</b>) 3D view (without parasitic element) and (<b>d</b>) 3D view (with parasitic element).</p> Full article ">Figure 8 Cont.
<p>E-field distribution for (<b>a</b>) Antenna 1, (<b>b</b>) Antenna 2, (<b>c</b>) 3D view (without parasitic element) and (<b>d</b>) 3D view (with parasitic element).</p> Full article ">Figure 9
<p>Current distribution for (<b>a</b>) Antenna 1 and (<b>b</b>) Antenna 2.</p> Full article ">Figure 10
<p>Current distribution for (<b>a</b>) Antenna 1 and (<b>b</b>) Antenna 2.</p> Full article ">Figure 11
<p>Measurement and simulation results of S-parameter for Antenna 1 and Antenna 2.</p> Full article ">Figure 12
<p>Measured and simulated radiation pattern for MIMO antenna: (<b>a</b>) E-plane and (<b>b</b>) H-plane.</p> Full article ">Figure 13
<p>Measurement of total average reflection coefficient (TARC).</p> Full article ">Figure 14
<p>Measurement and simulation of the enveloped correlation coefficient (ECC).</p> Full article ">Figure 15
<p>Measurement and simulation of the diversity gain (DG).</p> Full article ">Figure 16
<p>Measurement of the mean effective gain (MEG).</p> Full article ">
9 pages, 10159 KiB  
Article
Inverse Design of a Microstrip Meander Line Slow Wave Structure with XGBoost and Neural Network
by Yijun Zhu, Yang Xie, Ningfeng Bai and Xiaohan Sun
Electronics 2021, 10(19), 2430; https://doi.org/10.3390/electronics10192430 - 7 Oct 2021
Cited by 7 | Viewed by 2264
Abstract
We present a new machine learning (ML) deep learning (DL) synthesis algorithm for the design of a microstrip meander line (MML) slow wave structure (SWS). Exact numerical simulation data are used in the training of our network as a form of supervised learning. [...] Read more.
We present a new machine learning (ML) deep learning (DL) synthesis algorithm for the design of a microstrip meander line (MML) slow wave structure (SWS). Exact numerical simulation data are used in the training of our network as a form of supervised learning. The learning results show that the training mean squared error is as low as 5.23 × 10−2 when using 900 sets of data. When the desired performance is reached, workable geometry parameters can be obtained by this algorithm. A D-band MML SWS with 20 GHz bandwidth at 160 GHz center frequency is then designed using the auto-design neural network (ADNN). A cold test shows that its phase velocity varies by 0.005 c, and the transmission rate of a 50-period SWS is greater than −5 dB with the reflectivity below −15 dB when the frequency is from 150 to 170 GHz. Particle-in-cell (PIC) simulation also illustrates that a maximum power of 3.2 W is reached at 160 GHz with 34.66 dB gain and output power greater than 1 W from 152 to 168 GHz. Full article
(This article belongs to the Special Issue High-Frequency Vacuum Electron Devices)
►▼ Show Figures

Figure 1

Figure 1
<p>(<b>a</b>) Front view; and (<b>b</b>) top view of the MML unit structure.</p> Full article ">Figure 1 Cont.
<p>(<b>a</b>) Front view; and (<b>b</b>) top view of the MML unit structure.</p> Full article ">Figure 2
<p>Cold-test characteristic of the MML structure: (<b>a</b>) Dispersion characteristic; (<b>b</b>) Transmission (S21) and reflection (S11).</p> Full article ">Figure 3
<p>(<b>a</b>) MMLPM basic structure; (<b>b</b>) the corresponding optimal value with phase velocity; (<b>c</b>) the corresponding optimal value with phase velocity flatness; (<b>d</b>) the corresponding optimal value with <span class="html-italic">Smax;</span> (<b>e</b>) the corresponding optimal value with <span class="html-italic">Smin</span>.</p> Full article ">Figure 3 Cont.
<p>(<b>a</b>) MMLPM basic structure; (<b>b</b>) the corresponding optimal value with phase velocity; (<b>c</b>) the corresponding optimal value with phase velocity flatness; (<b>d</b>) the corresponding optimal value with <span class="html-italic">Smax;</span> (<b>e</b>) the corresponding optimal value with <span class="html-italic">Smin</span>.</p> Full article ">Figure 4
<p>(<b>a</b>) XGBoost-DNN composite structure; (<b>b</b>) basic structure of a fully connected deep neural network; (<b>c</b>) three different active functions; (<b>d</b>) training loss with different active functions.</p> Full article ">Figure 5
<p>(<b>a</b>) Transmission characteristics from 150 to 170 GHz of the designed MML-SWS; (<b>b</b>) dispersion characteristic and coupled impedance of the structure.</p> Full article ">Figure 6
<p>(<b>a</b>)Schematic model for PIC simulation; (<b>b</b>) output power and gain versus input signal power; (<b>c</b>) output power and gain versus frequency.</p> Full article ">
18 pages, 36122 KiB  
Article
Reconfigurable Morphological Processor for Grayscale Image Processing
by Bin Zhang
Electronics 2021, 10(19), 2429; https://doi.org/10.3390/electronics10192429 - 7 Oct 2021
Cited by 7 | Viewed by 2722
Abstract
Grayscale morphology is a powerful tool in image, video, and visual applications. A reconfigurable processor is proposed for grayscale image morphological processing. The architecture of the processor is a combination of a reconfigurable grayscale processing module (RGPM) and peripheral circuits. The RGPM, which [...] Read more.
Grayscale morphology is a powerful tool in image, video, and visual applications. A reconfigurable processor is proposed for grayscale image morphological processing. The architecture of the processor is a combination of a reconfigurable grayscale processing module (RGPM) and peripheral circuits. The RGPM, which consists of four grayscale computing units, conducts grayscale morphological operations and implements related algorithms of more than 100 f/s for a 1024 × 1024 image. The periphery circuits control the entire image processing and dynamic reconfiguration process. Synthesis results show that the proposed processor can provide 43.12 GOPS and achieve 8.87 GOPS/mm2 at a 220-MHz system clock. The simulation and experimental results show that the processor is suitable for high-performance embedded systems. Full article
(This article belongs to the Special Issue Design Tools and Architectures for Coarse-Grained Reconfigurable Computing)
►▼ Show Figures

Figure 1

Figure 1
<p>Example of maximum computing circuit.</p> Full article ">Figure 2
<p>Example of maximum computing circuit.</p> Full article ">Figure 3
<p>Architecture of grayscale image processor.</p> Full article ">Figure 4
<p>Implementation of RGPM.</p> Full article ">Figure 5
<p>Architecture of the GCU.</p> Full article ">Figure 6
<p>Block diagram of input control logic unit.</p> Full article ">Figure 7
<p>Circuit of maximum computing in MMCU2.</p> Full article ">Figure 8
<p>Two reconfigured architectures of GCU: (<b>a</b>) one 5 × 5 and two 3 × 3 operations are implemented simultaneously; (<b>b</b>) one 7 × 7 operation is implemented.</p> Full article ">Figure 9
<p>(<b>a</b>) 5 × 5 dilation; (<b>b</b>) 5 × 5 erosion; (<b>c</b>) 5 × 5 opening; (<b>d</b>) 5 × 5 closing.</p> Full article ">Figure 10
<p>Architecture of grayscale image processing system.</p> Full article ">Figure 11
<p>Implementation and processing results of eight pipelined 5 × 5 dilation operations.</p> Full article ">Figure 12
<p>Implementation and processing results of eight pipelined 5 × 5 erosion operations.</p> Full article ">Figure 13
<p>Implementation and processing results of 5 × 5 opening and closing operations.</p> Full article ">Figure 14
<p>Implementation and processing results of hit-and-miss operation.</p> Full article ">Figure 15
<p>Implementation and processing results of thinning and thickening.</p> Full article ">Figure 16
<p>Implementation and processing results of gradient.</p> Full article ">Figure 17
<p>Implementation and processing results of white top-hat transform.</p> Full article ">
30 pages, 40138 KiB  
Article
Per-Core Power Modeling for Heterogenous SoCs
by Ganapati Bhat, Sumit K. Mandal, Sai T. Manchukonda, Sai V. Vadlamudi, Ayushi Agarwal, Jun Wang and Umit Y. Ogras
Electronics 2021, 10(19), 2428; https://doi.org/10.3390/electronics10192428 - 7 Oct 2021
Cited by 1 | Viewed by 3193
Abstract
State-of-the-art mobile platforms, such as smartphones and tablets, are powered by heterogeneous system-on-chips (SoCs). These SoCs are composed of many processing elements, including multiple CPU core clusters (e.g., big.LITTLE cores), graphics processing units (GPUs), memory controllers and other on-chip resources. On the one [...] Read more.
State-of-the-art mobile platforms, such as smartphones and tablets, are powered by heterogeneous system-on-chips (SoCs). These SoCs are composed of many processing elements, including multiple CPU core clusters (e.g., big.LITTLE cores), graphics processing units (GPUs), memory controllers and other on-chip resources. On the one hand, mobile platforms need to provide a swift response time for interactive apps and high throughput for graphics-oriented workloads; on the other hand, the power consumption must be under tight control to prevent high skin temperatures and energy consumption. Therefore, commercial systems feature a range of mechanisms for dynamic power and temperature control. However, these techniques rely on simple indicators, such as core utilization and total power consumption. System architects are typically limited to the total power consumption, since multiple resources share the same power rail. More importantly, most of the power rails are not exposed to the input/output pins. To address this challenge, this paper presents a thorough methodology to model the power consumption of major resources in heterogeneous SoCs. The proposed models utilize a wide range of performance counters to capture the workload dynamics accurately. Experimental validation on a Nexus 6P phone, powered by an octa-core Snapdragon 810 SoC, showed that the proposed models can estimate the power consumption within a 10% error margin. Full article
(This article belongs to the Section Microelectronics)
►▼ Show Figures

Figure 1

Figure 1
<p>Overview of the power modeling.</p> Full article ">Figure 2
<p>Connection of external power supply to the phone.</p> Full article ">Figure 3
<p>(<b>a</b>) Frequency domain spectrum of the data before and after filtering. The black rectangle shows the major component of the power. (<b>b</b>) Time domain representation of the raw power, filtered power, and de-spiked power.</p> Full article ">Figure 4
<p>Display power variation with brightness and color.</p> Full article ">Figure 5
<p>Comparison of actual and predicted display power for different colors.</p> Full article ">Figure 6
<p>The test image.</p> Full article ">Figure 7
<p>Comparison of actual and predicted display power for the test image.</p> Full article ">Figure 8
<p>Power estimation when total power is dominated by leakage of A57 cores.</p> Full article ">Figure 9
<p>Reference and estimated dynamic power consumption for the A57 cluster running at 1.24 GHz.</p> Full article ">Figure 10
<p>(<b>a</b>) Behavior of the total power of the A53 cluster running at 860 MHz as a function of the temperature. The figure shows both measured and estimated power at two configurations. (<b>b</b>) Behavior of the dynamic power with temperature. The dynamic power is constant since the processor is idle. (<b>c</b>) Behavior of the leakage power with respect to the temperature. The leakage power shows an increase with temperature due to the temperature term in Equation (5).</p> Full article ">Figure 11
<p>Comparison of measured and estimated dynamic power for the A53 cluster running at 1.24 GHz. Each sample is 50 ms.</p> Full article ">Figure 12
<p>Variation of power with temperature and frequency.</p> Full article ">Figure 13
<p>The reference and estimated C_(dyn,gpu) at 600 MHz.</p> Full article ">Figure 14
<p>Actual power consumption and average execution time of PCA benchmark for all the bandwidths.</p> Full article ">Figure 15
<p>Comparison of actual and predicted memory power for PCA and Stream benchmarks.</p> Full article ">Figure 16
<p>Actual power consumption and execution time of the Candy Crush game for all the GPU memory bandwidths.</p> Full article ">Figure 17
<p>Comparison of actual and predicted memory power for the Candy Crush and Angry Birds games benchmarks.</p> Full article ">Figure 18
<p>(<b>a</b>). Instructions for the Angry Bird game app without alignment (<b>b</b>). Instructions for the Angry Bird game app with alignment of instructions.</p> Full article ">Figure 19
<p>Comparison of actual and predicted memory power for the Angry Birds game benchmark.</p> Full article ">Figure 20
<p>Comparison of the reference power consumption and the estimated power consumption for (<b>a</b>) BasicMath, (<b>b</b>) PCA, (<b>c</b>) MEL, (<b>d</b>) FFT, and (<b>e</b>) Spectral benchmarks using leave-one-out analysis.</p> Full article ">Figure 21
<p>(<b>a</b>). Wi-Fi packet transfer for an application download (<b>b</b>). Wi-Fi packet transfer for Google Hangouts Call.</p> Full article ">Figure 22
<p>Comparison of reference and estimated power for the WiFi power.</p> Full article ">Figure 23
<p>Comparison of reference and estimation of the Wi-Fi power.</p> Full article ">Figure 24
<p>Comparison of each power component when the WiFi is turned off randomly.</p> Full article ">
24 pages, 21801 KiB  
Article
Assessment and Improvement of the Pattern Recognition Performance of Memdiode-Based Cross-Point Arrays with Randomly Distributed Stuck-at-Faults
by Fernando L. Aguirre, Sebastián M. Pazos, Félix Palumbo, Antoni Morell, Jordi Suñé and Enrique Miranda
Electronics 2021, 10(19), 2427; https://doi.org/10.3390/electronics10192427 - 6 Oct 2021
Cited by 3 | Viewed by 2968
Abstract
In this work, the effect of randomly distributed stuck-at faults (SAFs) in memristive cross-point array (CPA)-based single and multi-layer perceptrons (SLPs and MLPs, respectively) intended for pattern recognition tasks is investigated by means of realistic SPICE simulations. The quasi-static memdiode model (QMM) is [...] Read more.
In this work, the effect of randomly distributed stuck-at faults (SAFs) in memristive cross-point array (CPA)-based single and multi-layer perceptrons (SLPs and MLPs, respectively) intended for pattern recognition tasks is investigated by means of realistic SPICE simulations. The quasi-static memdiode model (QMM) is considered here for the modelling of the synaptic weights implemented with memristors. Following the standard memristive approach, the QMM comprises two coupled equations, one for the electron transport based on the double-diode equation with a single series resistance and a second equation for the internal memory state of the device based on the so-called logistic hysteron. By modifying the state parameter in the current-voltage characteristic, SAFs of different severeness are simulated and the final outcome is analysed. Supervised ex-situ training and two well-known image datasets involving hand-written digits and human faces are employed to assess the inference accuracy of the SLP as a function of the faulty device ratio. The roles played by the memristor’s electrical parameters, line resistance, mapping strategy, image pixelation, and fault type (stuck-at-ON or stuck-at-OFF) on the CPA performance are statistically analysed following a Monte-Carlo approach. Three different re-mapping schemes to help mitigate the effect of the SAFs in the SLP inference phase are thoroughly investigated. Full article
(This article belongs to the Special Issue RRAM Devices: Multilevel State Control and Applications)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>(<b>a</b>) Sketch of the CPA structure. Red and blue arrows exemplify the electron flow through the memdiodes connecting the top (word lines (WL)) and bottom lines (bit lines (BL)). Different resistance states are schematically represented (high resistance state (HRS) to low resistance state (LRS)). The dashed blue line depicts the so-called sneakpath problem. The parasitic wire resistance is indicated for WL<span class="html-italic"><sub>i</sub></span> and BL<span class="html-italic"><sub>i</sub></span>. (<b>b</b>) Schematic representation of the MIM structure where the RS mechanism takes place, before the forming step and during the LRS-to-HRS alternate transition. Blue and red balls represent the metal ions and oxygen vacancies (VOs), respectively.</p> Full article ">Figure 2
<p>(<b>a</b>) Hysteron model with logistic ridge functions <span class="html-italic">Γ<sup>+</sup></span> (Equation (3)) and <span class="html-italic">Γ<sup>−</sup></span> (Equation (4)). Ω is the space of feasible states <span class="html-italic">S</span>. The black thick faded line superimposed on the hysteron model indicates the trajectory of the state variable λ inside Ω from an initial S<sub>1</sub> to a final S<sub>2</sub> state. Note that four transport mechanisms are considered for the pre-forming conduction, with the forming event taking place at the same voltage. The inset in the left shows the equivalent circuit model for the current equation (Equation 1) including the series resistance. The diodes are driven by the memory state of the device and one diode is activated at a time. Typical <span class="html-italic">I-V</span> characteristics for a memdiode [<a href="#B11-electronics-10-02427" class="html-bibr">11</a>] obtained via the simulation of the proposed model are superimposed. Current evolution is indicated by the blue arrows. The inset on the right side shows the exponential (HRS) to lineal (LRS) transition by varying the value of λ. The red shaded region indicates the possible voltages applied to the device. <span class="html-italic">I<sub>HRS</sub></span> and <span class="html-italic">I<sub>LRS</sub></span> currents are pinpointed at a fitting voltage with the grey and white circle markers, respectively. The overestimation of <span class="html-italic">I<sub>HRS</sub></span> may occur when considering a linear model [<a href="#B29-electronics-10-02427" class="html-bibr">29</a>] for the HRS regime, and lower applied voltages as indicated by the cyan, blue and black ball markers. (<b>b</b>) Experimental <span class="html-italic">I-V</span> loops of different materials reported in the literature, fitted with the QMM model: HfO<sub>2</sub> [<a href="#B64-electronics-10-02427" class="html-bibr">64</a>] and LMCO [<a href="#B65-electronics-10-02427" class="html-bibr">65</a>].</p> Full article ">Figure 3
<p>The change in the distribution of the elements of <span class="html-italic">W<sub>M</sub></span> (<b>a</b>) under different normalisation techniques is shown in (<b>b</b>–<b>f</b>). Inference accuracy as function of the FD ratio, considering different <span class="html-italic">W<sub>M</sub></span> normalisation approaches, is presented for (<b>g</b>) SA1, (<b>h</b>) SA0, and (<b>i</b>) SA0_nE faults. The SLP power consumption during the inference phase is indicated in the inset of (<b>h</b>) as a function of the SLP size. Similarly, the inference accuracy of the fault-free SLP under different normalisation methods is presented in (<b>i</b>). (<b>j</b>) Inference accuracy assuming different combinations of SA1 and SA0 faults. Note that the ratio of SAFs (containing both SA1 and SA0) is swept parametrically from 5% to 30%. (<b>k</b>) and (<b>l</b>) show the inference accuracy of an MLP ANN as a function of the ratio of SAFs, assuming SA1 and SA0, respectively.</p> Full article ">Figure 4
<p>Inference accuracy vs. FD ratio for different values or <span class="html-italic">R<sub>L</sub></span> is presented for the (<b>a</b>) SA1, (<b>b</b>) SA0, and (<b>c</b>) SA0_nE cases. (<b>d</b>) CPA Inference accuracy vs. ratio of biased stuck-at-ON devices. Each marker corresponds to an MC run. Data are codified in terms of the nominal fault ratio (marker type) and CPA size (marker colour), e.g., blue circle markers indicate the inference accuracy results for simulations of the 8 × 8 px. image CPAs with 1% of faulty devices, whereas red pentagon markers stand for the results obtained from 16 × 16 px. image CPAs and 30% of faulty devices. The case of 30% of faulty devices (pentagonal markers) have been highlighted for the three CPA sizes considered to provide a guide to the eye: As the CPA size increases, the ratio of biased faulty devices decreases from ~5% in the 1280 sys. CPA, to ~3.7% in the 5120 sys. and finally to ~2.3% in the 15,680 sys CPA.</p> Full article ">Figure 5
<p>(<b>a</b>) Sketched representation of the fault-tolerant adaptative mapping process (Algorithm 1) depicting the conductance compensation (top) that allows the tolerance of faults in the first and last rows (green-shaded cells) but which is uncapable of handling other SAFs (unrecoverable faults, grey-shaded cells). A row permutation approach (bottom) is required to turn unrecoverable faults into recoverable faults (See <a href="#electronics-10-02427-t002" class="html-table">Table 2</a>). (<b>b</b>) Row permutation is also used for Algorithms 2 and 3. In the latter, it is employed to re-map the faultiest CPA rows to the inactive image pixels. The MNIST case is shown as an example.</p> Full article ">Figure 6
<p>(<b>a</b>) Samples of Yale Face Database B showing 3 classes with 32 × 32 px. (top) and 16 × 16 px. (bottom) resolutions. In both cases, the <span class="html-italic">x</span> and <span class="html-italic">y</span> axis in the leftmost image stands for the pixel index. The re-mapping Algorithms 1–3 are tested with the MNIST dataset for the SA1 and SA0 faults in Figures (<b>b</b>) and (<b>c</b>), respectively. The corresponding trends for Yale Face Database B are shown in Figures (<b>d</b>) and (<b>e</b>). In both cases, Algorithm 1 shows the best results for SA1 faults and Algorithm 2 is the preferred one to tolerate SA0 faults.</p> Full article ">
19 pages, 3099 KiB  
Article
Nonlinear Model Predictive Control of Single-Link Flexible-Joint Robot Using Recurrent Neural Network and Differential Evolution Optimization
by Anlong Zhang, Zhiyun Lin, Bo Wang and Zhimin Han
Electronics 2021, 10(19), 2426; https://doi.org/10.3390/electronics10192426 - 6 Oct 2021
Cited by 16 | Viewed by 3768
Abstract
A recurrent neural network (RNN) and differential evolution optimization (DEO) based nonlinear model predictive control (NMPC) technique is proposed for position control of a single-link flexible-joint (FJ) robot. First, a simple three-layer recurrent neural network with rectified linear units as an activation function [...] Read more.
A recurrent neural network (RNN) and differential evolution optimization (DEO) based nonlinear model predictive control (NMPC) technique is proposed for position control of a single-link flexible-joint (FJ) robot. First, a simple three-layer recurrent neural network with rectified linear units as an activation function (ReLU-RNN) is employed for approximating the system dynamic model. Then, using the RNN predictive model and model predictive control (MPC) scheme, an RNN and DEO based NMPC controller is designed, and the DEO algorithm is used to solve the controller. Finally, comparing numerical simulation findings demonstrates the efficiency and performance of the proposed approach. The merit of this method is that not only is the control precision satisfied, but also the overshoots and the residual vibration are well suppressed. Full article
(This article belongs to the Collection Predictive and Learning Control in Engineering Applications)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>The architecture of single-link FJ robot system.</p> Full article ">Figure 2
<p>The ReLU-RNN architecture used to approximate system dynamic model.</p> Full article ">Figure 3
<p>The flow chart of DEO algorithm.</p> Full article ">Figure 4
<p>The architecture of RNN and DEO based NMPC controller.</p> Full article ">Figure 5
<p>The progress of multi-step prediction.</p> Full article ">Figure 6
<p>The absolute errors of multi-step prediction.</p> Full article ">Figure 7
<p>The cost values of the optimization process at five adjacent time steps.</p> Full article ">Figure 8
<p>The cost values of the optimization process at each time step.</p> Full article ">Figure 9
<p>The target tracking process of different controllers.</p> Full article ">Figure 10
<p>The control actions of different controllers.</p> Full article ">Figure 11
<p>The target tracking process with external disturbances.</p> Full article ">Figure 12
<p>The control actions with external disturbances.</p> Full article ">
14 pages, 4980 KiB  
Article
The Prediction of Capacity Trajectory for Lead–Acid Battery Based on Steep Drop Curve of Discharge Voltage and Gaussian Process Regression
by Qian Li, Guangzhen Liu, Ji’ang Zhang, Zhan Su, Chunyan Hao, Ju He and Ze Cheng
Electronics 2021, 10(19), 2425; https://doi.org/10.3390/electronics10192425 - 6 Oct 2021
Cited by 4 | Viewed by 2402
Abstract
In this paper, a method of capacity trajectory prediction for lead-acid battery, based on the steep drop curve of discharge voltage and improved Gaussian process regression model, is proposed by analyzing the relationship between the current available capacity and the voltage curve of [...] Read more.
In this paper, a method of capacity trajectory prediction for lead-acid battery, based on the steep drop curve of discharge voltage and improved Gaussian process regression model, is proposed by analyzing the relationship between the current available capacity and the voltage curve of short-time discharging. The battery under average charging is discharged for a short time, and the voltage drop of short-time discharging during equal time intervals, which has the highest relevance with presently available capacity, is extracted as the health feature (HF), and the ergodic method is used to search the optimal time interval. Then, a Gaussian process regression (GPR) model, which reflects the capacity degradation of lead–acid battery, is established (with the HF series as input and current available capacity series as output). Considering the complex trend of capacity trajectory, the rational quadratic covariance function is used as the kernel function of GPR model, and the conjugate gradient algorithm is used for optimization, in order to improve the nonlinear mapping ability of GPR. Finally, the experimental results of lead-acid batteries under different charging cut-off voltages and operating temperatures show that the proposed method can effectively predict the capacity change trajectory of lead–acid batteries with a small training sample, showing high prediction accuracy and wide applicability. Full article
(This article belongs to the Section Power Electronics)
►▼ Show Figures

Figure 1

Figure 1
<p>The experimental bench of lead-acid battery.</p> Full article ">Figure 2
<p>The capacity degradation trajectory under different working conditions.</p> Full article ">Figure 3
<p>Charge-discharge voltage and current curve for one cycle.</p> Full article ">Figure 4
<p>The discharge voltage curve of the battery under various working conditions and different cycle times.</p> Full article ">Figure 5
<p>The 3D graph of relevance and HF-capacity trajectory for each battery.</p> Full article ">Figure 5 Cont.
<p>The 3D graph of relevance and HF-capacity trajectory for each battery.</p> Full article ">Figure 6
<p>The flow chart of the proposed method.</p> Full article ">Figure 7
<p>The estimation results of capacity degradation trajectory and relatively error for each battery.</p> Full article ">Figure 7 Cont.
<p>The estimation results of capacity degradation trajectory and relatively error for each battery.</p> Full article ">
20 pages, 2669 KiB  
Article
Fractional Order Adaptive Fast Super-Twisting Sliding Mode Control for Steer-by-Wire Vehicles with Time-Delay Estimation
by Yong Yang, Yunbing Yan and Xiaowei Xu
Electronics 2021, 10(19), 2424; https://doi.org/10.3390/electronics10192424 - 5 Oct 2021
Cited by 7 | Viewed by 2171
Abstract
It is difficult to model and determine the parameters of the steer-by-wire (SBW) system accurately, and the perturbation is variable with complex and changeable tire–road conditions. In order to improve the control performance of the vehicle SBW system, an adaptive fast super-twisting sliding [...] Read more.
It is difficult to model and determine the parameters of the steer-by-wire (SBW) system accurately, and the perturbation is variable with complex and changeable tire–road conditions. In order to improve the control performance of the vehicle SBW system, an adaptive fast super-twisting sliding mode control (AFST-SMC) scheme with time-delay estimation (TDE) is proposed. The proposed scheme uses TDE to acquire the lumped dynamics in a simple way and establishes a practical model-free structure. Then, a fractional order (FO) sliding mode surface and a fast super-twisting sliding mode control structure were designed on the basic super-twisting sliding mode to ensure fast convergence and high control accuracy. Since the uncertain boundary information of the actual system is unknown, a novel adaptive algorithm is proposed to regulate the control gain based on the control errors. Theoretical analysis concerning system stability is given based on the Lyapunov theory. Finally, the effectiveness of the method is verified through comparative experiments. The results show that the proposed TDE-AFST-FOSMC control scheme has the advantages of model-free, fast response and high accuracy. Full article
(This article belongs to the Section Systems & Control Engineering)
►▼ Show Figures

Figure 1

Figure 1
<p>Convergence time comparison of different super-twisting structures.</p> Full article ">Figure 2
<p>Block diagram of TDE-based AFST-FOSMC control scheme.</p> Full article ">Figure 3
<p>Experimental platform of the SBW system.</p> Full article ">Figure 4
<p>Tracking performance and the control voltage: (<b>a</b>) ASMC algorithm; (<b>b</b>) TDE-STSMC algorithm; (<b>c</b>) TDE-FST-FOSMC algorithm; (<b>d</b>) TDE-AFST-FOSMC algorithm.</p> Full article ">Figure 4 Cont.
<p>Tracking performance and the control voltage: (<b>a</b>) ASMC algorithm; (<b>b</b>) TDE-STSMC algorithm; (<b>c</b>) TDE-FST-FOSMC algorithm; (<b>d</b>) TDE-AFST-FOSMC algorithm.</p> Full article ">Figure 5
<p>Estimation of disturbance torque in slalom path following.</p> Full article ">Figure 6
<p>Adaptive effect of control parameter <span class="html-italic">L</span>(t).</p> Full article ">Figure 7
<p>Front wheel deviation under shock disturbance.</p> Full article ">Figure 8
<p>Performance of estimation of shock disturbance torque.</p> Full article ">Figure 9
<p>Tracking performance of steering wheel angle.</p> Full article ">Figure 10
<p>Tracking error of steering wheel angle.</p> Full article ">Figure 11
<p>Estimation performance of disturbance torque.</p> Full article ">
16 pages, 6178 KiB  
Article
A Low-Power GPIO-Based Size Sensor to Monitor the Imbibition of Corn Seeds Beneath Soil
by Ehab A. Hamed, Jordan Athas, Xincheng Zhang, Noah Ashenden and Inhee Lee
Electronics 2021, 10(19), 2423; https://doi.org/10.3390/electronics10192423 - 4 Oct 2021
Viewed by 2290
Abstract
Seed imbibition, absorption of water by a dry seed, is an essential process in which embryo hydration and root establishment occur. In natural environments, this process occurs beneath the soil, making it difficult to observe preliminary growth of seeds. This paper presents a [...] Read more.
Seed imbibition, absorption of water by a dry seed, is an essential process in which embryo hydration and root establishment occur. In natural environments, this process occurs beneath the soil, making it difficult to observe preliminary growth of seeds. This paper presents a new technique for tracking the imbibition of corn seeds. The proposed system is designed to measure imbibition through seed expansion and wirelessly transmit data, permitting the system to remain beneath the soil with the subject seed. The system consists of low-cost commercial off-the-shelf components and 3D-printed probes. The proposed system is optimized to measure the size of multiple seeds with a single Analog-to-Digital Converter (ADC) pin by utilizing the General-Purpose Input Output (GPIO) pins of the microcontroller, to reconfigure connections to supply voltage or ground. The circuit design of the system shows low power consumption compared to other conventional circuits and utilizes fewer components by taking advantage of the microcontroller GPIOs. Additionally, the proposed circuit design shows less error and insensitivity to the supply voltage variations. Full article
(This article belongs to the Special Issue Application of Wireless Sensor Networks in Accredited Monitoring)
►▼ Show Figures

Figure 1

Figure 1
<p>Proposed system diagram with its main components.</p> Full article ">Figure 2
<p>Proposed circuit design: (<b>a</b>) Entire design with the connection of 3 FSRs and R0 to the GPIO pins; (<b>b</b>) GPIO configuration to get the value of V0; (<b>c</b>) GPIO configuration to get the value of V1.</p> Full article ">Figure 3
<p>Flowchart of the measurement process in the proposed design.</p> Full article ">Figure 4
<p>Flowchart of the measurement process in the proposed design.</p> Full article ">Figure 5
<p>Probe design: (<b>a</b>) Complete probe holding corn seed inside it; (<b>b</b>) Cylinder containing the spring inside it to protect it from the soil.</p> Full article ">Figure 6
<p>Waterproofed FSR: (<b>a</b>) FSR wires coated by marine epoxy; (<b>b</b>) FSR covered by black electrical tape.</p> Full article ">Figure 7
<p>A screenshot of the pulled FSR values from ESP32 through the Bluetooth connection.</p> Full article ">Figure 8
<p>Probes calibration using cuboids: (<b>a</b>) Different sizes of 3D-printed cuboids (4–11.2 mm); (<b>b</b>) How the cuboids are placed inside the probes for calibration.</p> Full article ">Figure 9
<p>FSRs resistance versus cuboids width for 9 different probes.</p> Full article ">Figure 10
<p>Hysteresis curves for 9 different probes showing the measured resistance while increasing and decreasing cuboid sizes: (<b>a</b>) Probe #1; (<b>b</b>) Probe #2; (<b>c</b>) Probe #3; (<b>d</b>) Probe #4; (<b>e</b>) Probe #5; (<b>f</b>) Probe #6; (<b>g</b>) Probe #7; (<b>h</b>) Probe #8; (<b>i</b>) Probe #9.</p> Full article ">Figure 10 Cont.
<p>Hysteresis curves for 9 different probes showing the measured resistance while increasing and decreasing cuboid sizes: (<b>a</b>) Probe #1; (<b>b</b>) Probe #2; (<b>c</b>) Probe #3; (<b>d</b>) Probe #4; (<b>e</b>) Probe #5; (<b>f</b>) Probe #6; (<b>g</b>) Probe #7; (<b>h</b>) Probe #8; (<b>i</b>) Probe #9.</p> Full article ">Figure 11
<p>Conventional methods to measure the resistance of three FSRs: (<b>a</b>) The voltage divider method; (<b>b</b>) The noninverting amplifier method; (<b>c</b>) The Wheatstone bridge method.</p> Full article ">Figure 11 Cont.
<p>Conventional methods to measure the resistance of three FSRs: (<b>a</b>) The voltage divider method; (<b>b</b>) The noninverting amplifier method; (<b>c</b>) The Wheatstone bridge method.</p> Full article ">Figure 12
<p>PCB for the proposed system with 9 probes connected as 3 groups.</p> Full article ">Figure 13
<p>Maximum measurement error of each method from different input resistances: (<b>a</b>) voltage divider, wheatstone bridge, and proposed methods; (<b>b</b>) noniverting amplifier method.</p> Full article ">Figure 14
<p>The corresponding voltage range (V0) that is read by the ADC for the proposed method with and without an amplifier.</p> Full article ">
17 pages, 6901 KiB  
Article
Supercapacitor Assisted Hybrid PV System for Efficient Solar Energy Harnessing
by Kasun Piyumal, Aruna Ranaweera, Sudath Kalingamudali and Nihal Kularatna
Electronics 2021, 10(19), 2422; https://doi.org/10.3390/electronics10192422 - 4 Oct 2021
Cited by 3 | Viewed by 4236
Abstract
In photovoltaic (PV) systems, maximum power point (MPP) is tracked by matching the load impedance to the internal impedance of the PV array by adjusting the duty cycle of the associated DC-DC converter. Scientists are trying to improve the efficiency of these converters [...] Read more.
In photovoltaic (PV) systems, maximum power point (MPP) is tracked by matching the load impedance to the internal impedance of the PV array by adjusting the duty cycle of the associated DC-DC converter. Scientists are trying to improve the efficiency of these converters by improving the performance of the power stage, while limited attention is given to finding alternative methods. This article describes a novel supercapacitor (SC) assisted technique to enhance the efficiency of a PV system without modifying the power stage of the charge controller. The proposed system is an SC—battery hybrid PV system where an SC bank is coupled in series with a PV array to enhance the overall system efficiency. Developed prototype of the proposed system with SC assisted loss circumvention embedded with a DC microgrid application detailed in the article showed that the average efficiency of the PV system is increased by 8%. This article further describes the theoretical and experimental investigation of the impedance matching technique for the proposed PV system, explaining how to adapt typical impedance matching for maximum power transfer. Full article
(This article belongs to the Section Power Electronics)
►▼ Show Figures

Figure 1

Figure 1
<p>Block diagram of a typical standalone photovoltaic system.</p> Full article ">Figure 2
<p>Capacitor charging circuit with series-connected useful Load.</p> Full article ">Figure 3
<p>Efficiency variation of the RC circuit vs. <span class="html-italic">k</span> and <span class="html-italic">m</span>: (<b>a</b>) at <span class="html-italic">P</span> = 0; (<b>b</b>) at <span class="html-italic">P</span> = 1; (<b>c</b>) at <span class="html-italic">P</span> = 10.</p> Full article ">Figure 4
<p>Simplified block diagram of the proposed PV system.</p> Full article ">Figure 5
<p>System block diagram of each operating mode: (<b>a</b>) Neutral; (<b>b</b>) SC charge recovery; (<b>c</b>) SC bypass.</p> Full article ">Figure 6
<p>I-V and P-V characteristics of a PV array.</p> Full article ">Figure 7
<p>Block diagram of a switch-mode DC-DC buck converter including a battery bank and a load connected to its output.</p> Full article ">Figure 8
<p>Variation of the input resistance of the DC-DC buck converter with duty ratio of control PWM signal.</p> Full article ">Figure 9
<p>Block diagram of the proposed system operating under: (<b>a</b>) Neutral mode; (<b>b</b>) SC charge recovery mode.</p> Full article ">Figure 10
<p>Variation of input resistance of the system vs. time and duty ratio when the system is operated under: (<b>a</b>) Neutral mode; (<b>b</b>) Neutral mode at different <math display="inline"><semantics> <mi>t</mi> </semantics></math> values; (<b>c</b>) SC charge recovery mode; (<b>d</b>) SC charge recovery mode at different <math display="inline"><semantics> <mi>t</mi> </semantics></math> values. (<math display="inline"><semantics> <mrow> <mi>C</mi> <mo>=</mo> <mn>108</mn> <mtext> </mtext> <mi mathvariant="normal">F</mi> <mo>,</mo> <mtext> </mtext> <msub> <mi>r</mi> <mi>C</mi> </msub> <mo>=</mo> <mn>4.8</mn> <mtext> </mtext> <mi mathvariant="normal">m</mi> <mo>Ω</mo> <mo>,</mo> <mtext> </mtext> <msub> <mi>R</mi> <mi>L</mi> </msub> <mo>=</mo> <mn>100</mn> <mtext> </mtext> <mo>Ω</mo> <mo>,</mo> <mtext> </mtext> <msub> <mi>R</mi> <mrow> <mi>L</mi> <mo>−</mo> <mi>D</mi> <mi>C</mi> </mrow> </msub> <mo>=</mo> <mn>20</mn> <mtext> </mtext> <mo>Ω</mo> </mrow> </semantics></math> and <math display="inline"><semantics> <mrow> <msub> <mi>r</mi> <mi>b</mi> </msub> <mo>=</mo> <mn>0.35</mn> <mtext> </mtext> <mo>Ω</mo> </mrow> </semantics></math>).</p> Full article ">Figure 11
<p>Variation of input resistance of the prototype system vs. duty ratio and SOC of SC bank when the system is operated under: (<b>a</b>) SC charge recovery mode; (<b>b</b>) SC charge recovery mode at different <math display="inline"><semantics> <mrow> <msub> <mi>V</mi> <mrow> <mi>S</mi> <mi>C</mi> </mrow> </msub> </mrow> </semantics></math> values; (<b>c</b>) Neutral mode; (<b>d</b>) Neutral mode at <math display="inline"><semantics> <mrow> <mn>10.5</mn> <mtext> </mtext> <mi mathvariant="normal">V</mi> <mo>&lt;</mo> <msub> <mi>V</mi> <mrow> <mi>S</mi> <mi>C</mi> </mrow> </msub> <mo>&lt;</mo> <mn>13.5</mn> <mtext> </mtext> <mi mathvariant="normal">V</mi> </mrow> </semantics></math>.</p> Full article ">Figure 12
<p>System prototype: (<b>a</b>) Detailed block diagram; (<b>b</b>) Experimental setup.</p> Full article ">Figure 13
<p>Variation of: (<b>a</b>) Input power from the PV array, power delivered to battery bank, SC bank and DC load, and voltage of the SC bank; (<b>b</b>) Energy input, used energy and system efficiency vs. time when system is operating under Neutral mode while DC load is driven directly by PV array.</p> Full article ">Figure 14
<p>Variation of: (<b>a</b>) Input power from the PV array, power delivered to battery bank, SC bank and DC load, and voltage of the SC bank; (<b>b</b>) Energy input, used energy and system efficiency vs. time when system is operating under Neutral mode while excess current required by DC load is provided by the SC bank.</p> Full article ">Figure 15
<p>Variation of: (<b>a</b>) Input power from the PV array, power delivered to battery bank + DC load and SC bank, and voltage of the SC bank; (<b>b</b>) Energy input, used energy and system efficiency vs. time when system is operating under SC charge recovery mode.</p> Full article ">Figure 16
<p>Variation of: (<b>a</b>) Input power from the PV array, power delivered to battery bank and DC load, and voltage of the SC bank; (<b>b</b>) Energy input, energy delivered to battery bank and charging efficiency; (<b>c</b>) Energy delivered by SC bank, consumed by DC load and efficiency vs. time when system is operating under SC Bypass mode.</p> Full article ">
4 pages, 161 KiB  
Editorial
Electronic Solutions for Artificial Intelligence Healthcare
by Hyeyoung Ko and Jun-Ho Huh
Electronics 2021, 10(19), 2421; https://doi.org/10.3390/electronics10192421 - 4 Oct 2021
Cited by 4 | Viewed by 2687
Abstract
At present, diverse, innovative technology is used in electronics and ubiquitous computing environments [...] Full article
(This article belongs to the Special Issue Electronic Solutions for Artificial Intelligence Healthcare)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">
20 pages, 5425 KiB  
Article
Edge Container for Speech Recognition
by Lukáš Beňo, Rudolf Pribiš and Peter Drahoš
Electronics 2021, 10(19), 2420; https://doi.org/10.3390/electronics10192420 - 4 Oct 2021
Cited by 4 | Viewed by 3549
Abstract
Containerization has been mainly used in pure software solutions, but it is gradually finding its way into the industrial systems. This paper introduces the edge container with artificial intelligence for speech recognition, which performs the voice control function of the actuator as a [...] Read more.
Containerization has been mainly used in pure software solutions, but it is gradually finding its way into the industrial systems. This paper introduces the edge container with artificial intelligence for speech recognition, which performs the voice control function of the actuator as a part of the Human Machine Interface (HMI). This work proposes a procedure for creating voice-controlled applications with modern hardware and software resources. The created architecture integrates well-known digital technologies such as containerization, cloud, edge computing and a commercial voice processing tool. This methodology and architecture enable the actual speech recognition and the voice control on the edge device in the local network, rather than in the cloud, like the majority of recent solutions. The Linux containers are designed to run without any additional configuration and setup by the end user. A simple adaptation of voice commands via configuration file may be considered as an additional contribution of the work. The architecture was verified by experiments with running containers on different devices, such as PC, Tinker Board 2, Raspberry Pi 3 and 4. The proposed solution and the practical experiment show how a voice-controlled system can be created, easily managed and distributed to many devices around the world in a few seconds. All this can be achieved by simple downloading and running two types of ready-made containers without any complex installations. The result of this work is a proven stable (network-independent) solution with data protection and low latency. Full article
(This article belongs to the Special Issue Digitization of Mechatronic Systems Using Modern Information and Communication Technologies)
►▼ Show Figures

Figure 1

Figure 1
<p>Difference between the cloud and edge computing [<a href="#B17-electronics-10-02420" class="html-bibr">17</a>].</p> Full article ">Figure 2
<p>Packing of all dependencies to the container and the following distribution of the container to different machines around the world.</p> Full article ">Figure 3
<p>The entire architecture of the controlling of the actuator using voice commands.</p> Full article ">Figure 4
<p>Sequential steps in the proposed architecture of voice control.</p> Full article ">Figure 5
<p>Module’s image and instances.</p> Full article ">Figure 6
<p>Routes in IoT Edge Hub deployed on the device.</p> Full article ">Figure 7
<p>IoT edge security daemon structure and communication.</p> Full article ">Figure 8
<p>Categories for edge devices.</p> Full article ">Figure 9
<p>IoT Hub in the cloud Azure for management and deployment of IoT edge devices.</p> Full article ">Figure 10
<p>Pseudocode of voice recognition module.</p> Full article ">Figure 11
<p>Three ways to connect many edge devices (Category I) on a local network to one single IoT edge device (Category II) running the Azure Speech Recognition service.</p> Full article ">Figure 12
<p>Possible location for running the Azure Speech Recognition service.</p> Full article ">Figure 13
<p>Visual Studio Code—debugging option. Possibility to see which modules are running on the specific devices. PC, Tinker Board 2, Raspberry Pi 3 and 4 are used in the experiment.</p> Full article ">Figure 14
<p>Comparison of time responses between performed speech recognition in the cloud Azure and IoT edge devices.</p> Full article ">Figure 15
<p>Real implementation of the voice control actuator using Bluetooth headphones for better comfort and freedom.</p> Full article ">
59 pages, 9802 KiB  
Article
Meta-Heuristic Optimization Techniques Used for Maximum Power Point Tracking in Solar PV System
by Preeti Verma, Afroz Alam, Adil Sarwar, Mohd Tariq, Hani Vahedi, Deeksha Gupta, Shafiq Ahmad and Adamali Shah Noor Mohamed
Electronics 2021, 10(19), 2419; https://doi.org/10.3390/electronics10192419 - 3 Oct 2021
Cited by 43 | Viewed by 5756
Abstract
A critical advancement in solar photovoltaic (PV) establishment has led to robust acceleration towards the evolution of new MPPT techniques. The sun-oriented PV framework has a non-linear characteristic in varying climatic conditions, which considerably impact the PV framework yield. Furthermore, the partial shading [...] Read more.
A critical advancement in solar photovoltaic (PV) establishment has led to robust acceleration towards the evolution of new MPPT techniques. The sun-oriented PV framework has a non-linear characteristic in varying climatic conditions, which considerably impact the PV framework yield. Furthermore, the partial shading condition (PSC) causes major problems, such as a drop in the output power yield and multiple peaks in the P–V attribute. Hence, following the global maximum power point (GMPP) under PSC is a demanding problem. Subsequently, different maximum power point tracking (MPPT) strategies have been utilized to improve the yield of a PV framework. However, the disarray lies in choosing the best MPPT technique from the wide algorithms for a particular purpose. Each algorithm has its benefits and drawbacks. Hence, there is a fundamental need for an appropriate audit of the MPPT strategies from time to time. This article presents new works done in the global power point tracking (GMPPT) algorithm field under the PSCs. It sums up different MPPT strategies alongside their working principle, mathematical representation, and flow charts. Moreover, tables depicted in this study briefly organize the significant attributes of algorithms. This work will serve as a reference for sorting an MPPT technique while designing PV systems. Full article
(This article belongs to the Special Issue Power Electronics in Automotive Industry Applications)
►▼ Show Figures

Figure 1

Figure 1
<p>Single diode model of the solar cell.</p> Full article ">Figure 2
<p>The I–V and P–V characteristic curves of the solar cell.</p> Full article ">Figure 3
<p>The influence of varying the temperature on the P–V characteristics of the solar cell.</p> Full article ">Figure 4
<p>The effect of varying the temperature on the I–V characteristics of the solar cell.</p> Full article ">Figure 5
<p>The stimulus of varying insolation on the I–V characteristics of the solar cell.</p> Full article ">Figure 6
<p>The impact of varying insolation on the P–V characteristics of the solar cell.</p> Full article ">Figure 7
<p>I–V curve under different solar insolation.</p> Full article ">Figure 8
<p>P–V curve at the different solar insolation.</p> Full article ">Figure 9
<p>The typical block design of the MPPT implementation for PV systems.</p> Full article ">Figure 10
<p>Flowchart to implement the P&amp;O strategy.</p> Full article ">Figure 11
<p>INC algorithm flowchart.</p> Full article ">Figure 12
<p>Fractional open circuit voltage MPPT strategy block diagram implementation.</p> Full article ">Figure 13
<p>Fractional short-circuit current MPPT strategy schematic diagram.</p> Full article ">Figure 14
<p>Categorization of reviewed meta-heuristic strategies.</p> Full article ">Figure 15
<p>Typical PSO MPPT technique flowchart.</p> Full article ">Figure 16
<p>Flowchart of the ACO strategy.</p> Full article ">Figure 17
<p>Flowchart of the ABC technique.</p> Full article ">Figure 18
<p>The hierarchical sequence of grey wolves.</p> Full article ">Figure 19
<p>Flowchart of the GWO MPPT algorithm.</p> Full article ">Figure 20
<p>Flowchart to implement the EPO strategy.</p> Full article ">Figure 21
<p>Salp chain (or a swarm of salps).</p> Full article ">Figure 22
<p>Flowchart of the SSA MPPT strategy.</p> Full article ">Figure 23
<p>Flowchart of the JA MPPT strategy.</p> Full article ">Figure 24
<p>Flowchart of Cuckoo Search technique.</p> Full article ">Figure 25
<p>Foraging conduct of flying squirrels.</p> Full article ">Figure 26
<p>Flowchart of the FSSO MPPT strategy.</p> Full article ">Figure 27
<p>Inverse-square law of sound intensity.</p> Full article ">Figure 28
<p>Flowchart of Firefly MPPT strategy.</p> Full article ">Figure 29
<p>Reviewed AI strategy categorization.</p> Full article ">Figure 30
<p>FLC MPPT strategy control block diagram.</p> Full article ">Figure 31
<p>Construction of the ANN.</p> Full article ">Figure 32
<p>Flowchart of the GA strategy.</p> Full article ">Figure 33
<p>The flowchart of the DE MPPT strategy.</p> Full article ">Figure 34
<p>Output power, voltage, and current curve of a PV panel with variations in irradiance as per the first set, and duty cycle curve controlled using CS algorithm.</p> Full article ">Figure 35
<p>Output power, voltage, and current curve of a PV panel with variation in irradiance as per the first set, and duty cycle curve controlled using the JA algorithm.</p> Full article ">Figure 36
<p>Output power, voltage, and current curve of a PV panel with variation in irradiance as per the second set, and the duty cycle curve was controlled using the CS algorithm.</p> Full article ">
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-129
Previous Issue
Next Issue
Back to TopTop