[go: up one dir, main page]
Next Issue
Volume 20, April-2
Previous Issue
Volume 20, March-2
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
7.3
3.4
Journals
Sensors
Volume 20
Issue 7
sensors-logo

Journal Menu

► Journal Menu

Journal Browser

► Journal Browser
share announcement

Sensors, Volume 20, Issue 7 (April-1 2020) – 354 articles

Cover Story (view full-size image): Vehicle localization using commercial monocular cameras can improve the robustness against degradation of global navigation satellite system (GNSS), and help with the dynamic update of high-definition (HD) maps using crowdsourcing cameras. Therefore, this paper proposes a vehicle localization method, called monocular localization with vector HD map (MLVHM). The method involves camera-based 6-DOF map-matching that aligns semantic-level geometric features that are robust against occlusion and lighting changes with the vector HD map. Experiments showed that MLVHM can achieve high-precision vehicle localization with a root mean square error (RMSE) of 24 cm within a 60 ms time delay using a vector HD map with a bandwidth of 50 kB/km. Compared with traditional monocular localizing with scale drift and error accumulation, the localization error is reduced by 86%.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:
18 pages, 6356 KiB  
Article
Influence of Mg Doping Levels on the Sensing Properties of SnO2 Films
by Bouteina Bendahmane, Milena Tomić, Nour El Houda Touidjen, Isabel Gràcia, Stella Vallejos and Farida Mansour
Sensors 2020, 20(7), 2158; https://doi.org/10.3390/s20072158 - 10 Apr 2020
Cited by 11 | Viewed by 3981
Abstract
This work presents the effect of magnesium (Mg) doping on the sensing properties of tin dioxide (SnO2) thin films. Mg-doped SnO2 films were prepared via a spray pyrolysis method using three doping concentrations (0.8 at.%, 1.2 at.%, and 1.6 at.%) [...] Read more.
This work presents the effect of magnesium (Mg) doping on the sensing properties of tin dioxide (SnO2) thin films. Mg-doped SnO2 films were prepared via a spray pyrolysis method using three doping concentrations (0.8 at.%, 1.2 at.%, and 1.6 at.%) and the sensing responses were obtained at a comparatively low operating temperature (160 °C) compared to other gas sensitive materials in the literature. The morphological, structural and chemical composition analysis of the doped films show local lattice disorders and a proportional decrease in the average crystallite size as the Mg-doping level increases. These results also indicate an excess of Mg (in the samples prepared with 1.6 at.% of magnesium) which causes the formation of a secondary magnesium oxide phase. The films are tested towards three volatile organic compounds (VOCs), including ethanol, acetone, and toluene. The gas sensing tests show an enhancement of the sensing properties to these vapors as the Mg-doping level rises. This improvement is particularly observed for ethanol and, thus, the gas sensing analysis is focused on this analyte. Results to 80 ppm of ethanol, for instance, show that the response of the 1.6 at.% Mg-doped SnO2 film is four times higher and 90 s faster than that of the 0.8 at.% Mg-doped SnO2 film. This enhancement is attributed to the Mg-incorporation into the SnO2 cell and to the formation of MgO within the film. These two factors maximize the electrical resistance change in the gas adsorption stage, and thus, raise ethanol sensitivity. Full article
(This article belongs to the Special Issue Application of Thin Film Materials in Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Schematic view of the gas sensing measurement system.</p> Full article ">Figure 2
<p>XRD patterns of the Mg-doped SnO<sub>2</sub> thin films, (<b>a</b>) 0.8 at.% magnesium-doped tin dioxide (MTO1); (<b>b</b>) 1.2 at.% magnesium-doped tin dioxide (MTO2); (<b>c</b>) 1.6 at.% magnesium-doped tin dioxide (MTO3).</p> Full article ">Figure 3
<p>SEM micrographs of the Mg-doped SnO<sub>2</sub> thin films, (<b>a</b>) MTO1; (<b>b</b>) MTO2; (<b>c</b>) MTO3.</p> Full article ">Figure 4
<p>XPS survey spectra of the Mg-doped SnO<sub>2</sub> thin films, (<b>a</b>) MTO1; (<b>b</b>) MTO2; (<b>c</b>) MTO3.</p> Full article ">Figure 5
<p>Deconvolution of O 1s (<b>left figure</b>) and Sn 3d (<b>right figure</b>) XPS core level spectra, ((<b>a</b>) and (<b>d</b>)) MTO1; ((<b>b</b>) and (<b>e</b>)) MTO2; ((<b>c</b>) and (<b>f</b>)) MTO3. The circles denote experimental data, colored lines demonstrate the deconvolution of peaks, and the black line corresponds to the sum of peaks fits (envelope).</p> Full article ">Figure 6
<p>XPS of (<b>a</b>) Mg 1s core level spectra and ((<b>b</b>)−(<b>d</b>)) Mg KLL Auger spectra at MTO1, MTO2, and MTO3. The circles denote experimental data, colored lines demonstrate the deconvolution of peaks, and the black line corresponds to the sum of peaks fits (envelope).</p> Full article ">Figure 7
<p>Response (<b>a</b>) and response time (<b>b</b>) towards 80 ppm of ethanol, acetone, and toluene for the non-doped and Mg-doped SnO<sub>2</sub> films.</p> Full article ">Figure 8
<p>Response of the (<b>a</b>) non-doped and Mg-doped SnO<sub>2</sub> films to 80 ppm of ethanol and (<b>b</b>) MTO3 response to 80 ppm of toluene, acetone, and ethanol recorded at 160 °C.</p> Full article ">Figure 9
<p>Dynamic response curves of Mg-doped SnO<sub>2</sub> films at different concentrations of ethanol (<b>left figure</b>) and toluene (<b>right figure</b>).</p> Full article ">Figure 10
<p>Response of Mg-doped SnO<sub>2</sub> films vs. (<b>a</b>) ethanol and (<b>b</b>) toluene concentration at 160 °C.</p> Full article ">Figure 11
<p>Sensitivity of the Mg-doped SnO<sub>2</sub> films to ethanol and toluene.</p> Full article ">Figure 12
<p>Response time to various concentrations of (<b>a</b>) ethanol, and (<b>b</b>) toluene.</p> Full article ">Figure 13
<p>Schematic representation of the gas sensing mechanisms of (<b>a</b>) non-doped and (<b>b</b> and <b>c</b>) Mg-doped SnO<sub>2</sub> films in air (left) and reductive gas (right). E<sub>CB</sub> is the bottom of conduction band; E<sub>F</sub> is the bulk Fermi level; E<sub>VB</sub> is the top of valence band; ΔΦ is the built-in potential barrier; Χn, and Χn<sub>2</sub> are depth of the depletion layer from the surface; Χn<sub>1</sub> is the depth of the accumulation layer from the surface (not to scale).</p> Full article ">
22 pages, 1363 KiB  
Article
Energy Management of Smart Home with Home Appliances, Energy Storage System and Electric Vehicle: A Hierarchical Deep Reinforcement Learning Approach
by Sangyoon Lee and Dae-Hyun Choi
Sensors 2020, 20(7), 2157; https://doi.org/10.3390/s20072157 - 10 Apr 2020
Cited by 90 | Viewed by 9412
Abstract
This paper presents a hierarchical deep reinforcement learning (DRL) method for the scheduling of energy consumptions of smart home appliances and distributed energy resources (DERs) including an energy storage system (ESS) and an electric vehicle (EV). Compared to Q-learning algorithms based on a [...] Read more.
This paper presents a hierarchical deep reinforcement learning (DRL) method for the scheduling of energy consumptions of smart home appliances and distributed energy resources (DERs) including an energy storage system (ESS) and an electric vehicle (EV). Compared to Q-learning algorithms based on a discrete action space, the novelty of the proposed approach is that the energy consumptions of home appliances and DERs are scheduled in a continuous action space using an actor–critic-based DRL method. To this end, a two-level DRL framework is proposed where home appliances are scheduled at the first level according to the consumer’s preferred appliance scheduling and comfort level, while the charging and discharging schedules of ESS and EV are calculated at the second level using the optimal solution from the first level along with the consumer environmental characteristics. A simulation study is performed in a single home with an air conditioner, a washing machine, a rooftop solar photovoltaic system, an ESS, and an EV under a time-of-use pricing. Numerical examples under different weather conditions, weekday/weekend, and driving patterns of the EV confirm the effectiveness of the proposed approach in terms of total cost of electricity, state of energy of the ESS and EV, and consumer preference. Full article
(This article belongs to the Special Issue Internet of Things for Smart Homes II)
►▼ Show Figures

Figure 1

Figure 1
<p>Conceptual architecture of the proposed deep reinforcement learning (DRL)-based home energy management system (HEMS) algorithm.</p> Full article ">Figure 2
<p>Architecture of the neural network model for the proposed actor–critic method.</p> Full article ">Figure 3
<p>Profiles of electricity price and weather. (<b>a</b>) time-of-use (TOU) price; (<b>b</b>) photovoltaic (PV) generation; (<b>c</b>) outdoor temperature.</p> Full article ">Figure 4
<p>DRL-based energy consumption schedule. (<b>a</b>) washing machine (WM); (<b>b</b>) air conditioner (AC); (<b>c</b>) WM+AC+Uncontrollable appliances at Level 1.</p> Full article ">Figure 5
<p>Performance comparison between Case 1 and Case 2. (<b>a</b>) energy consumption of the energy storage system (ESS); (<b>b</b>) state of energy (SOE) of the ESS; (<b>c</b>) energy consumption of the electric vehicle (EV); (<b>d</b>) SOE of the EV; (<b>e</b>) net consumption of household.</p> Full article ">Figure 6
<p>Performance comparison between Case 1 and Case 3. (<b>a</b>) SOE of the ESS; (<b>b</b>) SOE of the EV; (<b>c</b>) net consumption of household.</p> Full article ">Figure 7
<p>Performance comparison between Case 1 and Case 4. (<b>a</b>) SOE of the ESS; (<b>b</b>) SOE of the EV; (<b>c</b>) net consumption of household.</p> Full article ">Figure 8
<p>Comparison of a relative increase of the total electricity bill in Case 3 among the considered three cases.</p> Full article ">Figure 9
<p>Convergence of the total cost. (<b>a</b>) Level 1; (<b>b</b>) Level 2.</p> Full article ">Figure 10
<p>Comparison of the total cost convergence between the proposed two-level DRL approach and the single-level DRL approach.</p> Full article ">Figure 11
<p>Comparison of a relative increase of the total electricity bill in the proposed DRL method using building energy optimization tool (BEopt) and mixed-integer linear programming (MILP) programs for four cases.</p> Full article ">
26 pages, 646 KiB  
Article
P4UIoT: Pay-Per-Piece Patch Update Delivery for IoT Using Gradual Release
by Nachiket Tapas, Yechiav Yitzchak, Francesco Longo, Antonio Puliafito and Asaf Shabtai
Sensors 2020, 20(7), 2156; https://doi.org/10.3390/s20072156 - 10 Apr 2020
Cited by 6 | Viewed by 3672
Abstract
P 4 UIoT—pay-per-piece patch update delivery for IoT using gradual release—introduces a distributed framework for delivering patch updates to IoT devices. The framework facilitates distribution via peer-to-peer delivery networks and incentivizes the distribution operation. The peer-to-peer delivery network reduces load by delegating the [...] Read more.
P 4 UIoT—pay-per-piece patch update delivery for IoT using gradual release—introduces a distributed framework for delivering patch updates to IoT devices. The framework facilitates distribution via peer-to-peer delivery networks and incentivizes the distribution operation. The peer-to-peer delivery network reduces load by delegating the patch distribution to the nodes of the network, thereby protecting against a single point of failure and reducing costs. Distributed file-sharing solutions currently available in the literature are limited to sharing popular files among peers. In contrast, the proposed protocol incentivizes peers to distribute patch updates, which might be relevant only to IoT devices, using a blockchain-based lightning network. A manufacturer/owner named vendor of the IoT device commits a bid on the blockchain, which can be publicly verified by the members of the network. The nodes, called distributors, interested in delivering the patch update, compete among each other to exchange a piece of patch update with cryptocurrency payment. The pay-per-piece payments protocol addresses the problem of misbehavior between IoT devices and distributors as either of them may try to take advantage of the other. The pay-per-piece protocol is a form of a gradual release of a commodity like a patch update, where the commodity can be divided into small pieces and exchanged between the sender and the receiver building trust at each step as the transactions progress into rounds. The permissionless nature of the framework enables the proposal to scale as it incentivizes the participation of individual distributors. Thus, compared to the previous solutions, the proposed framework can scale better without any overhead and with reduced costs. A combination of the Bitcoin lightning network for cryptocurrency incentives with the BitTorrent delivery network is used to present a prototype of the proposed framework. Finally, a financial and scalability evaluation of the proposed framework is presented. Full article
(This article belongs to the Special Issue Blockchain in the Internet of Things: Opportunities, Challenges and Solutions)
►▼ Show Figures

Figure 1

Figure 1
<p>P<math display="inline"><semantics> <msup> <mrow/> <mn>4</mn> </msup> </semantics></math>UIoT High level architecture overview.</p> Full article ">Figure 2
<p>P<math display="inline"><semantics> <msup> <mrow/> <mn>4</mn> </msup> </semantics></math>UIoT sequence diagram.</p> Full article ">Figure 3
<p>The technology stack used in P<math display="inline"><semantics> <msup> <mrow/> <mn>4</mn> </msup> </semantics></math>UIoT implementation.</p> Full article ">Figure 4
<p>P<math display="inline"><semantics> <msup> <mrow/> <mn>4</mn> </msup> </semantics></math>UIoT state transition diagram.</p> Full article ">Figure 5
<p>Distribution latency for patch size (<b>a</b>) 10 kb, (<b>b</b>) 100 kb, and (<b>c</b>) 1 mb patch size.</p> Full article ">
21 pages, 3221 KiB  
Article
Design, Implementation, and Validation of a Piezoelectric Device to Study the Effects of Dynamic Mechanical Stimulation on Cell Proliferation, Migration and Morphology
by Dahiana Mojena-Medina, Marina Martínez-Hernández, Miguel de la Fuente, Guadalupe García-Isla, Julio Posada, José Luis Jorcano and Pablo Acedo
Sensors 2020, 20(7), 2155; https://doi.org/10.3390/s20072155 - 10 Apr 2020
Cited by 7 | Viewed by 4576
Abstract
Cell functions and behavior are regulated not only by soluble (biochemical) signals but also by biophysical and mechanical cues within the cells’ microenvironment. Thanks to the dynamical and complex cell machinery, cells are genuine and effective mechanotransducers translating mechanical stimuli into biochemical signals, [...] Read more.
Cell functions and behavior are regulated not only by soluble (biochemical) signals but also by biophysical and mechanical cues within the cells’ microenvironment. Thanks to the dynamical and complex cell machinery, cells are genuine and effective mechanotransducers translating mechanical stimuli into biochemical signals, which eventually alter multiple aspects of their own homeostasis. Given the dominant and classic biochemical-based views to explain biological processes, it could be challenging to elucidate the key role that mechanical parameters such as vibration, frequency, and force play in biology. Gaining a better understanding of how mechanical stimuli (and their mechanical parameters associated) affect biological outcomes relies partially on the availability of experimental tools that may allow researchers to alter mechanically the cell’s microenvironment and observe cell responses. Here, we introduce a new device to study in vitro responses of cells to dynamic mechanical stimulation using a piezoelectric membrane. Using this device, we can flexibly change the parameters of the dynamic mechanical stimulation (frequency, amplitude, and duration of the stimuli), which increases the possibility to study the cell behavior under different mechanical excitations. We report on the design and implementation of such device and the characterization of its dynamic mechanical properties. By using this device, we have performed a preliminary study on the effect of dynamic mechanical stimulation in a cell monolayer of an epidermal cell line (HaCaT) studying the effects of 1 Hz and 80 Hz excitation frequencies (in the dynamic stimuli) on HaCaT cell migration, proliferation, and morphology. Our preliminary results indicate that the response of HaCaT is dependent on the frequency of stimulation. The device is economic, easily replicated in other laboratories and can support research for a better understanding of mechanisms mediating cellular mechanotransduction. Full article
(This article belongs to the Section Biosensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Schematic representation of the completely mechanical stimulator.</p> Full article ">Figure 2
<p>(<b>a</b>) description of the conformational domains in the experimental stage based on a PVDF actuator; (<b>b</b>) lateral view of the schematic representation shown in (<b>a</b>). (i) the PVDF film is in the bottom of the Petri dish and a PDMS mold achieves the cell culture well. (ii) and (iii) show a cell culture with basal medium and cell cultures confined within the PDMS mold and with the PVDF film as a substrate (under static condition (ii) (0 voltage in the electronics stage) and dynamic condition (iii) (AC voltage in the electronics stage); (<b>c</b>) control set for all the experiments; (<b>d</b>) control set of the piezo device for static conditions; (<b>e</b>) dynamic set for cell culture stimulation at a determined frequency and amplitude (it is noted the external cable connections); (<b>f</b>) handling cell cultures on the piezoelectric devices inside a biosecurity cabin; (<b>g</b>) the piezoelectric system containing the HaCaT cells is placed inside the cell culture incubator with a lateral hole to allow electrical access (a signal generator, power supply, and the driver is observed outside the incubator).</p> Full article ">Figure 3
<p>(<b>a</b>) interferometric layout for analysis of the piezoelectric device; (<b>b</b>) plot showing a linear relation between piezo actuator input voltage and output amplitude for several frequencies; (<b>c</b>) response curve of the PVDF (mechanical deformation) vs. the applied frequency in the device. The curve shows negligible differences between the response of the PVDF functionalized with collagen and without functionalization.</p> Full article ">Figure 4
<p>(<b>a</b>) curve of the cytocompatibility assay for eight experiments repetition (<span class="html-italic">n</span> = 28 control samples (C) and <span class="html-italic">n</span> = 25 control material samples (CM)). No significance (<span class="html-italic">p</span> &lt; 0.05) was found. (<b>b</b>) curves for experiment 80 Hz applied. (<span class="html-italic">n</span> = 9 control samples (C), <span class="html-italic">n</span> = 9 control static samples (CM) and <span class="html-italic">n</span> = 8 dynamic samples at 80 Hz frequency) ** Significance (<span class="html-italic">p</span> &lt; 0.05) in days 3 and 6. (<b>c</b>) curves for two repetitions of experiment 1 Hz applied (<span class="html-italic">n</span> = 12 control samples (C), <span class="html-italic">n</span> = 12 control static samples (CM) and <span class="html-italic">n</span> = 12 dynamic samples at 1 Hz frequency) *** Significance (<span class="html-italic">p</span> &lt; 0.05) in days 6 and 8; (<b>d</b>) area decrease in scratch assay due to migration process in control devices (C). Value in the curve at 30 h represents that the healing closed completely at the observation, but not the closing time.; (<b>e</b>) area decrease in scratch assay due to migration process in control. devices with piezoelectric material without stimulation (CM); (<b>f</b>) area decrease in scratch assay due to migration process in piezoelectric device with 1 Hz frequency stimulation (FR 1 Hz); (<b>g</b>) area decrease in scratch assay due to migration process in piezoelectric devices with 80 Hz frequency stimulation (FR 80 Hz); (<b>h</b>) wound progression in a control (C) device, and experimental groups with 1 Hz and 80 Hz stimulation frequency at different times.</p> Full article ">Figure 4 Cont.
<p>(<b>a</b>) curve of the cytocompatibility assay for eight experiments repetition (<span class="html-italic">n</span> = 28 control samples (C) and <span class="html-italic">n</span> = 25 control material samples (CM)). No significance (<span class="html-italic">p</span> &lt; 0.05) was found. (<b>b</b>) curves for experiment 80 Hz applied. (<span class="html-italic">n</span> = 9 control samples (C), <span class="html-italic">n</span> = 9 control static samples (CM) and <span class="html-italic">n</span> = 8 dynamic samples at 80 Hz frequency) ** Significance (<span class="html-italic">p</span> &lt; 0.05) in days 3 and 6. (<b>c</b>) curves for two repetitions of experiment 1 Hz applied (<span class="html-italic">n</span> = 12 control samples (C), <span class="html-italic">n</span> = 12 control static samples (CM) and <span class="html-italic">n</span> = 12 dynamic samples at 1 Hz frequency) *** Significance (<span class="html-italic">p</span> &lt; 0.05) in days 6 and 8; (<b>d</b>) area decrease in scratch assay due to migration process in control devices (C). Value in the curve at 30 h represents that the healing closed completely at the observation, but not the closing time.; (<b>e</b>) area decrease in scratch assay due to migration process in control. devices with piezoelectric material without stimulation (CM); (<b>f</b>) area decrease in scratch assay due to migration process in piezoelectric device with 1 Hz frequency stimulation (FR 1 Hz); (<b>g</b>) area decrease in scratch assay due to migration process in piezoelectric devices with 80 Hz frequency stimulation (FR 80 Hz); (<b>h</b>) wound progression in a control (C) device, and experimental groups with 1 Hz and 80 Hz stimulation frequency at different times.</p> Full article ">Figure 5
<p>Electron microscopy images of samples under different conditions. The micrographics show difference in surface area and projections after stimulation. Non-significance is observed between controls (<b>A</b>) and static controls (<b>B</b>) at 48 h of observation. The morphology is clearly different between control groups (control (<b>A</b>) and control material (<b>B</b>)) with 1 Hz samples (<b>C</b>) and 80 Hz samples (<b>D</b>). The 1 Hz stimulated group cells show cytoplasmic projections and condensed nucleus at 48 h. The 80 Hz stimulated group shows cells with rounded shape at 48 h. Length scale: 20 µm.</p> Full article ">Figure 6
<p>Immunostaining images of f-actin and DAPI at 48 h after continuous stimulation. Changes in actin distribution is observed between control samples (control (<b>A</b>) and control material (<b>B</b>)) with 1 Hz stimulated samples (<b>C</b>). In the 80 Hz group (<b>D</b>), significant changes compared with control are not observed. Nucleus staining does not show significant difference in size in control samples compared with 80 Hz stimulated samples. In the 1 Hz group (<b>C</b>), in some fields, the nucleus sizes were oval and elongated consistent with directions of the stress fiber.</p> Full article ">Figure 7
<p>Average size of the diameter of nucleus normalized to control samples at 48 h after continuous mechanical stimulation (number of nuclei analyzed &gt; 300 for both control and experimental groups).</p> Full article ">
12 pages, 4122 KiB  
Article
Inlet Effect Caused by Multichannel Structure for Molecular Electronic Transducer Based on a Turbulent-Laminar Flow Model
by Qiuzhan Zhou, Qi He, Yuzhu Chen and Xue Bao
Sensors 2020, 20(7), 2154; https://doi.org/10.3390/s20072154 - 10 Apr 2020
Viewed by 2547
Abstract
The actual fluid form of an electrolyte in a molecular electronic converter is an important factor that causes a decrease in the accuracy of a molecular electronic transducer (MET) liquid motion sensor. To study the actual fluid morphology of an inertial electrolyte in [...] Read more.
The actual fluid form of an electrolyte in a molecular electronic converter is an important factor that causes a decrease in the accuracy of a molecular electronic transducer (MET) liquid motion sensor. To study the actual fluid morphology of an inertial electrolyte in molecular electron transducers, an inlet effect is defined according to the fluid morphology of turbulent-laminar flow, and a numerical simulation model of turbulent-laminar flow is proposed. Based on the turbulent-laminar flow model, this paper studies the variation of the inlet effect intensity when the thickness of the outermost insulating layer is 50 µm and 100 µm, respectively. Meanwhile, the changes of the inlet effect intensity and the error rate of central axial velocity field are also analyzed when the input signal intensity is different. Through the numerical experiment, it verifies that the thickness of the outermost insulating layer and the amplitude of the input signal are two important factors which can affect the inlet effect intensity and also the accuracy of the MET. Therefore, this study can provide a theoretical basis for the quantitative study on the performance optimization of a MET liquid sensor. Full article
(This article belongs to the Special Issue MET Angular and Linear Motion Seismic Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Integrated graph of the reaction cavity.</p> Full article ">Figure 2
<p>Schematic of a single channel model.</p> Full article ">Figure 3
<p>Electrolyte motion in reaction cavity at 5 s under b <math display="inline"><semantics> <mrow> <mi mathvariant="normal">a</mi> <mo>=</mo> <mn>0.01</mn> <mi mathvariant="normal">s</mi> <mi mathvariant="normal">i</mi> <mi mathvariant="normal">n</mi> <mo stretchy="false">(</mo> <mi>π</mi> <mi mathvariant="normal">t</mi> <mo stretchy="false">)</mo> <mo> </mo> <mi mathvariant="normal">m</mi> <mo>/</mo> <msup> <mi mathvariant="normal">s</mi> <mn>2</mn> </msup> </mrow> </semantics></math>.</p> Full article ">Figure 4
<p>Radial velocity field distribution of the xz plane in the sensing element under (<b>a</b>) turbulent-laminar flow model and (<b>b</b>) laminar flow model.</p> Full article ">Figure 5
<p>Axial velocity field distribution of the xy plane in the sensing element when the thickness of the outermost insulation layers is (<b>a</b>) 100 µm and (<b>b</b>) 50 µm.</p> Full article ">Figure 6
<p>Axial velocity field distribution of the yz plane in the sensing element under sinusoidal excitation signal in the quarter cycle.</p> Full article ">Figure 7
<p>Axial velocity field distribution of the xy plane in the sensing element under sinusoidal excitation signal in the quarter cycle.</p> Full article ">Figure 8
<p>Error rate of central axis velocity in sensing element.</p> Full article ">
18 pages, 710 KiB  
Article
Wireless Body Area Network (WBAN)-Based Telemedicine for Emergency Care
by Latha R and Vetrivelan P
Sensors 2020, 20(7), 2153; https://doi.org/10.3390/s20072153 - 10 Apr 2020
Cited by 10 | Viewed by 5452
Abstract
This paper is a collection of telemedicine techniques used by wireless body area networks (WBANs) for emergency conditions. Furthermore, Bayes’ theorem is proposed for predicting emergency conditions. With prior knowledge, the posterior probability can be found along with the observed evidence. The probability [...] Read more.
This paper is a collection of telemedicine techniques used by wireless body area networks (WBANs) for emergency conditions. Furthermore, Bayes’ theorem is proposed for predicting emergency conditions. With prior knowledge, the posterior probability can be found along with the observed evidence. The probability of sending emergency messages can be determined using Bayes’ theorem with the likelihood evidence. It can be viewed as medical decision-making, since diagnosis conditions such as emergency monitoring, delay-sensitive monitoring, and general monitoring are analyzed with its network characteristics, including data rate, cost, packet loss rate, latency, and jitter. This paper explains the network model with 16 variables, with one describing immediate consultation, as well as another three describing emergency monitoring, delay-sensitive monitoring, and general monitoring. The remaining 12 variables are observations related to latency, cost, packet loss rate, data rate, and jitter. Full article
(This article belongs to the Special Issue Wireless Body Area Networks for Health Applications)
►▼ Show Figures

Figure 1

Figure 1
<p>Schematic view of the methodology.</p> Full article ">Figure 2
<p>Bayes network model for telemedicine.</p> Full article ">
12 pages, 2157 KiB  
Article
Biosensors Platform Based on Chitosan/AuNPs/Phthalocyanine Composite Films for the Electrochemical Detection of Catechol. The Role of the Surface Structure
by Coral Salvo-Comino, Alfonso González-Gil, Javier Rodriguez-Valentin, Celia Garcia-Hernandez, Fernando Martin-Pedrosa, Cristina Garcia-Cabezon and Maria Luz Rodriguez-Mendez
Sensors 2020, 20(7), 2152; https://doi.org/10.3390/s20072152 - 10 Apr 2020
Cited by 29 | Viewed by 3634
Abstract
Biosensor platforms consisting of layer by layer films combining materials with different functionalities have been developed and used to obtain improved catechol biosensors. Tyrosinase (Tyr) or laccase (Lac) were deposited onto LbL films formed by layers of a cationic linker (chitosan, CHI) alternating [...] Read more.
Biosensor platforms consisting of layer by layer films combining materials with different functionalities have been developed and used to obtain improved catechol biosensors. Tyrosinase (Tyr) or laccase (Lac) were deposited onto LbL films formed by layers of a cationic linker (chitosan, CHI) alternating with layers of anionic electrocatalytic materials (sulfonated copper phthalocyanine, CuPcS or gold nanoparticles, AuNP). Films with different layer structures were successfully formed. Characterization of surface roughness and porosity was carried out using AFM. Electrochemical responses towards catechol showed that the LbL composites efficiently improved the electron transfer path between Tyr or Lac and the electrode surface, producing an increase in the intensity over the response in the absence of the LbL platform. LbL structures with higher roughness and pore size facilitated the diffusion of catechol, resulting in lower LODs. The [(CHI)-(AuNP)-(CHI)-(CuPcS)]2-Tyr showed an LOD of 8.55∙10−4 μM, which was one order of magnitude lower than the 9.55·10−3 µM obtained with [(CHI)-(CuPcS)-(CHI)-(AuNP)]2-Tyr, and two orders of magnitude lower than the obtained with other nanostructured platforms. It can be concluded that the combination of adequate materials with complementary activity and the control of the structure of the platform is an excellent strategy to obtain biosensors with improved performances. Full article
(This article belongs to the Section Biosensors)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>UV-Vis absorption spectra of LbL films of (<b>a</b>) [(CHI)-(CuPcS)-(CHI)-(AuNPs)]<sub>2</sub> (dotted line) and [(CHI)-(AuNPs)-(CHI)-(CuPcS)]<sub>2</sub> (solid line); (<b>b</b>) [(CHI)-(CuPcS)-(CHI)-(AuNPs)]<sub>2</sub>-Tyr. Inset shows a magnification of the band associated to the enzyme.</p> Full article ">Figure 2
<p>AFM topographic images of (<b>a</b>) [(CHI)-(CuPcS)-(CHI)-(AuNPs)]<sub>2</sub>, (<b>b</b>) [(CHI)-(AuNPs)-(CHI)-(CuPcS)]<sub>2</sub>; (<b>c</b>) [(CHI)-(CuPcS)-(CHI)-(AuNPs)]<sub>2</sub>–Tyr and (<b>d</b>) [(CHI)-(AuNPs)-(CHI)-(CuPcS)]<sub>2</sub>–Tyr.</p> Full article ">Figure 3
<p>CV curves registered in catechol 10<sup>−4</sup> mol∙L<sup>−1</sup> in 0.01 M phosphate buffer pH 7 as electrolyte at [(CHI)-(AuNPs)]<sub>2</sub> (dashed line), [(CHI)-(CuPcS)]<sub>2</sub> (solid grey line), [(CHI)-(CuPcS)-(CHI)-(AuNPs)]<sub>2</sub> (solid line), (CHI)-(AuNPs)-(CHI)-(CuPcS)]<sub>2</sub> (dotted line) electrodes at a scan rate of 0.1 V∙s<sup>−1</sup> . The inset shows a magnification of the cathodic peak.</p> Full article ">Figure 4
<p>Voltammetric response of catechol 10-4 M in 0.01 M phosphate buffer pH 7 at an (<b>a</b>) ITO-Tyr (solid line), [(CHI)-(CuPcS)-(CHI)-( AuNPs)]<sub>2</sub>-Tyr (dashed line), [(CHI)-(AuNPs)-(CHI)-(CuPcS)]<sub>2</sub>-Tyr (dotted line); and (<b>b</b>) ITO-Lac (solid line), [(CHI)-(CuPcS)-(CHI)-(AuNPs)]<sub>2</sub>-Lac (dashed line), [(CHI)-(AuNPs)-(CHI)-(CuPcS)]<sub>2</sub>-Lac (dotted line).</p> Full article ">Figure 5
<p>Calibration curves of the amperometric response for increasing the concentration of catechol at (<b>a</b>) [(CHI)-(CuPcS)-(CHI)-(AuNPs)]<sub>2</sub>-Tyr; (<b>b</b>) [(CHI)-(AuNPs)-(CHI)-(CuPcS)]<sub>2</sub>-Tyr; (<b>c</b>) [(CHI)-(CuPcS)-(CHI)-(AuNPs)]<sub>2</sub>-Lac; and (<b>d</b>) [(CHI)-(AuNPs)-(CHI)-(CuPcS)]<sub>2</sub>-Lac. The insets show the amperometric responses obtained for each biosensor.</p> Full article ">
16 pages, 1700 KiB  
Article
Specific Loss Power of Co/Li/Zn-Mixed Ferrite Powders for Magnetic Hyperthermia
by Gabriele Barrera, Marco Coisson, Federica Celegato, Luca Martino, Priyanka Tiwari, Roshni Verma, Shashank N. Kane, Frédéric Mazaleyrat and Paola Tiberto
Sensors 2020, 20(7), 2151; https://doi.org/10.3390/s20072151 - 10 Apr 2020
Cited by 16 | Viewed by 3476
Abstract
An important research effort on the design of the magnetic particles is increasingly required to optimize the heat generation in biomedical applications, such as magnetic hyperthermia and heat-assisted drug release, considering the severe restrictions for the human body’s exposure to an alternating magnetic [...] Read more.
An important research effort on the design of the magnetic particles is increasingly required to optimize the heat generation in biomedical applications, such as magnetic hyperthermia and heat-assisted drug release, considering the severe restrictions for the human body’s exposure to an alternating magnetic field. Magnetic nanoparticles, considered in a broad sense as passive sensors, show the ability to detect an alternating magnetic field and to transduce it into a localized increase of temperature. In this context, the high biocompatibility, easy synthesis procedure and easily tunable magnetic properties of ferrite powders make them ideal candidates. In particular, the tailoring of their chemical composition and cation distribution allows the control of their magnetic properties, tuning them towards the strict demands of these heat-assisted biomedical applications. In this work, Co0.76Zn0.24Fe2O4, Li0.375Zn0.25Fe2.375O4 and ZnFe2O4 mixed-structure ferrite powders were synthesized in a ‘dry gel’ form by a sol-gel auto-combustion method. Their microstructural properties and cation distribution were obtained by X-ray diffraction characterization. Static and dynamic magnetic measurements were performed revealing the connection between the cation distribution and magnetic behavior. Particular attention was focused on the effect of Co2+ and Li+ ions on the magnetic properties at a magnetic field amplitude and the frequency values according to the practical demands of heat-assisted biomedical applications. In this context, the specific loss power (SLP) values were evaluated by ac-hysteresis losses and thermometric measurements at selected values of the dynamic magnetic fields. Full article
(This article belongs to the Special Issue Biosensors with Magnetic Nanocomponents)
►▼ Show Figures

Figure 1

Figure 1
<p>Rietveld-refined XRD patterns of: (<b>a</b>) dry gel Co<sub>0.76</sub>Zn<sub>0.24</sub>Fe<sub>2</sub>O<sub>4</sub>, (<b>b</b>) Li<sub>0.375</sub>Zn<sub>0.25</sub>Fe<sub>2.375</sub>O<sub>4</sub> annealed at 450 °C/3 h, (<b>c</b>) ZnFe<sub>2</sub>O<sub>4</sub> annealed at 450 °C/3 h.</p> Full article ">Figure 2
<p>(<b>a</b>) Room-temperature major dc-hysteresis loops of all the studied samples; (<b>b</b>) dc-hysteresis loops areas as a function of the vertex field for all the studied ferrites.</p> Full article ">Figure 3
<p>(<b>a</b>) Room-temperature minor ac-hysteresis loops of all the studied samples (f = 69 kHz and <span class="html-italic">H<sub>v</sub></span> = 37 kA/m); (<b>b</b>) specific loss power (SLP) values for all the samples as a function of the vertex field obtained by the ac-hysteresis loops areas.</p> Full article ">Figure 4
<p>Time dependence of the temperature of the magnetic solution containing LiZn-ferrite powder under an applied field of 40 kA/m at 100 kHz. Black symbols: experimental data. Green line: best fit the by theoretical model.</p> Full article ">Figure 5
<p>SLP values for the LiZn sample as a function of the vertex field obtained by: ac-hysteresis loops areas (full red dots) at the operation frequency of 69 kHz and the thermometric measurements (empty red dots) at the operation frequency of 100 kHz.</p> Full article ">
19 pages, 11662 KiB  
Article
Bayesian Inversion for Geoacoustic Parameters in Shallow Sea
by Guangxue Zheng, Hanhao Zhu, Xiaohan Wang, Sartaj Khan, Nansong Li and Yangyang Xue
Sensors 2020, 20(7), 2150; https://doi.org/10.3390/s20072150 - 10 Apr 2020
Cited by 13 | Viewed by 3029
Abstract
Geoacoustic parameter inversion is a crucial issue in underwater acoustic research for shallow sea environments and has increasingly become popular in the recent past. This paper investigates the geoacoustic parameters in a shallow sea environment using a single-receiver geoacoustic inversion method based on [...] Read more.
Geoacoustic parameter inversion is a crucial issue in underwater acoustic research for shallow sea environments and has increasingly become popular in the recent past. This paper investigates the geoacoustic parameters in a shallow sea environment using a single-receiver geoacoustic inversion method based on Bayesian theory. In this context, the seabed is regarded as an elastic medium, the acoustic pressure at different positions under low-frequency is chosen as the study object, and the theoretical prediction value of the acoustic pressure is described by the Fast Field Method (FFM). The cost function between the measured and modeled acoustic fields is established under the assumption of Gaussian data errors using Bayesian methodology. The Bayesian inversion method enables the inference of the seabed geoacoustic parameters from the experimental data, including the optimal estimates of these parameters, such as density, sound speed and sound speed attenuation, and quantitative uncertainty estimates. The optimization is carried out by simulated annealing (SA), and the Posterior Probability Density (PPD) is given as the inversion result based on the Gibbs Sampler (GS) algorithm. Inversion results of the experimental data are in good agreement with both measured values and estimates from Genetic Algorithm (GA) inversion result in the same environment. Furthermore, the results also indicate that the sound speed and density in the seabed have fewer uncertainties and are more sensitive to acoustic pressure than the sound speed attenuation. The sea noise could increase the variance of PPD, which has less influence on the sensitive parameters. The mean value of PPD could still reflect the true values of geoacoustic parameters in simulation. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>(<b>a</b>) On-site measurement, (<b>b</b>) Geoacoustic parameters inversion.</p> Full article ">Figure 2
<p>A layered shallow sea waveguide model.</p> Full article ">Figure 3
<p>The diagram of the research.</p> Full article ">Figure 4
<p>The influence of the five geoacoustic parameters on acoustic propagation, (<b>a</b>)-(<b>e</b>) corresponds to the <span class="html-italic">c<sub>p,</sub> c<sub>s,</sub> ρ<sub>b,</sub></span> <span class="html-italic">α<sub>p</sub></span> and <span class="html-italic">α<sub>s</sub></span> respectively, (<b>f</b>) reveals the comparison of TLs’ anomalies when the value of each parameter was changed.</p> Full article ">Figure 5
<p>The sensitivity of the cost function (<span class="html-italic">E</span>(<b>m</b>)) for the five geoacoustic parameters. (<b>a</b>) to (<b>e</b>)corresponds to the <span class="html-italic">c<sub>p,</sub> c<sub>s,</sub> ρ<sub>b,</sub></span> <span class="html-italic">α<sub>p</sub></span>, and <span class="html-italic">α<sub>s</sub></span>, respectively. (<b>f</b>) corresponds to the comparison of five parameter’s influence on <span class="html-italic">E</span>(<b>m</b>).</p> Full article ">Figure 6
<p>Five geo-acoustic parameters’ Posterior Probability Density (PPD) in the noiseless sea environment. The red lines mean the true value of each parameter in simulation. The green segments represent the mean value of the inversion results and their variance.</p> Full article ">Figure 6 Cont.
<p>Five geo-acoustic parameters’ Posterior Probability Density (PPD) in the noiseless sea environment. The red lines mean the true value of each parameter in simulation. The green segments represent the mean value of the inversion results and their variance.</p> Full article ">Figure 7
<p>Five geo-acoustic parameters’ PPD in the noisy sea environment. The red lines mean the true value of each parameter in simulation. The green segments represent the mean value of the inversion results and their variance.</p> Full article ">Figure 8
<p>2D marginal PPDs between different parameters in the noiseless sea environment. White dashed lines mark the true values of each parameter in the simulation.</p> Full article ">Figure 9
<p>2D marginal PPDs between different parameters in the noisy sea environment. White dashed lines mark the true values of each parameter in the simulation.</p> Full article ">Figure 10
<p>The verification of simulation results. The solid blue line means the Transmission Loss (TL) measured in the simulation and the dashed red line represents the TL simulated by inversion result. (<b>a</b>) the comparison of TL in the noiseless sea environment; (<b>b</b>) the comparison of TL in the noisy sea environment.</p> Full article ">Figure 11
<p>Transmission losses under the four scaling conditions. The solid blue line represents the frequency 150 × 0.1 Hz, the dashed red line represents the frequency 150 × 1 Hz, the dotted black line represents the frequency 150 × 10 Hz, and the dashed dotted green line represents the frequency 150 × 100 Hz.</p> Full article ">Figure 12
<p>(<b>a</b>) Experimental measurement system, (<b>b</b>) The movable micro-worktable for measurement equipment.</p> Full article ">Figure 13
<p>(<b>a</b>) The TL measured in the experiment; (<b>b</b>) The arrival time of the signal. The red line represents the arrival time of the direct signal, the dashed red line means the arrival time of surface reflection signal, and the dotted red line means the arrival time of the bottom reflection signal.</p> Full article ">Figure 14
<p>Five geo-acoustic parameters’ PPD. Red lines mean the inversion result of each parameter by GA. Green segments represent the mean value of the inversion results and their variance.</p> Full article ">Figure 15
<p>2D marginal PPDs between different parameters. White dashed lines mark the Genetic Algorithm (GA) results of each parameter.</p> Full article ">Figure 16
<p>The verification approaches, (<b>a</b>) solid blue line means the TL measured in experiment and the dashed red line represents the TL simulated by inversion result; (<b>b</b>) solid blue line means the normalized signal amplitude in the time domain at the 470th reception point (one meter from the starting point), and the dashed red line represents the normalized signal amplitude in the time domain simulated by inversion result. The after off effects of the transmitter are highlighted by arrows.</p> Full article ">
15 pages, 4644 KiB  
Article
Monitoring of Interfacial Debonding of Concrete Filled Pultrusion-GFRP Tubular Column Based on Piezoelectric Smart Aggregate and Wavelet Analysis
by Wenwei Yang, Xia Yang and Shuntao Li
Sensors 2020, 20(7), 2149; https://doi.org/10.3390/s20072149 - 10 Apr 2020
Cited by 14 | Viewed by 2553
Abstract
The concrete filled pultrusion-GFRP (Glass Fiber Reinforced Polymer) tubular column (CFGC) is popular in hydraulic structures or regions with poor environmental conditions due to its excellent corrosion resistance. Considering the influence of concrete hydration heat, shrinkage, and creep, debonding may occur in the [...] Read more.
The concrete filled pultrusion-GFRP (Glass Fiber Reinforced Polymer) tubular column (CFGC) is popular in hydraulic structures or regions with poor environmental conditions due to its excellent corrosion resistance. Considering the influence of concrete hydration heat, shrinkage, and creep, debonding may occur in the interface between the GFRP tube and the concrete, which will greatly reduce the cooperation of the GFRP tube and concrete, and will weaken the mechanical property of CFGC. This paper introduces an active monitoring method based on the piezoelectric transducer. In the active sensing approach, the smart aggregate (SA) embedded in the concrete acted as a driver to transmit a modulated stress wave, and the PZT (Lead Zirconate Titanate) patches attached on the outer surface of CFGC serve as sensors to receive signals and transfer them to the computer for saving. Two groups of experiments were designed with the different debonding areas and thicknesses. The artificial damage of CFGC was identified and located by comparing the value of the delay under pulse excitation and the difference of wavelet-based energy under sweep excitation, and the damage indexes were defined based on the wavelet packet energy to quantify the level of the interface damage. The results showed that the debonding damage area of CFGC can be identified effectively through the active monitoring method, and the damage index can accurately reflect the damage level of the interface of GFRP tube and concrete. Therefore, this method can be used to identify and evaluate the interface debonding of CFGC in real time. In addition, if the method can be combined with remote sensing technology, it can be used as a real-time remote sensing monitoring technology to provide a solution for interface health monitoring of CFGC. Full article
(This article belongs to the Section Intelligent Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Active piezoelectric transducer detection system based on wave analysis.</p> Full article ">Figure 2
<p>Wavelet packet decomposition at level 3.</p> Full article ">Figure 3
<p>The position of SA embedded in concrete and the structural composition of SA.</p> Full article ">Figure 4
<p>PZT patches pasted on the column.</p> Full article ">Figure 5
<p>The arrangement of SA1 and the damage position of specimen A1. (<b>a</b>) Specimen A1; (<b>b</b>) Sectional view of A1; (<b>c</b>) Unfolded view of specimen A1 along the perimeter of the section.</p> Full article ">Figure 6
<p>The arrangement of SA2 and the damage position of specimen A2. (<b>a</b>) Specimen A2; (<b>b</b>) Sectional view of A2; (<b>c</b>) Unfolded view of specimen A2 along the perimeter of the section.</p> Full article ">Figure 7
<p>Composition of active detection system based on piezoelectric transducers.</p> Full article ">Figure 8
<p>Waveform of specimen A1 under sweep excitation.</p> Full article ">Figure 9
<p>Wavelet packet energy and damage index under four cases of specimen A1. (<b>a</b>) Wavelet packet energy, (<b>b</b>) Damage index DI.</p> Full article ">Figure 10
<p>Waveform of specimen A1 under pulse excitation.</p> Full article ">Figure 11
<p>Test result of specimen A1 under pulse excitation. (<b>a</b>) Arrival time of pulse wave; (<b>b</b>) Relationship between debonding area and pulse delay.</p> Full article ">Figure 12
<p>Waveform of specimen A2 under sweep excitation.</p> Full article ">Figure 13
<p>Wavelet packet energy and damage index under four cases of specimen A2. (<b>a</b>) Wavelet packet energy; (<b>b</b>) Damage index DI.</p> Full article ">Figure 14
<p>Waveform of specimen A2 under pulse excitation.</p> Full article ">Figure 15
<p>Test results of specimen A2 under pulse excitation. (<b>a</b>) Arrival time of pulse wave; (<b>b</b>) Relationship between debonding thickness and pulse delay.</p> Full article ">
19 pages, 3457 KiB  
Review
Parametric Acoustic Array and Its Application in Underwater Acoustic Engineering
by Hanyun Zhou, S.H. Huang and Wei Li
Sensors 2020, 20(7), 2148; https://doi.org/10.3390/s20072148 - 10 Apr 2020
Cited by 23 | Viewed by 6738
Abstract
As a sound transmitting device based on the nonlinear acoustic theory, parametric acoustic array (PAA) is able to generate high directivity and low frequency broadband signals with a small aperture transducer. Due to its predominant technical advantages, PAA has been widely used in [...] Read more.
As a sound transmitting device based on the nonlinear acoustic theory, parametric acoustic array (PAA) is able to generate high directivity and low frequency broadband signals with a small aperture transducer. Due to its predominant technical advantages, PAA has been widely used in a variety of application scenarios of underwater acoustic engineering, such as sub-bottom profile measurement, underwater acoustic communication, and detection of buried targets. In this review paper, we examine some of the important advances in the PAA since it was first proposed by Westervelt in 1963. These advances include theoretical modelling for the PAA, signal processing methods, design considerations and implementation issues, and applications of the PAA in underwater acoustic engineering. Moreover, we highlight some technical challenges which impede further development of the PAA, and correspondingly give a glimpse on its possible extension in the future. This article provides a comprehensive overview of some important works of the PAA and serves as a quick tutorial reference to readers who are interested to further explore and extend this technology, and bring this technology to other application areas. Full article
(This article belongs to the Special Issue Sensing and Probing by Acoustic Fields: Principles, Operation and Technology of Piezoelectric Sensors and Transducers)
►▼ Show Figures

Figure 1

Figure 1
<p>Generation of the secondary beam through the PAA.</p> Full article ">Figure 2
<p>Comparison of the original linear frequency modulated (LFM) and self-demodulated LFM in time domain.</p> Full article ">Figure 3
<p>Comparison of the original LFM and self-demodulated LFM in frequency domain.</p> Full article ">Figure 4
<p>Block diagram of the PAA system.</p> Full article ">Figure 5
<p>Block diagram of DSBAM.</p> Full article ">Figure 6
<p>Block diagram of SQRAM.</p> Full article ">Figure 7
<p>Block diagram of SSBAM.</p> Full article ">Figure 8
<p>System configuration of a parametric array sub-bottom profiler.</p> Full article ">
17 pages, 1680 KiB  
Article
A Novel Hybrid Algorithm Based on Grey Wolf Optimizer and Fireworks Algorithm
by Zhihang Yue, Sen Zhang and Wendong Xiao
Sensors 2020, 20(7), 2147; https://doi.org/10.3390/s20072147 - 10 Apr 2020
Cited by 36 | Viewed by 4658
Abstract
Grey wolf optimizer (GWO) is a meta-heuristic algorithm inspired by the hierarchy of grey wolves (Canis lupus). Fireworks algorithm (FWA) is a nature-inspired optimization method mimicking the explosion process of fireworks for optimization problems. Both of them have a strong optimal search capability. [...] Read more.
Grey wolf optimizer (GWO) is a meta-heuristic algorithm inspired by the hierarchy of grey wolves (Canis lupus). Fireworks algorithm (FWA) is a nature-inspired optimization method mimicking the explosion process of fireworks for optimization problems. Both of them have a strong optimal search capability. However, in some cases, GWO converges to the local optimum and FWA converges slowly. In this paper, a new hybrid algorithm (named as FWGWO) is proposed, which fuses the advantages of these two algorithms to achieve global optima effectively. The proposed algorithm combines the exploration ability of the fireworks algorithm with the exploitation ability of the grey wolf optimizer (GWO) by setting a balance coefficient. In order to test the competence of the proposed hybrid FWGWO, 16 well-known benchmark functions having a wide range of dimensions and varied complexities are used in this paper. The results of the proposed FWGWO are compared to nine other algorithms, including the standard FWA, the native GWO, enhanced grey wolf optimizer (EGWO), and augmented grey wolf optimizer (AGWO). The experimental results show that the FWGWO effectively improves the global optimal search capability and convergence speed of the GWO and FWA. Full article
(This article belongs to the Special Issue Intelligent Sensors for Positioning, Tracking, Monitoring, Navigation and Smart Sensing in Smart Cities)
►▼ Show Figures

Figure 1

Figure 1
<p>Position updating in GWO.</p> Full article ">Figure 2
<p>Adaptive balance coefficient.</p> Full article ">Figure 3
<p>Convergence curves of the unimodal functions.</p> Full article ">Figure 4
<p>Convergence curves of the multimodal functions.</p> Full article ">
15 pages, 8697 KiB  
Article
Tensor-Based Emotional Category Classification via Visual Attention-Based Heterogeneous CNN Feature Fusion
by Yuya Moroto, Keisuke Maeda, Takahiro Ogawa and Miki Haseyama
Sensors 2020, 20(7), 2146; https://doi.org/10.3390/s20072146 - 10 Apr 2020
Cited by 2 | Viewed by 3045
Abstract
The paper proposes a method of visual attention-based emotion classification through eye gaze analysis. Concretely, tensor-based emotional category classification via visual attention-based heterogeneous convolutional neural network (CNN) feature fusion is proposed. Based on the relationship between human emotions and changes in visual attention [...] Read more.
The paper proposes a method of visual attention-based emotion classification through eye gaze analysis. Concretely, tensor-based emotional category classification via visual attention-based heterogeneous convolutional neural network (CNN) feature fusion is proposed. Based on the relationship between human emotions and changes in visual attention with time, the proposed method performs new gaze-based image representation that is suitable for reflecting the characteristics of the changes in visual attention with time. Furthermore, since emotions evoked in humans are closely related to objects in images, our method uses a CNN model to obtain CNN features that can represent their characteristics. For improving the representation ability to the emotional categories, we extract multiple CNN features from our novel gaze-based image representation and enable their fusion by constructing a novel tensor consisting of these CNN features. Thus, this tensor construction realizes the visual attention-based heterogeneous CNN feature fusion. This is the main contribution of this paper. Finally, by applying logistic tensor regression with general tensor discriminant analysis to the newly constructed tensor, the emotional category classification becomes feasible. Since experimental results show that the proposed method enables the emotional category classification with the F1-measure of approximately 0.6, and about 10% improvement can be realized compared to comparative methods including state-of-the-art methods, the effectiveness of the proposed method is verified. Full article
(This article belongs to the Section Internet of Things)
►▼ Show Figures

Figure 1

Figure 1
<p>Overview of our new gaze-based image representation. Note that we handle color images in our method, but this figure shows a gray-scale version to visually explain our image representation. GIW matrices represent “gaze and image weight matrices”, which are explained in <a href="#sec3dot1-sensors-20-02146" class="html-sec">Section 3.1</a>.</p> Full article ">Figure 2
<p>Overview of our method. We construct the new gaze-based image representation and extract multiple convolutional neural network (CNN) features. By aligning these CNN features, we construct a CNN feature-based tensor (CFT) and apply both general tensor discriminant analysis and logistic tensor regression to the CFT. Finally, our method classifies images into emotional categories using outputs of the proposed network. Details of the procedures are shown in <a href="#sec3dot1-sensors-20-02146" class="html-sec">Section 3.1</a>, <a href="#sec3dot2-sensors-20-02146" class="html-sec">Section 3.2</a>, <a href="#sec3dot3-sensors-20-02146" class="html-sec">Section 3.3</a> and <a href="#sec3dot4-sensors-20-02146" class="html-sec">Section 3.4</a>.</p> Full article ">Figure 3
<p>This figure shows some experimental results of some test images and their ground truths. The areas that the subjects gazed at are shown in white at frames 1, 50, and 100. From these gaze data, PM (D-I-X) classifies this image into some categories. If the classified category is the same as the ground truth, the corresponding category is indicated in red. If the classified category is false, the corresponding category is indicated in black.</p> Full article ">
20 pages, 3332 KiB  
Article
YOLO-Tomato: A Robust Algorithm for Tomato Detection Based on YOLOv3
by Guoxu Liu, Joseph Christian Nouaze, Philippe Lyonel Touko Mbouembe and Jae Ho Kim
Sensors 2020, 20(7), 2145; https://doi.org/10.3390/s20072145 - 10 Apr 2020
Cited by 289 | Viewed by 21883
Abstract
Automatic fruit detection is a very important benefit of harvesting robots. However, complicated environment conditions, such as illumination variation, branch, and leaf occlusion as well as tomato overlap, have made fruit detection very challenging. In this study, an improved tomato detection model called [...] Read more.
Automatic fruit detection is a very important benefit of harvesting robots. However, complicated environment conditions, such as illumination variation, branch, and leaf occlusion as well as tomato overlap, have made fruit detection very challenging. In this study, an improved tomato detection model called YOLO-Tomato is proposed for dealing with these problems, based on YOLOv3. A dense architecture is incorporated into YOLOv3 to facilitate the reuse of features and help to learn a more compact and accurate model. Moreover, the model replaces the traditional rectangular bounding box (R-Bbox) with a circular bounding box (C-Bbox) for tomato localization. The new bounding boxes can then match the tomatoes more precisely, and thus improve the Intersection-over-Union (IoU) calculation for the Non-Maximum Suppression (NMS). They also reduce prediction coordinates. An ablation study demonstrated the efficacy of these modifications. The YOLO-Tomato was compared to several state-of-the-art detection methods and it had the best detection performance. Full article
(This article belongs to the Section Intelligent Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>YOLO model detection.</p> Full article ">Figure 2
<p>A 4-layer dense block. Each layer takes all preceding feature-maps as input and serves as input for all subsequent layers. <math display="inline"><semantics> <msub> <mi>H</mi> <mi>i</mi> </msub> </semantics></math> denotes the operation BN-ReLU-Conv<sub>1×1</sub>-BN-ReLU-Conv<sub>3×3</sub>.</p> Full article ">Figure 3
<p>Tomato samples with different growing circumstances: (<b>a</b>) two separated tomato, (<b>b</b>) a cluster of tomatoes, (<b>c</b>) occlusion case, (<b>d</b>) shading conditions, and (<b>e</b>) sunlight conditions.</p> Full article ">Figure 4
<p>Some examples of image augmentation operations: (<b>left</b>): original images, (<b>middle</b>): scaling and cropping, and (<b>right</b>): horizontal flip.</p> Full article ">Figure 5
<p>An overview of the proposed model.</p> Full article ">Figure 6
<p>A flowchart of training and detection process of YOLO-Tomato.</p> Full article ">Figure 7
<p>Dense architecture of the proposed model.</p> Full article ">Figure 8
<p>Overlap of two C-Bboxes.</p> Full article ">Figure 9
<p>C-Bbox prediction. The black dotted circle indicates the prior anchor, and the red circle is the prediction.</p> Full article ">Figure 10
<p>Clustering anchor dimensions of R-Bbox and C-Bbox. The k-means clustering was used to get the prior anchors. As in [<a href="#B24-sensors-20-02145" class="html-bibr">24</a>], the IoU was adopted instead of the Euclidean distance as the evaluation metric. As indicated by the dotted vertical line, nine clusters were adopted as the prior anchors, and were then divided into three parts and assigned to each of the three scales for detection.</p> Full article ">Figure 11
<p>P–R curves of different methods for ablation study. The markers indicate the points where recall and precision are obtained when the prediction confidence threshold equals 0.8.</p> Full article ">Figure 12
<p>(<b>a</b>) the 32 <math display="inline"><semantics> <mrow> <mn>3</mn> <mo>×</mo> <mn>3</mn> </mrow> </semantics></math> filters of Conv1 of the network, (<b>b</b>) the input image (cyan circles are marked manually for a better visualization), and (<b>c</b>–<b>e</b>), one of the feature activations from the 80th, 86th, and 92th convolutional layers, respectively.</p> Full article ">Figure 13
<p>(<b>a</b>) P–R curves, and (<b>b</b>) the F<sub>1</sub> scores of the models trained with different size of datasets.</p> Full article ">Figure 14
<p>Some examples of the detection results under different lighting conditions.</p> Full article ">Figure 15
<p>Some missed detection results due to severe occlusion by leaves or other tomatoes: (<b>a</b>) the green tomato which was largely covered by the red one was not detected, and (<b>b</b>) the red tomato was missed due to severe occlusion by leaves, stems and other tomatoes.</p> Full article ">Figure 16
<p>P–R curve for different detection methods.</p> Full article ">Figure 17
<p>The AP of different methods performed on the 30 sub-datasets displayed using boxplots. The red lines indicate the median values of AP, and ”+” indicates the outliers.</p> Full article ">
21 pages, 45267 KiB  
Article
Beam Deflection Monitoring Based on a Genetic Algorithm Using Lidar Data
by Michael Bekele Maru, Donghwan Lee, Gichun Cha and Seunghee Park
Sensors 2020, 20(7), 2144; https://doi.org/10.3390/s20072144 - 10 Apr 2020
Cited by 15 | Viewed by 5334
Abstract
The Light Detection And Ranging (LiDAR) system has become a prominent tool in structural health monitoring. Among such systems, Terrestrial Laser Scanning (TLS) is a potential technology for the acquisition of three-dimensional (3D) information to assess structural health conditions. This paper enhances the [...] Read more.
The Light Detection And Ranging (LiDAR) system has become a prominent tool in structural health monitoring. Among such systems, Terrestrial Laser Scanning (TLS) is a potential technology for the acquisition of three-dimensional (3D) information to assess structural health conditions. This paper enhances the application of TLS to damage detection and shape change analysis for structural element specimens. Specifically, estimating the deflection of a structural element with the aid of a Lidar system is introduced in this study. The proposed approach was validated by an indoor experiment by inducing artificial deflection on a simply supported beam. A robust genetic algorithm method is utilized to enhance the accuracy level of measuring deflection using lidar data. The proposed research primarily covers robust optimization of a genetic algorithm control parameter using the Taguchi experiment design. Once the acquired data is defined in terms of plane, which has minimum error, using a genetic algorithm and the deflection of the specimen can be extracted from the shape change analysis. Full article
(This article belongs to the Special Issue Sensors for Nondestructive Testing and Evaluation)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Flow chart for the proposed method.</p> Full article ">Figure 2
<p>(<b>a</b>) Point cloud before preprocessing; (<b>b</b>) Point cloud after preprocessing; (<b>c</b>) solid figure of a specimen.</p> Full article ">Figure 3
<p>GA Basic Features.</p> Full article ">Figure 4
<p>Genetic Algorithm Setup.</p> Full article ">Figure 5
<p>Plane representation for web and flange part.</p> Full article ">Figure 6
<p>(<b>a</b>) Photographic view of the experimental Setup; (<b>b</b>) Detailed drawing of the specimen and setup.</p> Full article ">Figure 7
<p>Illustration of LVDT sensor’s position.</p> Full article ">Figure 8
<p>Signal to noise ratio plot for each of the GA parameters.</p> Full article ">Figure 9
<p>Genetic Algorithm generation and individuals value for unloading case.</p> Full article ">Figure 10
<p>Flange point cloud before and after transformation of the coordinates.</p> Full article ">Figure 11
<p>Genetic algorithm generation and individual values for a deflected shape.</p> Full article ">Figure 12
<p>Deflected shape of flange entity with deflection curve at the enter line.</p> Full article ">Figure 13
<p>(<b>a</b>) model vs LVDT @ 0.55 m;(<b>b</b>) model vs LVDT @ 1.00 m; (<b>c</b>) model vs LVDT @ 1.45 m; (<b>d</b>) Absolute Error for model at mid-span.</p> Full article ">
13 pages, 5512 KiB  
Article
An Affordable Fabrication of a Zeolite-Based Capacitor for Gas Sensing
by Salvatore Andrea Pullano, Francesco Falcone, Davide C. Critello, Maria Giovanna Bianco, Michele Menniti and Antonino S. Fiorillo
Sensors 2020, 20(7), 2143; https://doi.org/10.3390/s20072143 - 10 Apr 2020
Cited by 12 | Viewed by 3417
Abstract
The development of even more compact, inexpensive, and highly sensitive gas sensors is widespread, even though their performances are still limited and technological improvements are in continuous evolution. Zeolite is a class of material which has received particular attention in different applications due [...] Read more.
The development of even more compact, inexpensive, and highly sensitive gas sensors is widespread, even though their performances are still limited and technological improvements are in continuous evolution. Zeolite is a class of material which has received particular attention in different applications due to its interesting adsorption/desorption capabilities. The behavior of a zeolite 4A modified capacitor has been investigated for the adsorption of nitrogen (N2), nitric oxide (NO) and 1,1-Difluoroethane (C2H4F2), which are of interest in the field of chemical, biological, radiological, and nuclear threats. Sample measurements were carried out in different environmental conditions, and the variation of the sensor electric capacitance was investigated. The dielectric properties were influenced by the type and concentration of gas species in the environment. Higher changes in capacitance were shown during the adsorption of dry air (+4.2%) and fluorinated gas (+7.3%), while lower dielectric variations were found upon exposure to N2 (−0.4%) and NO (−0.5%). The proposed approach pointed-out that a simple fabrication process may provide a convenient and affordable fabrication of reusable capacitive gas sensor. Full article
(This article belongs to the Special Issue Fabrication and Machining Technologies for Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Proof of concept of the zeolite-based gas sensor. A nanoporous layer acts as a molecules-sensitive dielectric, allowing a capacitive detection of the analyte, which results in an electrical signal correlated to the gas species adsorbed. The metal layers, together with the nanoporous dielectric, form a parallel plate capacitor whose electrodes are permeable to the gas stream, in order to allow a full diffusion of the molecules inside the zeolite. E(θ) represents the electric field between the two large parallel plates which is expected to be function of θ, defined as the number of adsorbed molecules compared to the total number of adsorption sites.</p> Full article ">Figure 2
<p>Fabrication procedure of gas sensitive nanoporous capacitor. (<b>a</b>) Starting from a cleaned glass substrate, (<b>b</b>) a thin Nickel electrode is deposited, and then (<b>c</b>) a zeolite composite layer is spun coated. (<b>d</b>) An upper Nickel electrode is deposited and, (<b>e</b>) a wet etching process is used for removing the zeolite excess, obtaining (<b>f</b>) a planar nanoporous capacitor.</p> Full article ">Figure 3
<p>SEM images of: (<b>a</b>) the glass substrate; (<b>b</b>) the nanoporous grain composing the mixture; (<b>c</b>) the deposited composite layer after annealing process; and (<b>d</b>) the upper view of the layer.</p> Full article ">Figure 4
<p>(<b>a</b>) FTIR Analysis of the zeolite composite layer, with the soybean oil spectrum in the inset. (<b>b</b>) Profilometric analysis of a sample of 10 μm in thickness.</p> Full article ">Figure 5
<p>Setup used for the characterization of the zeolite-based capacitor for gas sensing (not to scale).</p> Full article ">Figure 6
<p>Relationship inside the vacuum chamber at 20 mbar and 1 bar dry air saturated environment of: (<b>a</b>) electric capacitance vs. frequency; and (<b>b</b>) dissipation factor vs. frequency.</p> Full article ">Figure 7
<p>(<b>a</b>) Electric capacitance vs. pressure at three different frequencies onto the same sample. (<b>b</b>) Time evolution of the capacitance at constant pressure and temperature.</p> Full article ">Figure 8
<p>Adsorption of N<sub>2</sub>, NO, and C<sub>2</sub>H<sub>4</sub>F<sub>2</sub> by the Zeolite-based sensor. (<b>a</b>) Time behavior of capacitance and (<b>b</b>) frequency behavior of capacitance during the adsorption of the three different gases. In all cases, pressure was maintained at 1 bar.</p> Full article ">Figure 9
<p>Relation between capacitance of the sample maintained at 20 mbar for 1 h and the latter after injection of gas at 1 bar for 40 min.</p> Full article ">
22 pages, 13332 KiB  
Article
Surface Defect System for Long Product Manufacturing Using Differential Topographic Images
by F.J. delaCalle Herrero, Daniel F. García and Rubén Usamentiaga
Sensors 2020, 20(7), 2142; https://doi.org/10.3390/s20072142 - 10 Apr 2020
Cited by 7 | Viewed by 3132
Abstract
Current industrial products must meet quality requirements defined by international standards. Most commercial surface inspection systems give qualitative detections after a long, cumbersome and very expensive configuration process made by the seller company. In this paper, a new surface defect detection method is [...] Read more.
Current industrial products must meet quality requirements defined by international standards. Most commercial surface inspection systems give qualitative detections after a long, cumbersome and very expensive configuration process made by the seller company. In this paper, a new surface defect detection method is proposed based on 3D laser reconstruction. The method compares the long products, scan by scan, with their desired shape and produces differential topographic images of the surface at very high speeds. This work proposes a novel method where the values of the pixels in the images have a direct translation to real-world dimensions, which enables a detection based on the tolerances defined by international standards. These images are processed using computer vision techniques to detect defects and filter erroneous detections using both statistical distributions and a multilayer perceptron. Moreover, a systematic configuration procedure is proposed that is repeatable and can be performed by the manufacturer. The method has been tested using train track rails, which reports better results than two photometric systems including one commercial system, in both defect detection and erroneous detection rate. The method has been validated using a surface inspection rail pattern showing excellent performance. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Profile acquisition.</p> Full article ">Figure 2
<p>Acquisition and representation scheme.</p> Full article ">Figure 3
<p>Acquisition and representation example.</p> Full article ">Figure 4
<p>Detection pipeline.</p> Full article ">Figure 5
<p>Filter effect.</p> Full article ">Figure 6
<p>Segmentation procedure.</p> Full article ">Figure 7
<p>Region environment extraction.</p> Full article ">Figure 8
<p>Multilayer perceptron.</p> Full article ">Figure 9
<p>Rail profile acquisition.</p> Full article ">Figure 10
<p>Acquisition and representation example of a real surface defect.</p> Full article ">Figure 11
<p>Segmentation threshold exploration. (<b>a</b>) is the evolution of the Recall depending on the value of the segmentation threshold. (<b>b</b>) is the evoution of the number of False Positives per Rail depending on the value of the segmentation threshold.</p> Full article ">Figure 12
<p>Evaluated configurations for mask dimensions.</p> Full article ">Figure 13
<p>Environment filter distributions. Distribution of the values of the true-positive (TP) in red while distribution of the values of the false-positive (FP) in blue.</p> Full article ">Figure 14
<p>Neural network configurations.</p> Full article ">Figure 15
<p>Validation rail pattern.</p> Full article ">Figure 16
<p>Defects in the rail pattern.</p> Full article ">
15 pages, 5955 KiB  
Article
Wi-Fi CSI-Based Outdoor Human Flow Prediction Using a Support Vector Machine
by Masakatsu Ogawa and Hirofumi Munetomo
Sensors 2020, 20(7), 2141; https://doi.org/10.3390/s20072141 - 10 Apr 2020
Cited by 6 | Viewed by 3771
Abstract
This paper proposes a channel state information (CSI)-based prediction method of a human flow that includes activity. The objective of the paper is to predict a human flow in an outdoor road. This human flow prediction is useful for the prediction of the [...] Read more.
This paper proposes a channel state information (CSI)-based prediction method of a human flow that includes activity. The objective of the paper is to predict a human flow in an outdoor road. This human flow prediction is useful for the prediction of the number of passing people and their activity without privacy issues as a result of the absence of any camera systems. In this paper, we assume seven types of activities: one, two, and three people walking; one, two, and three people running; and one person cycling. Since the CSI can effectively express the effect of multipath fading in wireless signals, we expected the CSI to predict the various activities. In our proposed method, the amplitude and phase components are extracted from the measured CSI. The feature values for machine learning are the mean and variance of the maximum eigenvalue derived from the auto-correlation matrix and variance–covariance matrix composed of the amplitude or phase components and the passing time of flow. Using these feature values, we evaluated the prediction accuracy by leave-one-out cross-validation with a linear support vector machine (SVM). As a result, the proposed method achieved the maximum prediction accuracy of 100% for each direction and 99.5% for two directions. Full article
(This article belongs to the Special Issue Machine Learning Models for Wireless Network Monitoring and Data Analysis)
►▼ Show Figures

Figure 1

Figure 1
<p>Overview of human flow prediction.</p> Full article ">Figure 2
<p>Overview of the proposed system.</p> Full article ">Figure 3
<p>The maximum eigenvalue characteristics of channel state information (CSI) amplitude when the low pass filter is not applied.</p> Full article ">Figure 4
<p>The maximum eigenvalue characteristics of CSI amplitude when the low pass filter is applied.</p> Full article ">Figure 5
<p>Primary passing section in each path.</p> Full article ">Figure 6
<p>Final passing section in each path.</p> Full article ">Figure 7
<p>Summary of feature values.</p> Full article ">Figure 8
<p>Photo of the experimental environment.</p> Full article ">Figure 9
<p>Antenna deployment.</p> Full article ">Figure 10
<p>Confusion matrix for transmitter side to receiver side (TR) and receiver side to transmitter side (RT) directions when using only the lengths of the final passing section.</p> Full article ">
16 pages, 28099 KiB  
Article
A Method for Human Facial Image Annotation on Low Power Consumption Autonomous Devices
by Tomasz Hachaj
Sensors 2020, 20(7), 2140; https://doi.org/10.3390/s20072140 - 10 Apr 2020
Cited by 1 | Viewed by 31069
Abstract
This paper proposes a classifier designed for human facial feature annotation, which is capable of running on relatively cheap, low power consumption autonomous microcomputer systems. An autonomous system is one that depends only on locally available hardware and software—for example, it does not [...] Read more.
This paper proposes a classifier designed for human facial feature annotation, which is capable of running on relatively cheap, low power consumption autonomous microcomputer systems. An autonomous system is one that depends only on locally available hardware and software—for example, it does not use remote services available through the Internet. The proposed solution, which consists of a Histogram of Oriented Gradients (HOG) face detector and a set of neural networks, has comparable average accuracy and average true positive and true negative ratio to state-of-the-art deep neural network (DNN) architectures. However, contrary to DNNs, it is possible to easily implement the proposed method in a microcomputer with very limited RAM memory and without the use of additional coprocessors. The proposed method was trained and evaluated on a large 200,000 image face data set and compared with results obtained by other researchers. Further evaluation proves that it is possible to perform facial image attribute classification using the proposed algorithm on incoming video data captured by an RGB camera sensor of the microcomputer. The obtained results can be easily reproduced, as both the data set and source code can be downloaded. Developing and evaluating the proposed facial image annotation algorithm and its implementation, which is easily portable between various hardware and operating systems (virtually the same code works both on high-end PCs and microcomputers using the Windows and Linux platforms) and which is dedicated for low power consumption devices without coprocessors, is the main and novel contribution of this research. Full article
(This article belongs to the Section State-of-the-Art Sensors Technologies)
►▼ Show Figures

Figure 1

Figure 1
<p>Several random pictures from the CelebA dataset, which was used in this research. The images were aligned and cropped such that eyes of each person are approximately in the same position. The faces were also rescaled, such that they have similar size.</p> Full article ">Figure 2
<p>A mean face (in which pixels are averaged values of whole training data set of 50,000 images), then first four eigenfaces, sixth eigenface (six first eigenfaces explains 50% of variance), 26th (26 first eigenfaces explains 75% of variance), 156th (156 first eigenfaces explains 90% of variance), 473rd (473 first eigenfaces explains 95% of variance), and 3128th (3128 first eigenfaces explains 99% of variance).</p> Full article ">Figure 3
<p>Actual (input) image and its reconstruction using various numbers of eigenfaces, which describe 99%, 95%, 90%, and 75% of variance, respectively.</p> Full article ">Figure 4
<p>The assembled microcomputer that was used in this research. It consists of a Raspberry Pi 3 model B+ in a semi-transparent case with a 3.5″ LCD touchscreen, a Logitech Webcam C920 HD Pro, and a powerbank (5 V, 3 A).</p> Full article ">
21 pages, 11019 KiB  
Article
The Gas Fire Temperature Measurement for Detection of an Object’s Presence on Top of the Burner
by Andrzej Milecki and Dominik Rybarczyk
Sensors 2020, 20(7), 2139; https://doi.org/10.3390/s20072139 - 10 Apr 2020
Cited by 3 | Viewed by 11207
Abstract
This article covers the topic of temperature measurement on top of a gas burner fire in order to recognize pot removal from a gas burner and subsequently, to cut off the gas supply. The possibility of applying a factory-mounted thermocouple was investigated with [...] Read more.
This article covers the topic of temperature measurement on top of a gas burner fire in order to recognize pot removal from a gas burner and subsequently, to cut off the gas supply. The possibility of applying a factory-mounted thermocouple was investigated with the assumption that its output signal could be used to detect the presence of a pot on a gas burner. However, the characteristic of such a thermocouple is not fully linear and as the research has shown that such a thermocouple would not fit enough for the assumed purpose, thus another sensor needs to be used. Therefore, in this paper, the linear thermocouple and IR diode are used. The best localizations of theses sensors were investigated in order to obtain a signal suitable for the pot presence recognition over the burner. These investigations are supported by the use of an infrared camera. In the investigations, the temperature changes also caused by casual air blast or caused by increasing and decreasing the valve opening are recorded and analyzed. Finally, the changes of the thermocouple’s signals are used as an input signal to propose an algorithm for pot absence recognition over the burner. The microprocessor-based circuit with a control unit for detection of the pot absence is designed, built and investigated. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>The gas electromagnet valve with the thermocouple [<a href="#B6-sensors-20-02139" class="html-bibr">6</a>].</p> Full article ">Figure 2
<p>The photo of the SI sensors [<a href="#B22-sensors-20-02139" class="html-bibr">22</a>].</p> Full article ">Figure 3
<p>Scheme of the test stand.</p> Full article ">Figure 4
<p>Factory mounted thermocouple characteristics.</p> Full article ">Figure 5
<p>Temperature changes measured by K-type thermocouples while operating the valve (opening and closing): (<b>a</b>) without a pot; (<b>b</b>) with a pot.</p> Full article ">Figure 6
<p>Temperature changes during non-controlled air blow caused by: (<b>a</b>) opening the window when the flame is small; (<b>b</b>) quick and slow human movement when the flame is big.</p> Full article ">Figure 7
<p>Temperature changes measured by a factory mounted thermocouple when putting the pot on and removing it from the burner, when the flame is big (valve on max.) and small (valve on min.).</p> Full article ">Figure 8
<p>Extended characteristics measured by a factory mounted thermocouple if putting the pot on and removing it from the burner, when the flame is: (<b>a</b>) big; (<b>b</b>) small; explanation: red circle—start points of temperature changes: <span class="html-italic">T<sub>Fb</sub></span><sub>1</sub>, <span class="html-italic">T<sub>Fb</sub></span><sub>3</sub> and <span class="html-italic">T<sub>Fs</sub></span><sub>1</sub>, <span class="html-italic">T<sub>Fs</sub></span><sub>3</sub>; blue circle—points for recognition that the temperature changes significantly within 1s; green circle – recognition that the pot is placed on the burner <span class="html-italic">T<sub>Fb</sub></span><sub>2</sub>, <span class="html-italic">T<sub>Fs2</sub></span> (temperatures after 5 s); yellow circle—recognition that the pot is taken from the burner <span class="html-italic">T<sub>Fb</sub></span><sub>4</sub> and <span class="html-italic">T<sub>Fs</sub></span><sub>4.</sub></p> Full article ">Figure 9
<p>Temperature changes measured by linear thermocouple located in position 3.</p> Full article ">Figure 10
<p>Pictures taken by the infrared camera: (<b>a</b>) small flame without a pot, (<b>b</b>) big flame without a pot, (<b>c</b>) small flame with a pot of 220 mm diameter, (<b>d</b>) big flame with a pot of 220 mm diameter, (<b>e</b>) small flame with an enameled pot of 170 mm diameter, (<b>f</b>) big flame with an enameled pot of 170 mm diameter.</p> Full article ">Figure 11
<p>The photo of the used thermocouples.</p> Full article ">Figure 12
<p>The positions of the linear thermocouple.</p> Full article ">Figure 13
<p>Temperature changes for putting the pot and removing it from a burner when the flame is big and small measured by linear thermocouple located in: (<b>a</b>) position 1, (<b>b</b>) position 2.</p> Full article ">Figure 14
<p>Temperature changes measured by the linear thermocouple located in position 3.</p> Full article ">Figure 15
<p>Extended characteristics measured by the linear thermocouple in position 1 (<a href="#sensors-20-02139-f011" class="html-fig">Figure 11</a>) if putting the pot on and removing it from the burner when the flame is: (<b>a</b>) big, (<b>b</b>) small; explanation: red circles—start points for big flame <span class="html-italic">T<sub>Lb</sub></span><sub>1</sub> and <span class="html-italic">T<sub>Lb</sub></span><sub>3</sub> and for small flame <span class="html-italic">T<sub>Ls</sub></span><sub>1</sub> and <span class="html-italic">T<sub>Ls</sub></span><sub>3</sub>; blue circle—recognition of temperature changes, green circle—recognition that the pot is placed on the burner <span class="html-italic">T<sub>Lb</sub></span><sub>2</sub> and <span class="html-italic">T<sub>Ls</sub></span><sub>2</sub>; yellow circle—recognition that the pot is taken from the burner <span class="html-italic">T<sub>Lb</sub></span><sub>4</sub> and <span class="html-italic">T<sub>Ls</sub></span><sub>4.</sub></p> Full article ">Figure 16
<p>Temperature changes measured by the linear thermocouple located in position 3.</p> Full article ">Figure 17
<p>Temperature changes measured by the linear thermocouple in position 1, when the gas valve was closing and opening i.e., when reducing and increasing the gas flow to the burner if the pot is: (<b>a</b>) off, (<b>b</b>) on the burner.</p> Full article ">Figure 18
<p>The application of the IR diode for the temperature measurement.</p> Full article ">Figure 19
<p>IR diode output signal if the flame is on and off.</p> Full article ">Figure 20
<p>Scheme of the electronic controller.</p> Full article ">Figure 21
<p>Photo of the test stand with the electronics.</p> Full article ">Figure 22
<p>Signal changes measured by both thermocouples for the placing and removing of the pot from a burner if the flame is: (<b>a</b>) big, (<b>b</b>) small, and microcontroller binary output signal.</p> Full article ">
22 pages, 48850 KiB  
Article
Vacant Parking Slot Detection in the Around View Image Based on Deep Learning
by Wei Li, Libo Cao, Lingbo Yan, Chaohui Li, Xiexing Feng and Peijie Zhao
Sensors 2020, 20(7), 2138; https://doi.org/10.3390/s20072138 - 10 Apr 2020
Cited by 42 | Viewed by 10791
Abstract
Due to the complex visual environment, such as lighting variations, shadows, and limitations of vision, the accuracy of vacant parking slot detection for the park assist system (PAS) with a standalone around view monitor (AVM) needs to be improved. To address this problem, [...] Read more.
Due to the complex visual environment, such as lighting variations, shadows, and limitations of vision, the accuracy of vacant parking slot detection for the park assist system (PAS) with a standalone around view monitor (AVM) needs to be improved. To address this problem, we propose a vacant parking slot detection method based on deep learning, namely VPS-Net. VPS-Net converts the vacant parking slot detection into a two-step problem, including parking slot detection and occupancy classification. In the parking slot detection stage, we propose a parking slot detection method based on YOLOv3, which combines the classification of the parking slot with the localization of marking points so that various parking slots can be directly inferred using geometric cues. In the occupancy classification stage, we design a customized network whose size of convolution kernel and number of layers are adjusted according to the characteristics of the parking slot. Experiments show that VPS-Net can detect various vacant parking slots with a precision rate of 99.63% and a recall rate of 99.31% in the ps2.0 dataset, and has a satisfying generalizability in the PSV dataset. By introducing a multi-object detection network and a classification network, VPS-Net can detect various vacant parking slots robustly. Full article
(This article belongs to the Section Sensor Networks)
►▼ Show Figures

Figure 1

Figure 1
<p>Three typical kinds of parking slots. (<b>a</b>) perpendicular parking slots; (<b>b</b>) parallel parking slots; (<b>c</b>) slanted parking slots. A parking slot consists of four vertices, of which the paired marking points of the entrance line are marked with red dots, and the other two invisible vertices are marked with yellow dots. The entrance lines and the viewing range of an AVM system are also marked out.</p> Full article ">Figure 2
<p>Overview of the VPS-Net, which contains two modules: parking slot detection and occupancy classification. It takes the around view image as input and outputs the position of the vacant parking slot to the decision module of the PAS.</p> Full article ">Figure 3
<p>Marking points and parking slot heads. (<b>a</b>) shows the geometric relationship between the paired marking points and the parking slot head. Paired marking points are marked with green dots, and the parking slot head is marked with the red rectangle; (<b>b</b>) shows a variety of deformations of “T-shaped” or “L-shaped” marking points; (<b>c</b>) shows three kinds of the parking slot head belonging to classes “right-angled head”, “obtuse-angled head”, and “acute-angled head” respectively.</p> Full article ">Figure 4
<p>The bounding boxes of the parking slot head and marking points. Each bounding box consists of three parts: coordinates of the center point, width, and height.</p> Full article ">Figure 5
<p>The relationship between two marking points <math display="inline"><semantics> <msub> <mi mathvariant="bold">p</mi> <mn>1</mn> </msub> </semantics></math>, <math display="inline"><semantics> <msub> <mi mathvariant="bold">p</mi> <mn>2</mn> </msub> </semantics></math> and the bounding box of the parking slot head <math display="inline"><semantics> <mi mathvariant="bold">B</mi> </semantics></math>. (<b>a</b>) shows <math display="inline"><semantics> <mrow> <msub> <mi mathvariant="bold">p</mi> <mn mathvariant="bold">1</mn> </msub> <mo>⊆</mo> <mi mathvariant="bold">B</mi> </mrow> </semantics></math> and <math display="inline"><semantics> <mrow> <msub> <mi mathvariant="bold">p</mi> <mn mathvariant="bold">2</mn> </msub> <mo>⊆</mo> <mi mathvariant="bold">B</mi> </mrow> </semantics></math>; (<b>b</b>) shows <math display="inline"><semantics> <mrow> <msub> <mi mathvariant="bold">p</mi> <mn mathvariant="bold">1</mn> </msub> <mo>⊆</mo> <mi mathvariant="bold">B</mi> </mrow> </semantics></math> and <math display="inline"><semantics> <mrow> <msub> <mi mathvariant="bold">p</mi> <mn mathvariant="bold">2</mn> </msub> <mo>⊄</mo> <mi mathvariant="bold">B</mi> </mrow> </semantics></math>; (<b>c</b>) shows <math display="inline"><semantics> <mrow> <msub> <mi mathvariant="bold">p</mi> <mn mathvariant="bold">1</mn> </msub> <mo>⊄</mo> <mi mathvariant="bold">B</mi> </mrow> </semantics></math> and <math display="inline"><semantics> <mrow> <msub> <mi mathvariant="bold">p</mi> <mn mathvariant="bold">2</mn> </msub> <mo>⊆</mo> <mi mathvariant="bold">B</mi> <mspace width="4pt"/> </mrow> </semantics></math>; (<b>d</b>) shows <math display="inline"><semantics> <mrow> <msub> <mi mathvariant="bold">p</mi> <mn mathvariant="bold">1</mn> </msub> <mo>⊄</mo> <mi mathvariant="bold">B</mi> </mrow> </semantics></math> and <math display="inline"><semantics> <mrow> <msub> <mi mathvariant="bold">p</mi> <mn mathvariant="bold">2</mn> </msub> <mo>⊄</mo> <mi mathvariant="bold">B</mi> </mrow> </semantics></math>.</p> Full article ">Figure 6
<p>Complete parking slot inference. (<b>a</b>–<b>d</b>) are the perpendicular parking slot, the parallel parking slot, the slanted parking with an acute angle, and the slanted parking with an obtuse angle respectively. Their depth is <math display="inline"><semantics> <msub> <mi>d</mi> <mn>1</mn> </msub> </semantics></math>, <math display="inline"><semantics> <msub> <mi>d</mi> <mn>2</mn> </msub> </semantics></math> and <math display="inline"><semantics> <msub> <mi>d</mi> <mn>3</mn> </msub> </semantics></math> respectively, and their parking angle is <math display="inline"><semantics> <msub> <mi>α</mi> <mn>1</mn> </msub> </semantics></math>, <math display="inline"><semantics> <msub> <mi>α</mi> <mn>2</mn> </msub> </semantics></math> and <math display="inline"><semantics> <msub> <mi>α</mi> <mn>3</mn> </msub> </semantics></math> respectively. <math display="inline"><semantics> <msub> <mi mathvariant="bold">p</mi> <mn>1</mn> </msub> </semantics></math>, <math display="inline"><semantics> <msub> <mi mathvariant="bold">p</mi> <mn>2</mn> </msub> </semantics></math> are two visible paired marking points, and <math display="inline"><semantics> <msub> <mi mathvariant="bold">p</mi> <mn>3</mn> </msub> </semantics></math>, <math display="inline"><semantics> <msub> <mi mathvariant="bold">p</mi> <mn>4</mn> </msub> </semantics></math> are two invisible vertices.</p> Full article ">Figure 7
<p>The orientation of the parking slot when the vehicle is around it. Two rectangular boxes formed by the entrance line with a depth <span class="html-italic">d</span> are marked with red and orange dotted lines. The rectangular box formed by the car model is marked with gree dotted lines. The red arrow indicates the orientation of the parking slot.</p> Full article ">Figure 8
<p>The orientation of the parking slot when the vehicle is parking into it. (<b>a</b>) shows the orientation of the vertical parking slot. (<b>b</b>) shows the orientation of the parallel parking slot. The red arrow indicates the orientation of the parking slot. The yellow dotted line indicates the entrance line.</p> Full article ">Figure 9
<p>Training samples for vacant parking slot classification. (<b>a</b>) a negative sample: a non-vacant regularized parking slot. (<b>b</b>) a positive sample: a vacant regularized parking slot.</p> Full article ">Figure 10
<p>Cases of datasets used in evaluation. Rows 1 and 2 are the annotation information that was labeled for ps2.0 and PSV datasets. The green indicates the vacant parking slot. The red indicates the non-vacant parking slot. Rows 3 and 4 are parking slot samples that were cut and warped according to the annotation information.</p> Full article ">Figure 11
<p>AP<math display="inline"><semantics> <msub> <mrow/> <mn>50</mn> </msub> </semantics></math> histograms by three kinds of DCNN-based detectors.</p> Full article ">Figure 12
<p>Detection results by YOLOv3-based detector. The green bounding box indicates the “right-angled head”. The blue bounding box indicates the “acute-angled head”. The yellow bounding box indicates the “obtuse-angled head”. The red dot indicates the “marking point”.</p> Full article ">Figure 13
<p>(<b>a</b>,<b>b</b>) show representative images in the ps2.0 test dataset where the vehicle is across parking slots.</p> Full article ">Figure 14
<p>Precision-recall curves of different methods for parking slot occupancy classification.</p> Full article ">Figure 15
<p>VPS-Net detection results. Green indicates the vacant parking slot. Red indicates the non-vacant parking slot. Different rows show three kinds of parking slots in various imaging conditions like ’indoor’, ’outdoor daylight’, ’outdoor rainy’, ’outdoor shadow’, ’outdoor slanted’, ’outdoor street light’ respectively.</p> Full article ">Figure 16
<p>Representative images with the degraded image quality of marking points in the PSV dataset. (<b>a</b>) shows the marking point is far from cameras. (<b>b</b>) shows the marking point is on the stitching lines. The green bounding box indicates the parking slot head. The red dot indicates the detected marking point, and the purple dot indicates the inferred marking point based on the parking slot head.</p> Full article ">
20 pages, 11395 KiB  
Article
HKSiamFC: Visual-Tracking Framework Using Prior Information Provided by Staple and Kalman Filter
by Chenpu Li, Qianjian Xing and Zhenguo Ma
Sensors 2020, 20(7), 2137; https://doi.org/10.3390/s20072137 - 10 Apr 2020
Cited by 9 | Viewed by 3101
Abstract
In the field of visual tracking, trackers based on a convolutional neural network (CNN) have had significant achievements. The fully-convolutional Siamese (SiamFC) tracker is a typical representation of these CNN trackers and has attracted much attention. It models visual tracking as a similarity-learning [...] Read more.
In the field of visual tracking, trackers based on a convolutional neural network (CNN) have had significant achievements. The fully-convolutional Siamese (SiamFC) tracker is a typical representation of these CNN trackers and has attracted much attention. It models visual tracking as a similarity-learning problem. However, experiments showed that SiamFC was not so robust in some complex environments. This may be because the tracker lacked enough prior information about the target. Inspired by the key idea of a Staple tracker and Kalman filter, we constructed two more models to help compensate for SiamFC’s disadvantages. One model contained the target’s prior color information, and the other the target’s prior trajectory information. With these two models, we design a novel and robust tracking framework on the basis of SiamFC. We call it Histogram–Kalman SiamFC (HKSiamFC). We also evaluated HKSiamFC tracker’s performance on dataset of the online object tracking benchmark (OTB) and Temple Color (TC128), and it showed quite competitive performance when compared with the baseline tracker and several other state-of-the-art trackers. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Basic workflow of our Histogram–Kalman SiamFC (HKSiamFC) tracker. The two-branch convolutional neural network (CNN) has the same architecture as the original fully-convolutional Siamese (SiamFC) tracker. We adopted the histogram model from Staple in HKSiamFC and used it to generate a histogram-score map, and we then optimized it with a Gaussian filter; this map is used to provide prior information about target appearance. We also designed a Kalman model to generate a Gaussian likelihood map based on a Kalman filter to provide prior information about target motion. Finally, we combined all three maps to capture the target.</p> Full article ">Figure 2
<p>(<b>top</b>) Four frames (#0010, #0011, #0023, and #24) of SiamFC’s tracking failure in Human8 sequence; (<b>bottom</b>) histogram maps of each frame corresponding to first row (#0010, #0011, #0023, and #24), the red region in the map represents high score.</p> Full article ">Figure 3
<p>Effect of processing original histogram-score map by Gaussian filter.</p> Full article ">Figure 4
<p>Workflow of using Kalman filter to generate Gaussian likelihood maps. The green rectangle in Frame1 is the ground-truth bounding box containing the right target.</p> Full article ">Figure 5
<p>Comparison of HKSiamFC and five baseline trackers on Online Object Tracking Benchmark (OTB) dataset. Three plot pairs are results of (<b>top</b>–<b>bottom</b>) OTB100, OTB50, and OTB2013. This picture is best viewed on high-resolution displays.</p> Full article ">Figure 6
<p>Comparison of HKSiamFC and five baseline trackers using precision-plot metric under the 11 tracking scenarios. This picture is best viewed on high-resolution displays.</p> Full article ">Figure 7
<p>Comparison of HKSiamFC and five baseline trackers using success-plot metric under the 11 tracking scenarios. This picture is best viewed on high-resolution displays.</p> Full article ">Figure 8
<p>Comparison between HKSiamFC and 14 state-of-the-art trackers on OTB100. This picture is best viewed on high-resolution displays.</p> Full article ">Figure 9
<p>Tracking result of 15 compared trackers in precision plot in 11 different scenarios. This picture is best viewed on high-resolution displays.</p> Full article ">Figure 10
<p>Tracking result of 18 compared trackers on Temple Color (TC128). This picture is best viewed on high-resolution displays.</p> Full article ">Figure 11
<p>Qualitative tracking results of HKSiamFC and 10 state-of-the-art trackers on several typical sequences of OTB. Sequences (<b>top–bottom</b>) rows are Biker, Diving, Human 3, Jump, skating 2-1, and soccer. The color of each tracker is listed at the bottom of this figure.</p> Full article ">
17 pages, 2516 KiB  
Article
Detection of Atrial Fibrillation Using 1D Convolutional Neural Network
by Chaur-Heh Hsieh, Yan-Shuo Li, Bor-Jiunn Hwang and Ching-Hua Hsiao
Sensors 2020, 20(7), 2136; https://doi.org/10.3390/s20072136 - 10 Apr 2020
Cited by 75 | Viewed by 6834
Abstract
The automatic detection of atrial fibrillation (AF) is crucial for its association with the risk of embolic stroke. Most of the existing AF detection methods usually convert 1D time-series electrocardiogram (ECG) signal into 2D spectrogram to train a complex AF detection system, which [...] Read more.
The automatic detection of atrial fibrillation (AF) is crucial for its association with the risk of embolic stroke. Most of the existing AF detection methods usually convert 1D time-series electrocardiogram (ECG) signal into 2D spectrogram to train a complex AF detection system, which results in heavy training computation and high implementation cost. This paper proposes an AF detection method based on an end-to-end 1D convolutional neural network (CNN) architecture to raise the detection accuracy and reduce network complexity. By investigating the impact of major components of a convolutional block on detection accuracy and using grid search to obtain optimal hyperparameters of the CNN, we develop a simple, yet effective 1D CNN. Since the dataset provided by PhysioNet Challenge 2017 contains ECG recordings with different lengths, we also propose a length normalization algorithm to generate equal-length records to meet the requirement of CNN. Experimental results and analysis indicate that our method of 1D CNN achieves an average F1 score of 78.2%, which has better detection accuracy with lower network complexity, as compared with the existing deep learning-based methods. Full article
(This article belongs to the Special Issue Multimodal Data Fusion and Machine-Learning for Healthcare)
►▼ Show Figures

Figure 1

Figure 1
<p>Electrocardiogram.</p> Full article ">Figure 2
<p>Flowchart of the proposed atrial fibrillation (AF) detection method. ECG: electrocardiogram; CNN: convolutional neural network.</p> Full article ">Figure 3
<p>Data length histogram distribution.</p> Full article ">Figure 4
<p>ECG examples of four classes: Normal, AF, Other, and Noisy.</p> Full article ">Figure 5
<p>Normalized average confusion matrix of five folds.</p> Full article ">Figure 6
<p>AF record which is misclassified as other.</p> Full article ">Figure 7
<p>Training accuracy of All-BN.</p> Full article ">Figure 8
<p>Training accuracy of No-BN.</p> Full article ">
29 pages, 8755 KiB  
Review
Chipless RFID Sensors for the Internet of Things: Challenges and Opportunities
by Viviana Mulloni and Massimo Donelli
Sensors 2020, 20(7), 2135; https://doi.org/10.3390/s20072135 - 10 Apr 2020
Cited by 76 | Viewed by 10851
Abstract
Radio-frequency identification (RFID) sensors are one of the fundamental components of the internet of things that aims at connecting every physical object to the cloud for the exchange of information. In this framework, chipless RFIDs are a breakthrough technology because they remove the [...] Read more.
Radio-frequency identification (RFID) sensors are one of the fundamental components of the internet of things that aims at connecting every physical object to the cloud for the exchange of information. In this framework, chipless RFIDs are a breakthrough technology because they remove the cost associated with the chip, being at the same time printable, passive, low-power and suitable for harsh environments. After the important results achieved with multibit chipless tags, there is a clear motivation and interest to extend the chipless sensing functionality to physical, chemical, structural and environmental parameters. These potentialities triggered a strong interest in the scientific and industrial community towards this type of application. Temperature and humidity sensors, as well as localization, proximity, and structural health prototypes, have already been demonstrated, and many other sensing applications are foreseen soon. In this review, both the different architectural approaches available for this technology and the requirements related to the materials employed for sensing are summarized. Then, the state-of-the-art of categories of sensors and their applications are reported and discussed. Finally, an analysis of the current limitations and possible solution strategies for this technology are given, together with an overview of expected future developments. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Number of entries per year on Google Scholar using the keywords “chipless RFID” (Radio-frequency identification) and “chipless RFID sensors” as for 24 February 2020.</p> Full article ">Figure 2
<p>(<b>a</b>) Schema of a bistatic reader. (<b>b</b>) Photo of a bistatic reader prototype.</p> Full article ">Figure 3
<p>(<b>a</b>) Schema of a monostatic reader. (<b>b</b>) Photo of a monostatic reader with a waveguide ferrite circulator.</p> Full article ">Figure 4
<p>Chipless RFID sensor tag structure including both sensing and coding resonators. (<b>a</b>) Backscattered chipless tag. (<b>b</b>) Retransmission-based chipless tag.</p> Full article ">Figure 5
<p>Photo of a single-bit tag structure.</p> Full article ">Figure 6
<p>Photo of chipless tags equipped with different numbers of spiral resonators.</p> Full article ">Figure 7
<p>Photo of a single-bit chipless tag equipped with two patch antennas with modified corners.</p> Full article ">Figure 8
<p>Schema of a chipless tag system operating in circular polarization.</p> Full article ">Figure 9
<p>Photograph of a five-bit chipless tag equipped with two broadband circular polarized antennas.</p> Full article ">Figure 10
<p>(<b>a</b>) Schema of a five-bit chipless tag equipped with a sensing element placed on the fifth resonator and a simulated five-bit chipless tag equipped with different resistors placed on the fifth resonator. (<b>b</b>) |S<sub>21</sub>| and (<b>c</b>) phase response.</p> Full article ">Figure 11
<p>(<b>a</b>) Photo of a five-bit chipless tag equipped with a thermistor and (<b>b</b>) measured |S<sub>21</sub>| response of a five-bit chipless tag equipped with a thermistor.</p> Full article ">Figure 12
<p>(<b>a</b>) Single-bit chipless linear Van Atta array composed by four elements with a single C resonator. (<b>b</b>) Simulated backscattered response in the X band (8GHz–12GHz).</p> Full article ">Figure 13
<p>(<b>a</b>) Two-bit chipless linear Van Atta array composed of four elements with two C resonators. (<b>b</b>) Simulated backscattered response in the X band (8GHz–12GHz).</p> Full article ">Figure 14
<p>(<b>a</b>) Three-bit chipless linear Van Atta array composed of four elements with three C resonators. (<b>b</b>) Simulated backscattered response in the X band (8GHz–12GHz).</p> Full article ">Figure 15
<p>(<b>a</b>) Photograph of a single-bit chipless Van Atta array RFID prototype fabricated on an FR4 dielectric substrate and equipped with a C resonator placed vertically with respect to the connection microstrip. (<b>b</b>) Measured signal from the prototype reported in (<b>a</b>).</p> Full article ">Figure 16
<p>(<b>a</b>) Photo of a single-bit chipless Van Atta array RFID prototype fabricated on an FR4 dielectric substrate and equipped with a C resonator placed horizontally with respect to the connection microstrip. (<b>b</b>) Measured signal from the prototype reported in (<b>a</b>).</p> Full article ">Figure 17
<p>(<b>a</b>) Photo of a single-bit chipless Van Atta array RFID prototype fabricated on an FR4 dielectric substrate and equipped with a square patch resonator. (<b>b</b>) Measured signal from the prototype reported in (<b>a</b>).</p> Full article ">Figure 18
<p>Tags realized by lithography. (<b>a</b>) Flexible sensing resonator structure (copper on Kapton). (<b>b</b>) Spiral resonator arrays (copper on 168 μm-thick Rogers 4350B).</p> Full article ">Figure 19
<p>(<b>a</b>) Ink-jet printed resonators arrays on polyethylene terephthalate substrate realized with highly conductive Ag-nanoparticles ink. (<b>b</b>) Detail of the same sample.</p> Full article ">Figure 20
<p>Prototype a chipless RFID sensor for angular rotation monitoring comprising a set of 4 × 4 unit cells with four-dipole elements in each one. From ref. [<a href="#B154-sensors-20-02135" class="html-bibr">154</a>].</p> Full article ">
16 pages, 4404 KiB  
Article
Laser Patterning a Graphene Layer on a Ceramic Substrate for Sensor Applications
by Marcin Lebioda, Ryszard Pawlak, Witold Szymański, Witold Kaczorowski and Agata Jeziorna
Sensors 2020, 20(7), 2134; https://doi.org/10.3390/s20072134 - 10 Apr 2020
Cited by 9 | Viewed by 3732
Abstract
This paper describes a method for patterning the graphene layer and gold electrodes on a ceramic substrate using a Nd:YAG nanosecond fiber laser. The technique enables the processing of both layers and trimming of the sensor parameters. The main aim was to develop [...] Read more.
This paper describes a method for patterning the graphene layer and gold electrodes on a ceramic substrate using a Nd:YAG nanosecond fiber laser. The technique enables the processing of both layers and trimming of the sensor parameters. The main aim was to develop a technique for the effective and efficient shaping of both the sensory layer and the metallic electrodes. The laser shaping method is characterized by high speed and very good shape mapping, regardless of the complexity of the processing. Importantly, the technique enables the simultaneous shaping of both the graphene layer and Au electrodes in a direct process that does not require a complex and expensive masking process, and without damaging the ceramic substrate. Our results confirmed the effectiveness of the developed laser technology for shaping a graphene layer and Au electrodes. The ceramic substrate can be used in the construction of various types of sensors operating in a wide temperature range, especially the cryogenic range. Full article
(This article belongs to the Section Sensor Materials)
►▼ Show Figures

Figure 1

Figure 1
<p>Laser patterning system.</p> Full article ">Figure 2
<p>Scheme of the laser patterning procedure.</p> Full article ">Figure 3
<p>Method of sample preparation. HSMG <sup>®</sup>: High-strength metallurgical graphene.</p> Full article ">Figure 4
<p>Cryogenic cooling system.</p> Full article ">Figure 5
<p>Microscopic pictures of graphene structures (paths) shaped by laser ablation: (<b>a</b>,<b>b</b>) optical microscopy and (<b>c</b>,<b>d</b>) SEM.</p> Full article ">Figure 6
<p>Raman spectra of high strength metallurgical graphene (HSMG<sup>®</sup>) after laser machining: (<b>a</b>) Raman spectrum of laser irradiated graphene on a Au/Al<sub>2</sub>O<sub>3</sub> substrate, (<b>b</b>) the Raman spectrum of laser-irradiated graphene on Al<sub>2</sub>O<sub>3</sub>, and (<b>c</b>) sample marked with points of acquisition of the Raman spectra.</p> Full article ">Figure 7
<p>Sample after laser patterning: (<b>a</b>) the HSMG<sup>®</sup> graphene and gold electrodes were cut and (<b>b</b>) the HSMG<sup>®</sup> graphene and gold electrodes were shaped.</p> Full article ">Figure 8
<p>Temperature dependence of the resistance of the sample before and after laser cutting: (<b>a</b>) resistance <span class="html-italic">R</span> (kΩ) of sample and (<b>b</b>) relative change in the sample resistance (<span class="html-italic">R<sub>0</sub></span>—initial resistance of the sample at <span class="html-italic">T</span><sub>init</sub> = 293 K).</p> Full article ">Figure 9
<p>Temperature dependence of the resistance of the sample before and after laser cutting and reducing: (<b>a</b>) resistance <span class="html-italic">R</span> (kΩ) of the sample and (<b>b</b>) relative changes in the sample resistance (<span class="html-italic">R<sub>0</sub></span>—initial resistance of the sample at <span class="html-italic">T</span><sub>init</sub> = 293 K).</p> Full article ">Figure 10
<p>Relative changes in the sample resistances before and after the laser treatment (<span class="html-italic">R<sub>0</sub></span>—initial resistance of the sample at <span class="html-italic">T</span><sub>init</sub> = 293 K).</p> Full article ">
12 pages, 1226 KiB  
Article
Thin Film Encapsulation for RF MEMS in 5G and Modern Telecommunication Systems
by Anna Persano, Fabio Quaranta, Antonietta Taurino, Pietro Aleardo Siciliano and Jacopo Iannacci
Sensors 2020, 20(7), 2133; https://doi.org/10.3390/s20072133 - 10 Apr 2020
Cited by 13 | Viewed by 4961
Abstract
In this work, SiNx/a-Si/SiNx caps on conductive coplanar waveguides (CPWs) are proposed for thin film encapsulation of radio-frequency microelectromechanical systems (RF MEMS), in view of the application of these devices in fifth generation (5G) and modern telecommunication systems. Simplification and [...] Read more.
In this work, SiNx/a-Si/SiNx caps on conductive coplanar waveguides (CPWs) are proposed for thin film encapsulation of radio-frequency microelectromechanical systems (RF MEMS), in view of the application of these devices in fifth generation (5G) and modern telecommunication systems. Simplification and cost reduction of the fabrication process were obtained, using two etching processes in the same barrel chamber to create a matrix of holes through the capping layer and to remove the sacrificial layer under the cap. Encapsulating layers with etch holes of different size and density were fabricated to evaluate the removal of the sacrificial layer as a function of the percentage of the cap perforated area. Barrel etching process parameters also varied. Finally, a full three-dimensional finite element method-based simulation model was developed to predict the impact of fabricated thin film encapsulating caps on RF performance of CPWs. Full article
(This article belongs to the Special Issue RF-MEMS Solutions for Advanced Passive Components)
►▼ Show Figures

Figure 1

Figure 1
<p>Sketch of the process flow for the fabrication of thin film encapsulating caps.</p> Full article ">Figure 2
<p>Scanning electron microscopy (SEM) view of a fabricated encapsulating cap (<b>a</b>). SEM cross-section of the SiN<sub>x</sub>/a-Si/SiN<sub>x</sub> trilayer at lower (<b>b</b>) and higher magnification (<b>c</b>), along a cleaved hole. SEM view of a hole through the cap surface after barrel etching (<b>d</b>) and inductively coupled plasma (ICP) etching (<b>e</b>). 3D surface profile of the cap in (<b>a</b>) under the application of a stylus tip force of 4.9 µN (<b>f</b>).</p> Full article ">Figure 3
<p>Tilted SEM images (tilt angle of 60°) of the fabricated encapsulating caps with A = 26% (<b>a</b>), 50% (<b>b</b>), 57% (<b>c</b>), and 74% (<b>d</b>). Profiles measured for caps in (<b>a</b>–<b>d</b>) under the application of a stylus tip force of 4.9 µN and 490 µN (<b>e</b>). Deflection in the center of the caps in (<b>a</b>–<b>d</b>) as a function of the stylus tip force (<b>f</b>). The marker is the same for all SEM images. The green line in (a) is the profiler scan line. Barrel parameters, which were used for the release of caps in (<b>a</b>–<b>d</b>), are P (power) = 600 W and t = 8 min.</p> Full article ">Figure 4
<p>Magnitude of simulated reflection (<b>a</b>) and insertion (<b>b</b>) losses for gold coplanar waveguides (CPWs) uncapped and encapsulated with the fabricated caps having A = 26%, 50%, and 57% on low and high resistivity silicon substrates.</p> Full article ">
12 pages, 1485 KiB  
Article
Gait Characteristics under Imposed Challenge Speed Conditions in Patients with Parkinson’s Disease During Overground Walking
by Myeounggon Lee, Changhong Youm, Byungjoo Noh, Hwayoung Park and Sang-Myung Cheon
Sensors 2020, 20(7), 2132; https://doi.org/10.3390/s20072132 - 10 Apr 2020
Cited by 11 | Viewed by 3028
Abstract
Evaluating gait stability at slower or faster speeds and self-preferred speeds based on continuous steps may assist in determining the severity of motor symptoms in Parkinson’s disease (PD) patients. This study aimed to investigate the gait ability at imposed speed conditions in PD [...] Read more.
Evaluating gait stability at slower or faster speeds and self-preferred speeds based on continuous steps may assist in determining the severity of motor symptoms in Parkinson’s disease (PD) patients. This study aimed to investigate the gait ability at imposed speed conditions in PD patients during overground walking. Overall, 74 PD patients and 52 age-matched healthy controls were recruited. Levodopa was administered to patients in the PD group, and all participants completed imposed slower, preferred, and faster speed walking tests along a straight 15-m walkway wearing shoe-type inertial measurement units. Reliability of the slower and faster conditions between the estimated and measured speeds indicated excellent agreement for PD patients and controls. PD patients demonstrated higher gait asymmetry (GA) and coefficient of variance (CV) for stride length and stance phase than the controls at slower speeds and higher CVs for phases for single support, double support, and stance. CV of the double support phase could distinguish between PD patients and controls at faster speeds. The GA and CVs of stride length and phase-related variables were associated with motor symptoms in PD patients. Speed conditions should be considered during gait analysis. Gait variability could evaluate the severity of motor symptoms in PD patients. Full article
(This article belongs to the Special Issue Smart Sensors: Applications and Advances in Human Motion Analysis)
►▼ Show Figures

Figure 1

Figure 1
<p>Recruitment process flowchart.</p> Full article ">Figure 2
<p>Schematic of the data collection and analysis phase under steady-state conditions. (<b>a</b>) Data collection and analysis phase; the blue arrows indicate acceleration to steady-state and deceleration steps after measurements are completed. (<b>b</b>) Detection of gait events with the shoe-type inertial measurement unit (IMU) system. Data is collected at 100 Hz. HS, heel strike; TO, toe-off.</p> Full article ">Figure 3
<p>Bland-Altman plots for data agreement between the estimated and measured overground walking speeds. (<b>a</b>) and (<b>b</b>) are the slower and faster speed results for PD patients; (<b>c</b>) and (<b>d</b>) are the slower and faster speed results for control patients. PD, Parkinson’s disease; LOA, limits of agreement.</p> Full article ">
14 pages, 12221 KiB  
Article
Counting Cattle in UAV Images—Dealing with Clustered Animals and Animal/Background Contrast Changes
by Jayme Garcia Arnal Barbedo, Luciano Vieira Koenigkan, Patrícia Menezes Santos and Andrea Roberto Bueno Ribeiro
Sensors 2020, 20(7), 2126; https://doi.org/10.3390/s20072126 - 10 Apr 2020
Cited by 43 | Viewed by 5738
Abstract
The management of livestock in extensive production systems may be challenging, especially in large areas. Using Unmanned Aerial Vehicles (UAVs) to collect images from the area of interest is quickly becoming a viable alternative, but suitable algorithms for extraction of relevant information from [...] Read more.
The management of livestock in extensive production systems may be challenging, especially in large areas. Using Unmanned Aerial Vehicles (UAVs) to collect images from the area of interest is quickly becoming a viable alternative, but suitable algorithms for extraction of relevant information from the images are still rare. This article proposes a method for counting cattle which combines a deep learning model for rough animal location, color space manipulation to increase contrast between animals and background, mathematical morphology to isolate the animals and infer the number of individuals in clustered groups, and image matching to take into account image overlap. Using Nelore and Canchim breeds as a case study, the proposed approach yields accuracies over 90% under a wide variety of conditions and backgrounds. Full article
(This article belongs to the Section Remote Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Steps adopted to generate the binary masks used for animal segmentation.</p> Full article ">Figure 2
<p>Examples of images resulting from each processing step shown in <a href="#sensors-20-02126-f001" class="html-fig">Figure 1</a>. (<b>A</b>) Original image; (<b>B</b>) region of interest; (<b>C</b>) quadrants of the image; (<b>D</b>) color channels used in the algorithm; (<b>E</b>) masks associated to each channel; (<b>F</b>) six resulting masks with all quadrants united.</p> Full article ">Figure 3
<p>Mask combination and estimation of the number of animals.</p> Full article ">Figure 4
<p>Confusion matrix crossing actual and estimated counts for different cluster sizes.</p> Full article ">Figure 5
<p>Example of image containing many calves.</p> Full article ">Figure 6
<p>Examples of images containing trees (<b>A</b>), shed (<b>B</b>) and feeder (<b>C</b>).</p> Full article ">Figure 7
<p>Examples of animals in different positions.</p> Full article ">Figure 8
<p>Examples of images with underexposure (<b>A</b>) and overexposure (<b>B</b>).</p> Full article ">
24 pages, 1772 KiB  
Article
Mining Massive E-Health Data Streams for IoMT Enabled Healthcare Systems
by Affan Ahmed Toor, Muhammad Usman, Farah Younas, Alvis Cheuk M. Fong, Sajid Ali Khan and Simon Fong
Sensors 2020, 20(7), 2131; https://doi.org/10.3390/s20072131 - 9 Apr 2020
Cited by 32 | Viewed by 5198
Abstract
With the increasing popularity of the Internet-of-Medical-Things (IoMT) and smart devices, huge volumes of data streams have been generated. This study aims to address the concept drift, which is a major challenge in the processing of voluminous data streams. Concept drift refers to [...] Read more.
With the increasing popularity of the Internet-of-Medical-Things (IoMT) and smart devices, huge volumes of data streams have been generated. This study aims to address the concept drift, which is a major challenge in the processing of voluminous data streams. Concept drift refers to overtime change in data distribution. It may occur in the medical domain, for example the medical sensors measuring for general healthcare or rehabilitation, which may switch their roles for ICU emergency operations when required. Detecting concept drifts becomes trickier when the class distributions in data are skewed, which is often true for medical sensors e-health data. Reactive Drift Detection Method (RDDM) is an efficient method for detecting long concepts. However, RDDM has a high error rate, and it does not handle class imbalance. We propose an Enhanced Reactive Drift Detection Method (ERDDM), which systematically generates strategies to handle concept drift with class imbalance in data streams. We conducted experiments to compare ERDDM with three contemporary techniques in terms of prediction error, drift detection delay, latency, and ability to handle data imbalance. The experimentation was done in Massive Online Analysis (MOA) on 48 synthetic datasets customized to possess the capabilities of data streams. ERDDM can handle abrupt and gradual drifts and performs better than all benchmarks in almost all experiments. Full article
(This article belongs to the Special Issue Advanced Machine Learning Tools and Methods for IoMT Sensor Applications)
►▼ Show Figures

Figure 1

Figure 1
<p>Infrastructure of Internet of Medical Things (IoMT).</p> Full article ">Figure 2
<p>Average prediction error in datasets with abrupt drift.</p> Full article ">Figure 3
<p>Average prediction error in datasets with gradual drift.</p> Full article ">Figure 4
<p>Average detection delay in datasets with abrupt drift.</p> Full article ">Figure 5
<p>Average detection delay in datasets with gradual drift.</p> Full article ">Figure 6
<p>Mean evaluation time of datasets with abrupt drift.</p> Full article ">Figure 7
<p>Mean evaluation time of datasets with gradual drift.</p> Full article ">Figure 8
<p>Detected drifts in datasets with abrupt drift.</p> Full article ">Figure 9
<p>Detected drifts in datasets with gradual drift.</p> Full article ">
14 pages, 6212 KiB  
Article
Classification for Penicillium expansum Spoilage and Defect in Apples by Electronic Nose Combined with Chemometrics
by Zhiming Guo, Chuang Guo, Quansheng Chen, Qin Ouyang, Jiyong Shi, Hesham R. El-Seedi and Xiaobo Zou
Sensors 2020, 20(7), 2130; https://doi.org/10.3390/s20072130 - 9 Apr 2020
Cited by 28 | Viewed by 3766
Abstract
It is crucial for the efficacy of the apple storage to apply methods like electronic nose systems for detection and prediction of spoilage or infection by Penicillium expansum. Based on the acquisition of electronic nose signals, selected sensitive feature sensors of spoilage [...] Read more.
It is crucial for the efficacy of the apple storage to apply methods like electronic nose systems for detection and prediction of spoilage or infection by Penicillium expansum. Based on the acquisition of electronic nose signals, selected sensitive feature sensors of spoilage apple and all sensors were analyzed and compared by the recognition effect. Principal component analysis (PCA), principle component analysis-discriminant analysis (PCA-DA), linear discriminant analysis (LDA), partial least squares discriminate analysis (PLS-DA) and K-nearest neighbor (KNN) were used to establish the classification model of apple with different degrees of corruption. PCA-DA has the best prediction, the accuracy of training set and prediction set was 100% and 97.22%, respectively. synergy interval (SI), genetic algorithm (GA) and competitive adaptive reweighted sampling (CARS) are three selection methods used to accurately and quickly extract appropriate feature variables, while constructing a PLS model to predict plaque area. Among them, the PLS model with unique variables was optimized by CARS method, and the best prediction result of the area of the rotten apple was obtained. The best results are as follows: Rc = 0.953, root mean square error of calibration (RMSEC) = 1.28, Rp = 0.972, root mean square error of prediction (RMSEP) = 1.01. The results demonstrated that the electronic nose has a potential application in the classification of rotten apples and the quantitative detection of spoilage area. Full article
(This article belongs to the Special Issue Non-destructive Sensors and Machine Learning for Food Safety & Quality Inspection)
►▼ Show Figures

Figure 1

Figure 1
<p>Schematic procedures of classification and prediction of defects of <span class="html-italic">Penicillium expansum</span> in apples by electronic nose combined with chemometrics.</p> Full article ">Figure 2
<p>(<b>a</b>) Data of each sensor of a single corrupt apple; (<b>b</b>) response signals of various sensors to apple gases with different degrees of corruption.</p> Full article ">Figure 3
<p>(<b>a</b>) Principle component analysis (PCA) results using data from all sensors; (<b>b</b>) feature sensors.</p> Full article ">Figure 4
<p>(<b>a</b>) Linear discriminant analysis (LDA) results using data from all sensors; (<b>b</b>) feature sensors.</p> Full article ">Figure 5
<p>(<b>a</b>,<b>b</b>) K-nearest neighbor (KNN) results using data from all sensors; (<b>c</b>,<b>d</b>) feature sensors.</p> Full article ">Figure 6
<p>(<b>a</b>,<b>b</b>) Quantitative model results for all sensor versus characteristic sensor information using PLS; (<b>c</b>,<b>d</b>) Si-PLS; (<b>e</b>,<b>f</b>) GA-PLS; (<b>g</b>,<b>h</b>) CARS-PLS.</p> Full article ">Figure 6 Cont.
<p>(<b>a</b>,<b>b</b>) Quantitative model results for all sensor versus characteristic sensor information using PLS; (<b>c</b>,<b>d</b>) Si-PLS; (<b>e</b>,<b>f</b>) GA-PLS; (<b>g</b>,<b>h</b>) CARS-PLS.</p> Full article ">
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-354
Previous Issue
Next Issue
Back to TopTop