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

Next Issue
Volume 22, December-1
Previous Issue
Volume 22, November-1
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
sensors-logo

Journal Browser

Journal Browser

Sensors, Volume 22, Issue 22 (November-2 2022) – 443 articles

Cover Story (view full-size image): In this work, a measurement head for frequency mixing magnetic detection (FMMD) is introduced which involves the utilization of two ring-shaped permanent magnets to generate a static offset magnetic field. A steel cylinder in the ring bores homogenizes the field. Through variation of the distance between the ring magnets and of the thickness of the steel cylinder, the magnitude of the magnetic field at the sample position can be adjusted. Furthermore, the measurement setup is compared to the electromagnet offset module based on measured signals and temperature behavior. View this paper
  • Issues are regarded as officially published after their release is announced to the table of contents alert mailing list.
  • You may sign up for e-mail alerts to receive table of contents of newly released issues.
  • PDF is the official format for papers published in both, html and pdf forms. To view the papers in pdf format, click on the "PDF Full-text" link, and use the free Adobe Reader to open them.
Order results
Result details
Section
Select all
Export citation of selected articles as:
11 pages, 2512 KiB  
Article
A Study on Object Detection Performance of YOLOv4 for Autonomous Driving of Tram
by Joo Woo, Ji-Hyeon Baek, So-Hyeon Jo, Sun Young Kim and Jae-Hoon Jeong
Sensors 2022, 22(22), 9026; https://doi.org/10.3390/s22229026 - 21 Nov 2022
Cited by 13 | Viewed by 3502
Abstract
Recently, autonomous driving technology has been in the spotlight. However, autonomous driving is still in its infancy in the railway industry. In the case of railways, there are fewer control elements than autonomous driving of cars due to the characteristics of running on [...] Read more.
Recently, autonomous driving technology has been in the spotlight. However, autonomous driving is still in its infancy in the railway industry. In the case of railways, there are fewer control elements than autonomous driving of cars due to the characteristics of running on railways, but there is a disadvantage in that evasive maneuvers cannot be made in the event of a dangerous situation. In addition, when braking, it cannot be decelerated quickly for the weight of the body and the safety of the passengers. In the case of a tram, one of the railway systems, research has already been conducted on how to generate a profile that plans braking and acceleration as a base technology for autonomous driving, and to find the location coordinates of surrounding objects through object recognition. In pilot research about the tram’s automated driving, YOLOv3 was used for object detection to find object coordinates. YOLOv3 is an artificial intelligence model that finds coordinates, sizes, and classes of objects in an image. YOLOv3 is the third upgrade of YOLO, which is one of the most famous object detection technologies based on CNN. YOLO’s object detection performance is characterized by ordinary accuracy and fast speed. For this paper, we conducted a study to find out whether the object detection performance required for autonomous trams can be sufficiently implemented with the already developed object detection model. For this experiment, we used the YOLOv4 which is the fourth upgrade of YOLO. Full article
(This article belongs to the Special Issue Navigation Filters for Autonomous Vehicles)
Show Figures

Figure 1

Figure 1
<p>MS COCO dataset sample.</p>
Full article ">Figure 2
<p>BDD 100K dataset sample.</p>
Full article ">Figure 3
<p>The architecture of YOLOv4.</p>
Full article ">Figure 4
<p>Camera installed inside tram: (<b>a</b>) camera (acA2000-165uc, Basler ace), (<b>b</b>) installed camera, and (<b>c</b>) installed camera (enlarged).</p>
Full article ">Figure 5
<p>Loss graph while training YOLOv4 with BDD 100K dataset.</p>
Full article ">Figure 6
<p>mAP graph while training YOLOv4 with BDD 100K dataset.</p>
Full article ">Figure 7
<p>Experimental Video of Test Bed: (<b>a</b>) YOLOv4 Trained on BDD 100K Dataset (case1), (<b>b</b>) YOLOv4 Trained on BDD 100K Dataset (case2), (<b>c</b>) YOLOv4 Trained on MS COCO Dataset (case1), and (<b>d</b>) YOLOv4 Trained on MS COCO Dataset (case2).</p>
Full article ">
27 pages, 7444 KiB  
Article
An Experimental Assessment of People’s Location Efficiency Using Low-Energy Communications-Based Movement Tracking
by Marius Minea
Sensors 2022, 22(22), 9025; https://doi.org/10.3390/s22229025 - 21 Nov 2022
Cited by 2 | Viewed by 1860
Abstract
(1) Background: public transport demand dynamics represents important information for fleet managers and is also a key factor in making public transport attractive to reduce the environmental footprint of urban traffic. This research presents some experimental results on the assessment of low-energy communication [...] Read more.
(1) Background: public transport demand dynamics represents important information for fleet managers and is also a key factor in making public transport attractive to reduce the environmental footprint of urban traffic. This research presents some experimental results on the assessment of low-energy communication technologies, such as Wi-Fi and Bluetooth, as support for people density and/or movement tracking sensing technologies. (2) Methods: the research is based on field measurements to determine the percentage of discoverable devices carried by people, in relation to the total number of physical persons in interest, different scenarios of mobile devices usage and evaluation of influences on radio signals’ propagation, RSSI / RX read values, and efficiency of indoor localization, or in similar GPS-denied environments. Different situations are investigated, especially public transport-related ones, such as subway stations, indoors of commuting hubs, railway stations and trains. (3) Results: diagrams and experiments are presented, and models of signal behavior are also proposed. (4) Conclusions: recommendations on the efficiency of these non-conventional traveler and passenger flow tracking solutions and models are presented at the end of the paper. Full article
(This article belongs to the Special Issue Multi-Radio and/or Multi-Sensor Integrated Navigation System)
Show Figures

Figure 1

Figure 1
<p>Aspect of Station 1 test bed setup (red dots representing placement of human observers).</p>
Full article ">Figure 2
<p>Aspect of Station 2 test bed setup (red dots represent placement of human observers).</p>
Full article ">Figure 3
<p>Detected devices (represented in blue—Series 1) versus real number of persons (represented in orange—Series 2) present on platforms—no trains in station (Subway Station 1).</p>
Full article ">Figure 4
<p>Detected devices (represented in blue—Series 1) versus real number of persons (represented in orange—Series 2) present on platforms—no trains in station (Subway Station 2 subway station).</p>
Full article ">Figure 5
<p>Detected devices (represented in blue—Series 1) versus real number of persons (represented in orange—Series 2) present on platforms—one train in station (Subway Station 1).</p>
Full article ">Figure 6
<p>Detected devices (represented in blue—Series 1) versus real number of persons (represented in orange—Series 2) present on platforms—one train in station (Subway Station 2).</p>
Full article ">Figure 7
<p>Detected devices (represented in blue—Series 1) versus real number of persons (represented in orange—Series 2) present in trains (traveling in tunnel).</p>
Full article ">Figure 8
<p>Setup for RX-based location tests.</p>
Full article ">Figure 9
<p>Variation of signal strength with distance (<b>left</b>), and absolute error in distance measurement (<b>right</b>)—equipment Huawei AP, mixed indoor environment (LoS + NloS).</p>
Full article ">Figure 10
<p>Variation in received signal strength with distance (<b>left</b>) and absolute error in distance measurements (<b>right</b>), Xiaomi AP, mixed indoor environment (LoS + NLoS).</p>
Full article ">Figure 11
<p>Test bed setup for signal strength variation measurement.</p>
Full article ">Figure 12
<p>Distance to Wi-Fi AP computed from measured signal strength vs. real distance distribution, plotted in Weka environment. Colors represent belonging of distances to different clusters. Lower figure: Horizontal axis: computed distance, vertical axis: number of determinations.</p>
Full article ">Figure 13
<p>Weka Linear Regression classifier errors plotted for experimental results sequence 4 of tests (Vertical Axis—predicted values, Horizontal Axis—real distances). Colors represent belonging of errors to different clusters.</p>
Full article ">Figure 14
<p>Expectation-Maximization clustering of recorded measurements—Wi-Fi constancy of received signal strength information, LoS condition—colors represent grouping of recorded values performed by the EM algorithm (horizontal axis: distance to AP, vertical axis: RSS level).</p>
Full article ">Figure 15
<p>Test bed setup for BT signal strength indicator variation in stationary conditions—NLoS connection.</p>
Full article ">Figure 16
<p>Error distribution in RSSI-based prediction of distance to AP, NLoS scenario—behind an armored concrete wall—colors represent belonging of errors to different clusters, as computed by linear regression algorithm (Vertical axis: predicted distance, horizontal axis: real distance).</p>
Full article ">Figure 17
<p>Distance to BT device computed from measured RSSI, vs. real distance distribution, plotted in Weka (horizontal axis: distance [m], vertical axis: number of events). The colors represent belonging to the different clusters, as classified by Weka with linear regression.</p>
Full article ">Figure 18
<p>Variation of RTT (<b>left</b>) and Maximum deviation of recorded RTT (<b>right</b>) according to distance to AP, indoor, mixed path (0–8m LoS, 8–12m NLoS).</p>
Full article ">Figure 19
<p>Variation in signal strength for fixed position of mobile station, LoS, d = 1 m. Equipment: Huawei Mate 20 Lite (Histogram + errors in computed distance).</p>
Full article ">Figure 20
<p>Variation in signal strength for fixed position of mobile station, LoS, d = 1 m. Equipment: Samsung Galaxy A50 (Histogram + errors in computed distance).</p>
Full article ">Figure 21
<p>Variation in signal strength for fixed position of mobile station, LoS, d = 1 m. Equipment: Motorola Z3.</p>
Full article ">Figure 22
<p>Variation in signal strength for fixed position of mobile station, LoS, d = 1m. Equipment: Samsung Galaxy A12.</p>
Full article ">Figure 23
<p>Regularized histogram for Motorola Z3, stationary conditions, indoors scenario, 1 m distance to transmitter, LoS.</p>
Full article ">Figure 24
<p>Transitions diagram (Markov Chains model—based on Motorola Z3 measurements, indoor conditions, LoS, 1 m distance to transmitter).</p>
Full article ">Figure 25
<p>Dispersion of received signal strength recorded values during the test with BT source, LoS condition, 2 m from transmitter (horizontal axis—level [dBm], vertical axis: number of receptions). Colors in the above figure show belonging to time clusters when values were recorded.</p>
Full article ">Figure 26
<p>Dispersion of received signal strength recorded levels, with the BT source behind a wall of concrete (NLoS), 2 m distance (horizontal axis—level [dBm], vertical axis: number of receptions). The colors represent belonging to the different classified clusters.</p>
Full article ">
23 pages, 4364 KiB  
Article
SGGformer: Shifted Graph Convolutional Graph-Transformer for Traffic Prediction
by Shilin Pu, Liang Chu, Jincheng Hu, Shibo Li, Jihao Li and Wen Sun
Sensors 2022, 22(22), 9024; https://doi.org/10.3390/s22229024 - 21 Nov 2022
Viewed by 2381
Abstract
Accurate traffic prediction is significant in intelligent cities’ safe and stable development. However, due to the complex spatiotemporal correlation of traffic flow data, establishing an accurate traffic prediction model is still challenging. Aiming to meet the challenge, this paper proposes SGGformer, an advanced [...] Read more.
Accurate traffic prediction is significant in intelligent cities’ safe and stable development. However, due to the complex spatiotemporal correlation of traffic flow data, establishing an accurate traffic prediction model is still challenging. Aiming to meet the challenge, this paper proposes SGGformer, an advanced traffic grade prediction model which combines a shifted window operation, a multi-channel graph convolution network, and a graph Transformer network. Firstly, the shifted window operation is used for coarsening the time series data, thus, the computational complexity can be reduced. Then, a multi-channel graph convolutional network is adopted to capture and aggregate the spatial correlations of the roads in multiple dimensions. Finally, the improved graph Transformer based on the advanced Transformer model is proposed to extract the long-term temporal correlation of traffic data effectively. The prediction performance is evaluated by using actual traffic datasets, and the test results show that the SGGformer proposed exceeds the state-of-the-art baseline. Full article
(This article belongs to the Special Issue Sensing, Optimization, and Navigation on Vehicle Control)
Show Figures

Figure 1

Figure 1
<p>The topological adjacency matrix.</p>
Full article ">Figure 2
<p>Distribution of sample points under different features after SOM clustering.</p>
Full article ">Figure 3
<p>Distribution of traffic grades for 24 h of the day for the full graph in the test set.</p>
Full article ">Figure 4
<p>Overall architecture diagram of SSGformer. Historical traffic data is firstly divided into different segments by shifted window operation, and then multi-channel GCN operation is performed on different historical segments in parallel based on the topological adjacency matrix to obtain spatial correlation characteristics. Then, the Graph Transformer module is used to further complete the extraction of time-related features. Finally, the traffic grade of all regions is output in the whole graph corresponding to the time of practice <math display="inline"><semantics> <mrow> <mi>t</mi> <mo>+</mo> <mi>h</mi> </mrow> </semantics></math>. Specifically, the Chinese characters in the regional grade map represent the names of roads and landmark sites.</p>
Full article ">Figure 5
<p>The process of the shifted window operation.</p>
Full article ">Figure 6
<p>The process of spatial correlation modeling.</p>
Full article ">Figure 7
<p>The process of temporal correlation modeling.</p>
Full article ">Figure 8
<p>The calculation process of the high-dimensional self-attention.</p>
Full article ">Figure 9
<p>The raw trajectory data of the floating vehicles.</p>
Full article ">Figure 10
<p>The performance comparison among SGGformer and baselines.</p>
Full article ">Figure 11
<p>The grade difference under prediction length 1 h.</p>
Full article ">Figure 12
<p>The grade difference under prediction length 3 h.</p>
Full article ">Figure 13
<p>The grade difference under prediction length 6 h.</p>
Full article ">Figure 14
<p>Comparison of results of ablation experiments.</p>
Full article ">
25 pages, 832 KiB  
Article
Joint Optimization of Energy Consumption and Data Transmission in Smart Body Area Networks
by Limiao Li, Junyao Long, Wei Zhou, Alireza Jolfaei and Mohammad Sayad Haghighi
Sensors 2022, 22(22), 9023; https://doi.org/10.3390/s22229023 - 21 Nov 2022
Cited by 2 | Viewed by 2137
Abstract
In Wireless Body Area Networks (BAN), energy consumption, energy harvesting, and data communication are the three most important issues. In this paper, we develop an optimal allocation algorithm (OAA) for sensor devices, which are carried by or implanted in human body, harvest energy [...] Read more.
In Wireless Body Area Networks (BAN), energy consumption, energy harvesting, and data communication are the three most important issues. In this paper, we develop an optimal allocation algorithm (OAA) for sensor devices, which are carried by or implanted in human body, harvest energy from their surroundings, and are powered by batteries. Based on the optimal allocation algorithm that uses a two-timescale Lyapunov optimization approach, we design a framework for joint optimization of network service cost and network utility to study energy, communication, and allocation management at the network edge. Then, we formulate the utility maximization problem of network service cost management based on the framework. Specifically, we use OAA, which does not require prior knowledge of energy harvesting to decompose the problem into three subproblems: battery management, data collection amount control and transmission energy consumption control. We solve these through OAA to achieve three main goals: (1) balancing the cost of energy consumption and the cost of data transmission on the premise of minimizing the service cost of the devices; (2) keeping the balance of energy consumption and energy collection under the condition of stable queue; and (3) maximizing network utility of the device. The simulation results show that the proposed algorithm can actually optimize the network performance. Full article
Show Figures

Figure 1

Figure 1
<p>Common sensor node locations in BAN.</p>
Full article ">Figure 2
<p>Diagram of three sub problems and data interaction among them.</p>
Full article ">Figure 3
<p>Relationship between <span class="html-italic">V</span> and network utility.</p>
Full article ">Figure 4
<p>Relationship between <span class="html-italic">V</span> and network service cost.</p>
Full article ">Figure 5
<p>Dynamic change of energy queue under different <span class="html-italic">V</span> values.</p>
Full article ">Figure 6
<p>Dynamic change of the energy queue after amplification and corresponding sleep/wake decisions at <math display="inline"><semantics> <mrow> <mi>V</mi> <mo>=</mo> <mn>30</mn> </mrow> </semantics></math>.</p>
Full article ">Figure 7
<p>Dynamic change of data queue under different <span class="html-italic">V</span> values.</p>
Full article ">Figure 8
<p>Dynamic change of the data queue after amplification.</p>
Full article ">Figure 9
<p>The effect of parameter changes on network utility at different <span class="html-italic">V</span> values. (<b>a</b>) The relationship between network utility and wake probability; (<b>b</b>) The relationship between network utility and channel idle probability.</p>
Full article ">Figure 10
<p>The effect of parameter changes on network service cost at different <span class="html-italic">V</span> values. (<b>a</b>) The relationship between network service cost and maximum harvestable energy; (<b>b</b>) The relationship between network service cost and maximum transmission energy consumption.</p>
Full article ">
15 pages, 8466 KiB  
Article
A Wideband and Low-Power Distributed Cascode Mixer Using Inductive Feedback
by Jihoon Kim
Sensors 2022, 22(22), 9022; https://doi.org/10.3390/s22229022 - 21 Nov 2022
Cited by 3 | Viewed by 2022
Abstract
A wideband and low-power distributed cascode mixer is implemented for future mobile communications. The distributed design inspired by the distributed amplifier (DA) enables a mixer to operate in a wide band. In addition, the cascode structure and inductive positive feedback design allow high [...] Read more.
A wideband and low-power distributed cascode mixer is implemented for future mobile communications. The distributed design inspired by the distributed amplifier (DA) enables a mixer to operate in a wide band. In addition, the cascode structure and inductive positive feedback design allow high conversion gain with low-power consumption. The proposed mixer is fabricated using a 130 nm commercial complementary metal-oxide-semiconductor (CMOS) process. It consists of three cascode gain cells and operates with a drain voltage of 1.5 V and a gate voltage of 0.5 to 0.7 V. The fabricated mixer exhibits conversion gain of −2.9 to 3.1 dB at the radio frequencies (RFs) of 4 to 30 GHz and −1.9 to 0.4 dB at RFs of 54 to 66 GHz under the conditions of 8 to 10 dBm of local oscillator (LO) power and 650 MHz of intermediate frequency (IF). The LO-RF isolation is more than 15 dB over the entire measurement band (0.2 to 67 GHz) as the RF and LO signals are applied to different transistors owing to the cascode structure. The total power consumption is only within 12 mW, and the chip size is 0.056 mm2, making it possible to implement a compact mixer. The proposed mixer shows broadband characteristics covering from ultra-wideband (UWB) and the 28 GHz fifth-generation (5G) communication band to the 60 GHz wireless gigabit alliance (WiGig) band. Full article
(This article belongs to the Special Issue Integrated Circuits for Sensor Systems)
Show Figures

Figure 1

Figure 1
<p>Block diagram of a 5G antenna module.</p>
Full article ">Figure 2
<p>Frequency spectrum distribution in commercial use today.</p>
Full article ">Figure 3
<p>Simulated derivative of g<sub>m</sub> with respect to V<sub>ds</sub> according to V<sub>gs</sub> and V<sub>ds</sub>.</p>
Full article ">Figure 4
<p>(<b>a</b>) Circuit schematic of the cascode mixer and (<b>b</b>) DC-IV of each transistor constituting the cascode structure.</p>
Full article ">Figure 5
<p>Macro model equivalent circuit of RF CMOS used in this work.</p>
Full article ">Figure 6
<p>Comparison of (<b>a</b>) S11 and (<b>b</b>) S21 according to transistor (TR) size with current consumption.</p>
Full article ">Figure 7
<p>Simulated magnitudes of (<b>a</b>) S21 and (<b>b</b>) S11 in the total gate lines of a distributed cascode mixer versus RF according to gate line inductance.</p>
Full article ">Figure 8
<p>Transistor equivalent circuits of the cascode mixer when the first positive feedback inductor (L<sub>p1</sub>) is added to the gate node of the upper transistor ((<b>a</b>) two-port representation and (<b>b</b>) simplified one).</p>
Full article ">Figure 9
<p>Transistor equivalent circuit of the cascode mixer when the second positive feedback inductor (L<sub>p2</sub>) is added to the source-drain connection between M1 and M2 ((<b>a</b>) two-port representation and (<b>b</b>) simplified one).</p>
Full article ">Figure 10
<p>Simulated CG at high frequency according to (<b>a</b>) L<sub>p1</sub> and (<b>b</b>) L<sub>p2</sub>.</p>
Full article ">Figure 11
<p>Circuit schematic of distributed cascode mixer.</p>
Full article ">Figure 12
<p>Layout of distributed cascode mixer.</p>
Full article ">Figure 13
<p>Simulated (<b>a</b>) CG versus RF frequency and (<b>b</b>) return loss versus frequency of the distributed cascode mixer.</p>
Full article ">Figure 14
<p>(<b>a</b>) Fundamental signal (fund), IMD3 signal (third) and the difference between the two signals (imd) according to the power sweep of the two-tone RF signal at 20 GHz (<b>b</b>) simulated imd according to RF Frequency (RF power = −30 dBm).</p>
Full article ">Figure 15
<p>Chip photograph of the fabricated mixer (size: 0.8 × 0.7 mm<sup>2</sup>).</p>
Full article ">Figure 16
<p>Measurement setup of the fabricated mixer ((<b>a</b>) CG, (<b>b</b>) return loss and isolation).</p>
Full article ">Figure 17
<p>Measured CG versus RF (bias condition1: V<sub>DD</sub> = 1.5 V, V<sub>GG1</sub> = 0.6 V, V<sub>GG2</sub> = 0.7 V, bias condition2: V<sub>DD</sub> = 1.5 V, V<sub>GG1</sub> = 0.5 V, V<sub>GG2</sub> = 0.5 V, LO power = 8–10 dBm, RF power = −30 dBm, IF = 650 MHz).</p>
Full article ">Figure 18
<p>Measured RF and LO return loss under bias conditions of (<b>a</b>) V<sub>DD</sub> = 1.5 V, V<sub>GG1</sub> = 0.6 V, V<sub>GG2</sub> = 0.7 V and (<b>b</b>) V<sub>DD</sub> = 1.5 V, V<sub>GG1</sub> = 0.5 V, V<sub>GG2</sub> = 0.5 V.</p>
Full article ">Figure 19
<p>Measured LO-RF isolation under bias conditions of (<b>a</b>) V<sub>DD</sub> = 1.5 V, V<sub>GG1</sub> = 0.6 V, V<sub>GG2</sub> = 0.7 V and (<b>b</b>) V<sub>DD</sub> = 1.5 V, V<sub>GG1</sub> = 0.5 V, V<sub>GG2</sub> = 0.5 V.</p>
Full article ">
19 pages, 4816 KiB  
Article
A Dynamic Deployment Method of Security Services Based on Malicious Behavior Knowledge Base
by Qi Guo, Man Li, Weilin Wang and Ying Liu
Sensors 2022, 22(22), 9021; https://doi.org/10.3390/s22229021 - 21 Nov 2022
Cited by 3 | Viewed by 1876
Abstract
In view of various security requirements, there are various security services in the network. In particular, DDoS attacks have various types and detection methods. How to flexibly combine security services and make full use of the information provided by security services have become [...] Read more.
In view of various security requirements, there are various security services in the network. In particular, DDoS attacks have various types and detection methods. How to flexibly combine security services and make full use of the information provided by security services have become urgent problems to be solved. This paper combines the reasoning ability of the malicious behavior knowledge base to realize the dynamic deployment of the service function chain and dynamic configuration of the security service function. The method feeds back the information generated by the security service to the knowledge base. After the analysis of the knowledge base, the service function chain path and the security service configuration policies are generated, and these policies will be dynamically distributed to the security service function. Finally, security services can be dynamically arranged for different network traffic, realizing the coordinated use of various security services and improving the overall detection rate of the network. The experimental results show that by arranging the paths under the UDP and the TCP, the overall detection rate of the network can reach 99% and 88%, respectively, indicating that it has a good overall detection performance for multiple distributed denial of service (DDoS) attacks. Full article
(This article belongs to the Special Issue Security and Privacy for IoT Networks and the Mobile Internet)
Show Figures

Figure 1

Figure 1
<p>Block diagram of dynamic deployment of security services.</p>
Full article ">Figure 2
<p>Processing flow of malicious behavior knowledge base block diagram.</p>
Full article ">Figure 3
<p>Orchestration Layer Module.</p>
Full article ">Figure 4
<p>Data layer module.</p>
Full article ">Figure 5
<p>Advanced Configuration Policy.</p>
Full article ">Figure 6
<p>Advanced Path Policy Information Model.</p>
Full article ">Figure 7
<p>Information Model for Low-Level Path Policy.</p>
Full article ">Figure 8
<p>Information Model for Feedback of NSF Capability.</p>
Full article ">Figure 9
<p>The information model of the registration interface.</p>
Full article ">Figure 10
<p>Data Layer Security Function Dynamic Deployment Experimental Topology.</p>
Full article ">Figure 11
<p>Path policies under UDP and TCP.</p>
Full article ">Figure 12
<p>Classifier flow table.</p>
Full article ">Figure 13
<p>UDPs path and TCPs path.</p>
Full article ">Figure 14
<p>Path Delay Comparison.</p>
Full article ">Figure 15
<p>Data display of malicious behavior knowledge base. (<b>a</b>) Malicious traffic detection graph. (<b>b</b>) NSF Capability Graph.</p>
Full article ">Figure 16
<p>Advanced policy for blocking traffic. (<b>a</b>) Policy name and event. (<b>b</b>) Condition, actions and custom rules.</p>
Full article ">Figure 17
<p>Dynamically adjusting the path policy.</p>
Full article ">
20 pages, 1165 KiB  
Article
A Privacy-Preserving, Two-Party, Secure Computation Mechanism for Consensus-Based Peer-to-Peer Energy Trading in the Smart Grid
by Zhihu Li, Haiqing Xu, Feng Zhai, Bing Zhao, Meng Xu and Zhenwei Guo
Sensors 2022, 22(22), 9020; https://doi.org/10.3390/s22229020 - 21 Nov 2022
Cited by 5 | Viewed by 2310
Abstract
Consumers in electricity markets are becoming more proactive because of the rapid development of demand–response management and distributed energy resources, which boost the transformation of peer-to-peer (P2P) energy-trading mechanisms. However, in the P2P negotiation process, it is a challenging task to prevent private [...] Read more.
Consumers in electricity markets are becoming more proactive because of the rapid development of demand–response management and distributed energy resources, which boost the transformation of peer-to-peer (P2P) energy-trading mechanisms. However, in the P2P negotiation process, it is a challenging task to prevent private information from being attacked by malicious agents. In this paper, we propose a privacy-preserving, two-party, secure computation mechanism for consensus-based P2P energy trading. First, a novel P2P negotiation mechanism for energy trading is proposed based on the consensus + innovation (C + I) method and the power transfer distribution factor (PTDF), and this mechanism can simultaneously maximize social welfare and maintain physical network constraints. In addition, the C + I method only requires a minimum set of information to be exchanged. Then, we analyze the strategy of malicious neighboring agents colluding to attack in order to steal private information. To defend against this attack, we propose a two-party, secure computation mechanism in order to realize safe negotiation between each pair of prosumers based on Paillier homomorphic encryption (HE), a smart contract (SC), and zero-knowledge proof (ZKP). The energy price is updated in a safe way without leaking any private information. Finally, we simulate the functionality of the privacy-preserving mechanism in terms of convergence performance, computational efficiency, scalability, and SC operations. Full article
(This article belongs to the Special Issue Cryptographic Technologies for Securing Blockchain)
Show Figures

Figure 1

Figure 1
<p>P2P energy-trading market architecture.</p>
Full article ">Figure 2
<p>Blockchain-based P2P energy-trading market architecture.</p>
Full article ">Figure 3
<p>Collusion attack for malicious neighboring agents and two-party, secure computation between two agents.</p>
Full article ">Figure 4
<p>Convergence of the algorithm.</p>
Full article ">Figure 5
<p>Power flow in different lines under different line capacities.</p>
Full article ">Figure 6
<p>Encryption algorithm computation performance. (<b>a</b>) Agent encryption and decryption time. (<b>b</b>) The size of the public/private keys and ciphertext.</p>
Full article ">Figure 7
<p>Impact of number of agents on computation time and number of iterations for convergence.</p>
Full article ">Figure 8
<p>Test system schematic.</p>
Full article ">Figure 9
<p>Computational time of different P2P negotiation mechanisms.</p>
Full article ">Figure 10
<p>P2P energy-trading blockchain, smart contracts, and transactions results stored on the blockchain.</p>
Full article ">
34 pages, 1493 KiB  
Review
An Exploration of Recent Intelligent Image Analysis Techniques for Visual Pavement Surface Condition Assessment
by Waqar S. Qureshi, Syed Ibrahim Hassan, Susan McKeever, David Power, Brian Mulry, Kieran Feighan and Dympna O’Sullivan
Sensors 2022, 22(22), 9019; https://doi.org/10.3390/s22229019 - 21 Nov 2022
Cited by 13 | Viewed by 4921
Abstract
Road pavement condition assessment is essential for maintenance, asset management, and budgeting for pavement infrastructure. Countries allocate a substantial annual budget to maintain and improve local, regional, and national highways. Pavement condition is assessed by measuring several pavement characteristics such as roughness, surface [...] Read more.
Road pavement condition assessment is essential for maintenance, asset management, and budgeting for pavement infrastructure. Countries allocate a substantial annual budget to maintain and improve local, regional, and national highways. Pavement condition is assessed by measuring several pavement characteristics such as roughness, surface skid resistance, pavement strength, deflection, and visual surface distresses. Visual inspection identifies and quantifies surface distresses, and the condition is assessed using standard rating scales. This paper critically analyzes the research trends in the academic literature, professional practices and current commercial solutions for surface condition ratings by civil authorities. We observe that various surface condition rating systems exist, and each uses its own defined subset of pavement characteristics to evaluate pavement conditions. It is noted that automated visual sensing systems using intelligent algorithms can help reduce the cost and time required for assessing the condition of pavement infrastructure, especially for local and regional road networks. However, environmental factors, pavement types, and image collection devices are significant in this domain and lead to challenging variations. Commercial solutions for automatic pavement assessment with certain limitations exist. The topic is also a focus of academic research. More recently, academic research has pivoted toward deep learning, given that image data is now available in some form. However, research to automate pavement distress assessment often focuses on the regional pavement condition assessment standard that a country or state follows. We observe that the criteria a region adopts to make the evaluation depends on factors such as pavement construction type, type of road network in the area, flow and traffic, environmental conditions, and region’s economic situation. We summarized a list of publicly available datasets for distress detection and pavement condition assessment. We listed approaches focusing on crack segmentation and methods concentrating on distress detection and identification using object detection and classification. We segregated the recent academic literature in terms of the camera’s view and the dataset used, the year and country in which the work was published, the F1 score, and the architecture type. It is observed that the literature tends to focus more on distress identification (“presence/absence” detection) but less on distress quantification, which is essential for developing approaches for automated pavement rating. Full article
(This article belongs to the Special Issue Sensors for Smart Vehicle Applications)
Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Pavement condition rating process.</p>
Full article ">Figure 2
<p>Different forms of data acquisition for pavement condition assessment that can be adopted.</p>
Full article ">Figure 3
<p>Picture of a particular commercial vehicle, typical sensors attached for capturing pavement images, and output shown by the software. This image is of a customizable vehicle reproduced from the website: <a href="https://romdas.com/romdas-dataview.html" target="_blank">https://romdas.com/romdas-dataview.html</a> [<a href="#B54-sensors-22-09019" class="html-bibr">54</a>] (accessed on 15 November 2022).</p>
Full article ">
15 pages, 1028 KiB  
Article
Vehicular Environment Identification Based on Channel State Information and Deep Learning
by Soheyb Ribouh, Rahmad Sadli, Yassin Elhillali, Atika Rivenq and Abdenour Hadid
Sensors 2022, 22(22), 9018; https://doi.org/10.3390/s22229018 - 21 Nov 2022
Cited by 3 | Viewed by 2027
Abstract
This paper presents a novel vehicular environment identification approach based on deep learning. It consists of exploiting the vehicular wireless channel characteristics in the form of Channel State Information (CSI) in the receiver side of a connected vehicle in order to identify the [...] Read more.
This paper presents a novel vehicular environment identification approach based on deep learning. It consists of exploiting the vehicular wireless channel characteristics in the form of Channel State Information (CSI) in the receiver side of a connected vehicle in order to identify the environment type in which the vehicle is driving, without any need to implement specific sensors such as cameras or radars. We consider environment identification as a classification problem, and propose a new convolutional neural network (CNN) architecture to deal with it. The estimated CSI is used as the input feature to train the model. To perform the identification process, the model is targeted for implementation in an autonomous vehicle connected to a vehicular network (VN). The proposed model is extensively evaluated, showing that it can reliably recognize the surrounding environment with high accuracy (96.48%). Our model is compared to related approaches and state-of-the-art classification architectures. The experiments show that our proposed model yields favorable performance compared to all other considered methods. Full article
(This article belongs to the Topic Machine Learning in Communication Systems and Networks)
Show Figures

Figure 1

Figure 1
<p>Vehicular environment identification process.</p>
Full article ">Figure 2
<p>IEEE 802.11p PHY layer frame structure.</p>
Full article ">Figure 3
<p>Flow chart describing vehicular environment identification process.</p>
Full article ">Figure 4
<p>Proposed CNN Architecture.</p>
Full article ">Figure 5
<p>Confusion matrix based on LTS approach for the proposed CNN.</p>
Full article ">Figure 6
<p>Confusion matrix for ANN based on LTS approach.</p>
Full article ">Figure 7
<p>Confusion matrix for KNN based on LTS approach.</p>
Full article ">Figure 8
<p>Confusion matrix for RF based on LTS approach.</p>
Full article ">Figure 9
<p>Confusion matrix for GNB based on LTS approach.</p>
Full article ">Figure 10
<p>Confusion matrix for SVM based on LTS approach.</p>
Full article ">Figure 11
<p>Confusion matrix for the proposed CNN based on CSI approach.</p>
Full article ">Figure 12
<p>Comparison of our model’s accuracy to state-of-the-art alternatives.</p>
Full article ">
9 pages, 5456 KiB  
Article
Flat Multi-Wavelength Brillouin Erbium-Doped Fiber Laser Based on a Sagnac Loop for High-Sensitivity Sensor
by Liang Chen, Jian He and Yi Liu
Sensors 2022, 22(22), 9017; https://doi.org/10.3390/s22229017 - 21 Nov 2022
Cited by 1 | Viewed by 2341
Abstract
We have demonstrated the use of a flat multi-wavelength Brillouin erbium-doped fiber laser (MWBEFL) based on a Sagnac loop with an unpumped erbium-doped fiber (Un-EDF) as a high-sensitivity sensor. A Sagnac loop with a Un-EDF was used as power equalizer to achieve multi-wavelength [...] Read more.
We have demonstrated the use of a flat multi-wavelength Brillouin erbium-doped fiber laser (MWBEFL) based on a Sagnac loop with an unpumped erbium-doped fiber (Un-EDF) as a high-sensitivity sensor. A Sagnac loop with a Un-EDF was used as power equalizer to achieve multi-wavelength power flatness by adjusting the birefringence beat length properly. In the experiments, the best result obtained in terms of Brillouin Stokes lines and output power flatness was ±0.315 dB and the optical signal-to-noise ratio (OSNR) was 18.97 dB within a 33 nm bandwidth range from 1532.0 nm to 1565.0 nm. The flatness of the 33 nm bandwidth range varied from ±0.315 dB to ±1.38 dB and the average OSNR was about 17.51 dB. The peak power values of Brillouin Stokes lines observed under different wavelengths were extremely close and their range of fluctuation was about ±0.37 dB. These experimental results were close to our previous experimental values obtained using a passive Sagnac loop with a Un-EDF. The flat range covering almost the entire C-band has broad application prospects in high-sensitivity distributed optical fiber sensing and wavelength-division multiplexing. Full article
(This article belongs to the Special Issue Chip-Based MEMS Platforms)
Show Figures

Figure 1

Figure 1
<p>Schematic diagram of the flat MWBEFL based on the use of a Sagnac loop with a Un-EDF. WDM: wavelength-division multiplexer. TLS: tunable laser source. PC: polarization controller. OC: optical coupler. Cir: three-port circulator. OSA: optical spectrum analyzer.</p>
Full article ">Figure 2
<p>Experimental setup of the gain-flattened EDFA with a Un-EDF.</p>
Full article ">Figure 3
<p>Comparison diagram of ASE gain spectra before flattening and after flattening at the reflection port (<b>a</b>) and at the transmission port (<b>b</b>).</p>
Full article ">Figure 4
<p>(<b>a</b>) ASE and self-excited light without the use of the TLS pump; (<b>b</b>) the output spectrum obtained when the 980 nm pump power was 54 mWl (<b>c</b>) the output spectrum obtained when the TLS pump power was 14 dBm and the 980 nm pump power was 500 mW.</p>
Full article ">Figure 5
<p>Output spectra of different TLS wavelengths. (<b>a</b>) TLS wavelength = 1532 nm, (<b>b</b>) TLS wavelength = 1542.5 nm, (<b>c</b>) TLS wavelength = 1547 nm, (<b>d</b>) TLS wavelength = 1553 nm, (<b>e</b>) TLS wavelength = 1559 nm, (<b>f</b>) TLS wavelength = 1565 nm.</p>
Full article ">Figure 6
<p>Flat output spectra of different TLS wavelengths in the 33 nm range.</p>
Full article ">Figure 7
<p>Output power flatness at different TLS wavelengths and the number of Brillouin Stokes lines in the entire flat range.</p>
Full article ">Figure 8
<p>Brillouin Stokes lines’ average OSNR and peak power at different TLS wavelengths.</p>
Full article ">
13 pages, 1240 KiB  
Article
Realization of Crowded Pipes Climbing Locomotion of Snake Robot Using Hybrid Force–Position Control Method
by Yongdong Wang and Tetsushi Kamegawa
Sensors 2022, 22(22), 9016; https://doi.org/10.3390/s22229016 - 21 Nov 2022
Cited by 1 | Viewed by 2391
Abstract
The movement capabilities of snake robots allow them to be applied in a variety of applications. We realized a snake robot climbing in crowded pipes. In this paper, we implement a sinusoidal curve control method that allows the snake robot to move faster. [...] Read more.
The movement capabilities of snake robots allow them to be applied in a variety of applications. We realized a snake robot climbing in crowded pipes. In this paper, we implement a sinusoidal curve control method that allows the snake robot to move faster. The control method is composed of a hybrid force–position controller that allows the snake robot to move more stably. We conducted experiments to confirm the effectiveness of the proposed method. The experimental results show that the proposed method is stable and effective compared to the previous control method that we had implemented in the snake robot. Full article
(This article belongs to the Special Issue Advances in Snake Robots of Bio-Inspired Robotics)
Show Figures

Figure 1

Figure 1
<p>Structure of the snake robot model (<b>upper</b>) and the snake robot for the experiment (<b>bottom</b>) [<a href="#B14-sensors-22-09016" class="html-bibr">14</a>].</p>
Full article ">Figure 2
<p>Structure of the crowded pipes model (<b>left</b>) and the crowded pipes environment for the experiment (<b>right</b>) [<a href="#B14-sensors-22-09016" class="html-bibr">14</a>].</p>
Full article ">Figure 3
<p>Block diagram of the system dynamics of the snake robot joint.</p>
Full article ">Figure 4
<p>System structure of the snake robot [<a href="#B14-sensors-22-09016" class="html-bibr">14</a>].</p>
Full article ">Figure 5
<p>Comparison of travel distances (<b>left graph</b>) and load magnitude (<b>right graph</b>) of control methods for crawling crowded pipes upward using the snake robot. Ten trials were conducted individually for each control method. Hybrid control achieved reaching whole distance in all ten trials.</p>
Full article ">Figure 6
<p>Comparison of travel distances (<b>left graph</b>) and load magnitude (<b>right graph</b>) of control methods for crawling crowded pipes downward using the snake robot. Ten trials were conducted individually for each control method. Hybrid control achieved reaching whole distance in all ten trials.</p>
Full article ">Figure 7
<p>Commands and feedback data from the robot’s 3rd pitch joint are used for the trial of the pure position controller.</p>
Full article ">Figure 8
<p>Commands and feedback data from the robot’s 3rd pitch joint are used for the trial of the admittance controller.</p>
Full article ">Figure 9
<p>Commands and feedback data from the robot’s 3rd pitch joint are used for the trial of the impedance controller. The input angle difference was scaled 10 times for easier viewing.</p>
Full article ">Figure 10
<p>Commands and feedback data from the robot’s 3rd pitch joint are used for the trial of the hybrid controller. The input angle difference was scaled 10 times for easier viewing.</p>
Full article ">Figure 11
<p>Velocity comparison of 10 upward motion experiments and 10 downward motion experiments.</p>
Full article ">Figure 12
<p>Experimental results of upward motion using hybrid force–position control method.</p>
Full article ">Figure 13
<p>Experimental results of downward motion using hybrid force–position control method.</p>
Full article ">
19 pages, 3791 KiB  
Article
Removing False Targets for Cyclic Prefixed OFDM Sensing with Extended Ranging
by Kai Wu, J. Andrew Zhang, Xiaojing Huang and Y. Jay Guo
Sensors 2022, 22(22), 9015; https://doi.org/10.3390/s22229015 - 21 Nov 2022
Viewed by 2136
Abstract
Employing a cyclic prefixed OFDM (CP-OFDM) communication waveform for sensing has attracted extensive attention in vehicular integrated sensing and communications (ISAC). A unified sensing framework was developed recently, enabling CP-OFDM sensing to surpass the conventional limits imposed by underlying communications. However, a false [...] Read more.
Employing a cyclic prefixed OFDM (CP-OFDM) communication waveform for sensing has attracted extensive attention in vehicular integrated sensing and communications (ISAC). A unified sensing framework was developed recently, enabling CP-OFDM sensing to surpass the conventional limits imposed by underlying communications. However, a false target issue still remains unsolved. In this paper, we investigate and solve this issue. Specifically, we unveil that false targets are caused by periodic cyclic prefixes (CPs) in CP-OFDM waveforms. We also derive the relation between the locations of false and true targets, and other features, e.g., strength, of false targets. Moreover, we develop an effective solution to remove false targets. Simulations are provided to confirm the validity of our analysis and the effectiveness of the proposed solution. In particular, our design can reduce the false alarm rate caused by false targets by over 50% compared with the prior art. Full article
(This article belongs to the Special Issue Integrated Sensing and Communications for Next-Generation Networks)
Show Figures

Figure 1

Figure 1
<p>Illustrating the RDMs of UFS (<b>a</b>) and COS (<b>b</b>). Please note that the symbol “-” before all numbers, as automatically generated by MATLAB during plotting the figure, denotes the negative sign.</p>
Full article ">Figure 2
<p>Doppler (<b>a</b>) and range (<b>b</b>) cuts of the RDM obtained by COS, as plotted in <a href="#sensors-22-09015-f001" class="html-fig">Figure 1</a>b. Doppler (<b>c</b>) and range (<b>d</b>) cuts of the RDM obtained by USF, as plotted in <a href="#sensors-22-09015-f001" class="html-fig">Figure 1</a>a. Please note that the symbol “-” before all numbers, as automatically generated by MATLAB during plotting the figure, denotes the negative sign.</p>
Full article ">Figure 3
<p>(<b>a</b>) The block diagram for the solution developed in <a href="#sec5dot1-sensors-22-09015" class="html-sec">Section 5.1</a>; (<b>b</b>) the block diagram for the method designed in <a href="#sec5dot2-sensors-22-09015" class="html-sec">Section 5.2</a>.</p>
Full article ">Figure 4
<p>RDM of Algorithm 3 in (<b>a</b>); RDM of Algorithm 4 in (<b>b</b>). A single target is set here. Please note that the symbol “-” before all numbers, as automatically generated by MATLAB during plotting the figure, denotes the negative sign.</p>
Full article ">Figure 5
<p>(<b>a</b>) Doppler cuts of the RDMs shown in <a href="#sensors-22-09015-f004" class="html-fig">Figure 4</a>; (<b>b</b>) range cuts of the same RDMs. Note that Algorithm 1 as developed in [<a href="#B32-sensors-22-09015" class="html-bibr">32</a>], is simulated as a benchmark. It suffers from the false target issue. In contrast, the new algorithms, Algorithms 3 and 4, are able to remove the false targets. Please note that the symbol “-” before all numbers, as automatically generated by MATLAB during plotting the figure, denotes the negative sign.</p>
Full article ">Figure 6
<p>RDM of COS in (<b>a</b>); RDM of Algorithm 1 in (<b>b</b>); RDM of Algorithm 3 in (<b>c</b>). Four targets are set, as highlighted in the figures. Please note that the symbol “-” before all numbers, as automatically generated by MATLAB during plotting the figure, denotes the negative sign.</p>
Full article ">Figure 7
<p>RDMs obtained by Algorithm 4 with four targets simulated. The RDMs from the first three iterations in the algorithms are given in (<b>a</b>–<b>c</b>), respectively. Please note that the symbol “-” before all numbers, as automatically generated by MATLAB during plotting the figure, denotes the negative sign.</p>
Full article ">Figure 8
<p>Detecting performance versus SNR (with respect to <math display="inline"><semantics> <mi mathvariant="bold">y</mi> </semantics></math>), where a single target is set with randomly generated parameters over <math display="inline"><semantics> <msup> <mn>10</mn> <mn>4</mn> </msup> </semantics></math> independent trials. The detecting probability is given in (<b>a</b>) and the false alarm rate is given in (<b>b</b>). Note that Algorithms 1 and 2, as developed in [<a href="#B21-sensors-22-09015" class="html-bibr">21</a>,<a href="#B32-sensors-22-09015" class="html-bibr">32</a>], respectively, are simulated as benchmark methods. Additionally note that Algorithm 2 [<a href="#B21-sensors-22-09015" class="html-bibr">21</a>] cannot properly sense a target with the echo delay over the CP duration, and hence its detecting probability is much lower than our algorithms. Please note that the symbol “-” before all numbers, as automatically generated by MATLAB during plotting the figure, denotes the negative sign.</p>
Full article ">Figure 9
<p>Detecting performance versus SNR (with respect to <math display="inline"><semantics> <mi mathvariant="bold">y</mi> </semantics></math>), where four targets are set with randomly generated parameters over <math display="inline"><semantics> <msup> <mn>10</mn> <mn>4</mn> </msup> </semantics></math> independent trials. The detecting probability is given in (<b>a</b>) and the false alarm rate is given in (<b>b</b>). Please note that the symbol “-” before all numbers, as automatically generated by MATLAB during plotting the figure, denotes the negative sign.</p>
Full article ">
10 pages, 2030 KiB  
Article
Demonstration of Spatial Modulation Using a Novel Active Transmitter Detection Scheme with Signal Space Diversity in Optical Wireless Communications
by Tingting Song, Ampalavanapillai Nirmalathas and Christina Lim
Sensors 2022, 22(22), 9014; https://doi.org/10.3390/s22229014 - 21 Nov 2022
Cited by 2 | Viewed by 1583
Abstract
Line-of-sight (LOS) indoor optical wireless communications (OWC) enable a high data rate transmission while potentially suffering from optical channel obstructions. Additional LOS links using diversity techniques can tackle the received signal performance degradation, where channel gains often differ in multiple LOS channels. In [...] Read more.
Line-of-sight (LOS) indoor optical wireless communications (OWC) enable a high data rate transmission while potentially suffering from optical channel obstructions. Additional LOS links using diversity techniques can tackle the received signal performance degradation, where channel gains often differ in multiple LOS channels. In this paper, a novel active transmitter detection scheme in spatial modulation (SM) is proposed to be incorporated with signal space diversity (SSD) technique to enable an increased OWC system throughput with an improved bit-error-rate (BER). This transmitter detection scheme is composed of a signal pre-distortion technique at the transmitter and a power-based statistical detection method at the receiver, which can address the problem of power-based transmitter detection in SM using carrierless amplitude and phase modulation waveforms with numerous signal levels. Experimental results show that, with the proposed transmitter detection scheme, SSD can be effectively provided with ~0.61 dB signal-to-noise-ratio (SNR) improvement. Additionally, an improved data rate ~7.5 Gbit/s is expected due to effective transmitter detection in SM. The SSD performances at different constellation rotation angles and under different channel gain distributions are also investigated, respectively. The proposed scheme provides a practical solution to implement power-based SM and thus aids the SSD realization for improving system performance. Full article
Show Figures

Figure 1

Figure 1
<p>The architecture of the overall indoor OWC system.</p>
Full article ">Figure 2
<p>Signal pre-distortion technique. (<b>a</b>) Signal flow for pre-distortion. (<b>b</b>) Example of the generation of a pre-distorted CAP waveform.</p>
Full article ">Figure 3
<p>Experimental setup. (<b>a</b>) Schematic diagram. (<b>b</b>) Photo.</p>
Full article ">Figure 4
<p>BER of the SM with SSD using the proposed novel active transmitter detection scheme. (<b>a</b>) ideal detection vs. the proposed detection. (<b>b</b>) Different channel gain distributions. (<b>c</b>) Different suppress ratios <span class="html-italic">τ</span>.</p>
Full article ">
13 pages, 5137 KiB  
Article
Low-Cost Nitric Oxide Sensors: Assessment of Temperature and Humidity Effects
by Steven Owen, Lachlan H. Yee and Damien T. Maher
Sensors 2022, 22(22), 9013; https://doi.org/10.3390/s22229013 - 21 Nov 2022
Cited by 2 | Viewed by 2173
Abstract
High equipment cost is a significant entry barrier to research for small organizations in developing solutions to air pollution problems. Low-cost electrochemical sensors show sensitivity at parts-per-billion by volume mixing ratios but are subject to variation due to changing environmental conditions, in particular [...] Read more.
High equipment cost is a significant entry barrier to research for small organizations in developing solutions to air pollution problems. Low-cost electrochemical sensors show sensitivity at parts-per-billion by volume mixing ratios but are subject to variation due to changing environmental conditions, in particular temperature. In this study, we demonstrate a low-cost Internet of Things (IoT)-based sensor system for nitric oxide analysis. The sensor system used a four-electrode electrochemical sensor exposed to a series of isothermal/isohume conditions. When deployed under these conditions, stable baseline responses were achieved, in contrast to ambient air conditions where temperature and humidity conditions may be variable. The interrelationship between working and auxiliary electrodes was linear within an environmental envelope of 20–40 °C and 30–80% relative humidity, with correlation coefficients from 0.9980 to 0.9999 when measured under isothermal/isohume conditions. These data enabled the determination of surface functions that describe the working to auxiliary electrode offsets and calibration curve gradients and intercepts. The linear and reproducible nature of individual calibration curves for stepwise nitric oxide (NO) additions under isothermal/isohume environments suggests the suitability of these sensors for applications aside from their role in air quality monitoring. Such applications would include nitric oxide kinetic studies for atmospheric applications or measurement of the potential biocatalytic activity of nitric oxide consuming enzymes in biocatalytic coatings, both of which currently employ high-capital-cost chemiluminescence detectors. Full article
(This article belongs to the Section Electronic Sensors)
Show Figures

Figure 1

Figure 1
<p>(<b>a</b>) Environmental chamber constructed from a Chemglass CG-1929-29 5-L double-walled reaction vessel fitted with a CG-1941 4-neck reaction vessel lid. (<b>b</b>) Low-cost electrochemical sensor showing Alphasense sensors attached to South Coast Science DFE.</p>
Full article ">Figure 2
<p>Stable electrochemical responses under isothermal/isohume conditions.</p>
Full article ">Figure 3
<p>Working electrode responses to humidity transient.</p>
Full article ">Figure 4
<p>Temporal changes in NO electrochemical electrode responses due to changing temperature (dT = ±1.2 °C).</p>
Full article ">Figure 5
<p>Temperature-dependence of NO electrochemical electrode responses (dT = ±1.2 °C).</p>
Full article ">Figure 6
<p>Interrelationship between NO electrodes to minor temperature fluctuations (dT = ±1.2 °C).</p>
Full article ">Figure 7
<p>(<b>a</b>) MATLAB surface function of offset between working to auxiliary electrode across an environmental envelope. (<b>b</b>) Residuals plot of actual values to surface function.</p>
Full article ">Figure 8
<p>MATLAB surface function describing the gradients (<b>a</b>) and intercepts (<b>b</b>) resulting from the calibration curves measured across an environmental envelope.</p>
Full article ">Figure 9
<p>MATLAB residuals plots resulting from the calibration curves measured across an environmental envelope for, (<b>a</b>) Gradient and (<b>b</b>) Intercept.</p>
Full article ">Figure 10
<p>Calculated NO concentrations for three sensors under four different environmental conditions using the initially measured calibration curves, (<b>a</b>)—Original sensor, (<b>c</b>)—Sensor 2, (<b>e</b>)—(Sensor 3) and using the NO working electrode to NO auxiliary electrode offset, gradient, and intercept surface functions from the initially measured calibration curves (<b>b</b>)—Original sensor, (<b>d</b>)—Sensor 2, (<b>f</b>)—(Sensor 3).</p>
Full article ">
9 pages, 17362 KiB  
Article
Fabrication and Characterization of Novel Silk Fiber-Optic SERS Sensor with Uniform Assembly of Gold Nanoparticles
by Taeyoung Kang, Yongjun Cho, Kyeong Min Yuk, Chan Yeong Yu, Seung Ho Choi and Kyung Min Byun
Sensors 2022, 22(22), 9012; https://doi.org/10.3390/s22229012 - 21 Nov 2022
Cited by 3 | Viewed by 2300
Abstract
Biocompatible optical fibers and waveguides are gaining attention as promising platforms for implantable biophotonic devices. Recently, the distinct properties of silk fibroin were extensively explored because of its unique advantages, including flexibility, process compatibility, long-term biosafety, and controllable biodegradability for in vitro and [...] Read more.
Biocompatible optical fibers and waveguides are gaining attention as promising platforms for implantable biophotonic devices. Recently, the distinct properties of silk fibroin were extensively explored because of its unique advantages, including flexibility, process compatibility, long-term biosafety, and controllable biodegradability for in vitro and in vivo biomedical applications. In this study, we developed a novel silk fiber for a sensitive optical sensor based on surface-enhanced Raman spectroscopy (SERS). In contrast to conventional plasmonic nanostructures, which employ expensive and time-consuming fabrication processes, gold nanoparticles were uniformly patterned on the top surface of the fiber employing a simple and cost-effective convective self-assembly technique. The fabricated silk fiber-optic SERS probe presented a good performance in terms of detection limit, sensitivity, and linearity. In particular, the uniform pattern of gold nanoparticles contributed to a highly linear sensing feature compared to the commercial multi-mode fiber sample with an irregular and aggregated distribution of gold nanoparticles. Through further optimization, silk-based fiber-optic probes can function as useful tools for highly sensitive, cost-effective, and easily tailored biophotonic platforms, thereby offering new capabilities for future implantable SERS devices. Full article
(This article belongs to the Section Optical Sensors)
Show Figures

Figure 1

Figure 1
<p>Fabrication process of silk optical fibers. (<b>a</b>) Preparing <span class="html-italic">Bombyx mori</span> cocoons. (<b>b</b>) Cutting. (<b>c</b>) Degumming. (<b>d</b>) Removing sericin. (<b>e</b>) Drying. (<b>f</b>) Dissolving in LiBr. (<b>g</b>) Dialysis. (<b>h</b>) Freeze-drying. (<b>i</b>) Dissolving in HFIP. (<b>j</b>) Wet spinning. (<b>k</b>) Cladding coating. (<b>l</b>) Photo of the silk optical fiber.</p>
Full article ">Figure 2
<p>Characteristics of the silk fiber-optic probe. (<b>a</b>) The outer diameter of the silk core and cladding layers. (<b>b</b>) Refractive index measure for core and cladding structures made of silk fibroin proteins. The dotted line at λ = 785 nm indicates the condition for the SERS experiment. (<b>c</b>) Optical transmittance test results for silk fibers with and without a cladding layer and multi-mode fibers. The fiber length was fixed at 3 cm.</p>
Full article ">Figure 3
<p>Silk-based fiber-optic SERS probe with gold nanoparticles (<b>a</b>). A schematic diagram of the SERS measurement setup and FE-SEM images of gold nanoparticles distributed on top of (<b>b</b>) a silk fiber and (<b>c</b>) a multi-mode fiber are presented. Histograms of the cluster size distribution were computed for (<b>d</b>) the silk and (<b>e</b>) the silica multi-mode fibers, respectively.</p>
Full article ">Figure 4
<p>Measured SERS signals of 4-ABT molecules when laser light is incident to the fiber-top surface directly (Measure A) and when the light is coupled through the core/cladding structure (Measure B) for (<b>a</b>) a silk fiber and (<b>b</b>) a multi-mode fiber. The gray bar indicates the primary peak of 4-ABT molecules. The fiber length was fixed as 1 cm for both samples.</p>
Full article ">Figure 5
<p>SERS experiments for a varied concentration of 4-ABT molecules. (<b>a</b>) SERS measurement data of 4-ABT from 1 µM to 10 mM and (<b>b</b>) Linear regression analysis (the red line) of the intensity at 1076 cm<sup>−1</sup> for silk fiber samples. (<b>c</b>) SERS measurement data of 4-ABT and (<b>d</b>) Linear regression analysis (the red line) of the intensity at 1076 cm<sup>−1</sup> for multi-mode fiber samples.</p>
Full article ">
13 pages, 447 KiB  
Article
Leveraging Self-Attention Mechanism for Attitude Estimation in Smartphones
by James Brotchie, Wei Shao, Wenchao Li and Allison Kealy
Sensors 2022, 22(22), 9011; https://doi.org/10.3390/s22229011 - 21 Nov 2022
Cited by 3 | Viewed by 2389
Abstract
Inertial attitude estimation is a crucial component of many modern systems and applications. Attitude estimation from commercial-grade inertial sensors has been the subject of an abundance of research in recent years due to the proliferation of Inertial Measurement Units (IMUs) in mobile devices, [...] Read more.
Inertial attitude estimation is a crucial component of many modern systems and applications. Attitude estimation from commercial-grade inertial sensors has been the subject of an abundance of research in recent years due to the proliferation of Inertial Measurement Units (IMUs) in mobile devices, such as the smartphone. Traditional methodologies involve probabilistic, iterative-state estimation; however, these approaches do not generalise well over changing motion dynamics and environmental conditions, as they require context-specific parameter tuning. In this work, we explore novel methods for attitude estimation from low-cost inertial sensors using a self-attention-based neural network, the Attformer. This paper proposes to part ways from the traditional cycle of continuous integration algorithms, and formulate it as an optimisation problem. This approach separates itself by leveraging attention operations to learn the complex patterns and dynamics associated with inertial data, allowing for the linear complexity in the dimension of the feature vector to account for these patterns. Additionally, we look at combining traditional state-of-the-art approaches with our self-attention method. These models were evaluated on entirely unseen sequences, over a range of different activities, users and devices, and compared with a recent alternate deep learning approach, the unscented Kalman filter and the iOS CoreMotion API. The inbuilt iOS had a mean angular distance from the true attitude of 117.31, the GRU 21.90, the UKF 16.38, the Attformer 16.28 and, finally, the UKF–Attformer had mean angular distance of 10.86. We show that this plug-and-play solution outperforms previous approaches and generalises well across different users, devices and activities. Full article
(This article belongs to the Special Issue Feature Papers in Physical Sensors 2022)
Show Figures

Figure 1

Figure 1
<p>Attformer/UKF–Attformer Structure. The Attformer was trained solely with the input features of the raw three-axis measurements from the accelerometer (Equation (<a href="#FD3-sensors-22-09011" class="html-disp-formula">3</a>)), gyroscope (Equation (<a href="#FD1-sensors-22-09011" class="html-disp-formula">1</a>)) and magnetometer (Equation (<a href="#FD4-sensors-22-09011" class="html-disp-formula">4</a>)). The UKF–Attformer was trained with the additional input feature of the prior UKF attitude estimate from <a href="#sec4dot3dot2-sensors-22-09011" class="html-sec">Section 4.3.2</a>.</p>
Full article ">Figure 2
<p>Structure of the 2-Layer GRU with 200 neurons per layer.</p>
Full article ">
22 pages, 6751 KiB  
Article
Dynamic Phase and Polarization Modulation Using Two-Beam Parallel Coding for Optical Storage in Transparent Materials
by Jintao Hong, Jin Li and Daping Chu
Sensors 2022, 22(22), 9010; https://doi.org/10.3390/s22229010 - 21 Nov 2022
Cited by 2 | Viewed by 2497
Abstract
In this paper, we propose and experimentally demonstrate a parallel coding and two-beam combining approach for the simultaneous implementation of dynamically generating holographic patterns at their arbitrary linear polarization states. Two orthogonal input beams are parallelly and independently encoded with the same target [...] Read more.
In this paper, we propose and experimentally demonstrate a parallel coding and two-beam combining approach for the simultaneous implementation of dynamically generating holographic patterns at their arbitrary linear polarization states. Two orthogonal input beams are parallelly and independently encoded with the same target image information but there is different amplitude information by using two-phase computer-generated holograms (CGH) on two Liquid-Crystal-on-Silicon-Spatial-Light Modulators (LCOS SLMs). Two modulated beams are then considered as two polarization components and are spatially superposed to form the target polarization state. The final linear vector beam is created by the spatial superposition of the two base beams, capable of controlling the vector angle through the phase depth of the phase-only CGHs. Meanwhile, the combined holographic patterns can be freely encoded by the holograms of two vector components. Thus, this allows us to tailor the optical fields endowed with arbitrary holographic patterns and the linear polarization states at the same time. This method provides a more promising approach for laser data writing generation systems in the next-generation optical data storage technology in transparent materials. Full article
(This article belongs to the Section Optical Sensors)
Show Figures

Figure 1

Figure 1
<p>Schematic of the proposed method using two-beam parallel coding in combination. The information of the target image and amplitude are independently encoded onto the input beams to form the target polarization state.</p>
Full article ">Figure 2
<p>The principle of the proposed method using a tunable optical vector, where LP is a linear polarizer, LC is liquid crystals, HWP is a half-wave plate, BS is a beam splitter, and BC is a beam combiner.</p>
Full article ">Figure 3
<p>An example of the different polarization states determined by the amplitude ratio and phase shift between two polarization components.</p>
Full article ">Figure 4
<p>The partial phase profile of one hologram under three different phase depths.</p>
Full article ">Figure 5
<p>The partial phase profile of the hologram with three different grey levels displayed on the active area of an LCOS SLM: a target image (<b>a</b>), the corresponding CGH hologram (<b>b</b>), a selected area of the hologram (<b>c</b>), and the grey levels (<b>d</b>).</p>
Full article ">Figure 6
<p>Phase profile of a blazed phase stepped grating on an LCOS SLM with period d, maximum phase depth M2π, and N steps.</p>
Full article ">Figure 7
<p>The intensity of different diffraction orders (<math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>1</mn> <mo>,</mo> <mtext> </mtext> <mn>2</mn> <mo>,</mo> <mtext> </mtext> <mn>3</mn> </mrow> </semantics></math>) versus the phase depth <span class="html-italic">M</span> of a continuous phase grating (<math display="inline"><semantics> <mrow> <mi>N</mi> <mo>→</mo> <mo>∞</mo> </mrow> </semantics></math>).</p>
Full article ">Figure 8
<p>Schematic of the experimental setup, where LP is a linear polarizer, HWP is a half-wave plate, BS is a beam splitter, BC is a beam combiner, and SF is a spatial filter.</p>
Full article ">Figure 9
<p>Schematic of the proposed method using spatial superposition of two holographic beams with orthogonal polarization states. The output of combining beams will have a 45° polarization angle when the amplitude ratio of two input beams is 1:1.</p>
Full article ">Figure 10
<p>When two basis vector beams have an alignment error, two orthogonal vector beams will have a small angle at the image reconstruction plane. Due to the high coherence of two beams, they will spatially interfere. Therefore, the interference fringes can be observed under an analyzer.</p>
Full article ">Figure 11
<p>Interference patterns of two orthogonal vector beams with the same shape are observed when the analyzer is placed at 45° (<b>a</b>) and −45° (<b>b</b>). The combined beam includes two original polarization states (<b>c</b>). (<b>d</b>,<b>e</b>) show the superposition of two orthogonal vector beams with different shapes (circle and cross) that is observed under different polarization analyzer angles (blue double-end arrow). Interference patterns are observed when the analyzer is placed at 45° (<b>f</b>).</p>
Full article ">Figure 12
<p>Experimental results (<b>a</b>–<b>h</b>) of unparallel two vector beams with the same shape combining in free space. The width-of-stripes pattern varies with the degree of parallelization of the two beams.</p>
Full article ">Figure 13
<p>The two orthogonal vector beams are completely parallel after the high-accuracy aligning process and the interference pattern disappears due to the combination of two beams.</p>
Full article ">Figure 14
<p>The two orthogonal vector beams are completely aligned. The superposition area of two beams has a single polarization state. (<b>a</b>) is the result when the polarizer is placed at 45°, (<b>b</b>) is the result when the polarizer is placed at –45°, and (<b>c</b>) is the polarization state of the combined beams.</p>
Full article ">Figure 15
<p>The 3D plots of the intensity of images provide a comparison of the overlapped area and area of input beams under an analyzer at 45°. (<b>a</b>,<b>b</b>) and (<b>c</b>,<b>d</b>) show the corresponding 3D intensity and 2D intensity, respectively.</p>
Full article ">Figure 16
<p>Holographic images with a linear polarization state of 30°: recorded holographic images when the analyzer is set at different angles (<b>a</b>), the average intensity of four holographic images (<b>b</b>), the 3D intensity distribution at 30° (<b>c</b>), and the 3D intensity distribution at 0° (<b>d</b>).</p>
Full article ">Figure 17
<p>Holographic images with a linear polarization state of 45°: recorded holographic images when the analyzer is set at different angles (<b>a</b>), the average intensity of four holographic images (<b>b</b>), the 3D intensity distribution at 45° (<b>c</b>), and the 3D intensity distribution at 0° (<b>d</b>).</p>
Full article ">Figure 18
<p>Holographic images with a linear polarization state of 60°: recorded holographic images when the analyzer is set at different angles (<b>a</b>), the average intensity of four holographic images (<b>b</b>), the 3D intensity distribution at 60° (<b>c</b>), and the 3D intensity distribution at 90° (<b>d</b>).</p>
Full article ">
16 pages, 2357 KiB  
Article
Effects of Dynamic IMU-to-Segment Misalignment Error on 3-DOF Knee Angle Estimation in Walking and Running
by Chao Jiang, Yan Yang, Huayun Mao, Dewei Yang and Wei Wang
Sensors 2022, 22(22), 9009; https://doi.org/10.3390/s22229009 - 21 Nov 2022
Cited by 3 | Viewed by 2212
Abstract
The inertial measurement unit (IMU)-to-segment (I2S) alignment is an important part of IMU-based joint angle estimation, and the accurate estimation of the three degree of freedom (3-DOF) knee angle can provide practical support for the evaluation of motions. In this paper, we introduce [...] Read more.
The inertial measurement unit (IMU)-to-segment (I2S) alignment is an important part of IMU-based joint angle estimation, and the accurate estimation of the three degree of freedom (3-DOF) knee angle can provide practical support for the evaluation of motions. In this paper, we introduce a dynamic weight particle swarm optimization (DPSO) algorithm with crossover factor based on the joint constraint to obtain the dynamic alignment vectors of I2S, and use them to perform the quaternion-based 3-DOF knee angle estimation algorithm. The optimization algorithm and the joint angle estimation algorithm were evaluated by comparing with the optical motion capture system. The range of 3-DOF knee angle root mean square errors (RMSEs) is 1.6°–5.9° during different motions. Furthermore, we also set up experiments of human walking (3 km/h), jogging (6 km/h) and ordinary running (9 km/h) to investigate the effects of dynamic I2S misalignment errors on 3-DOF knee angle estimation during different motions by artificially adding errors to I2S alignment parameters. The results showed differences in the effects of I2S misalignment errors on the estimation of knee abduction, internal rotation and flexion, which indicate the differences in knee joint kinematics among different motions. The IMU to thigh misalignment error has the greatest effect on the estimation of knee internal rotation. The effect of IMU to thigh misalignment error on the estimation of knee abduction angle becomes smaller and then larger during the two processes of switching from walking to jogging and then speeding up to ordinary running. The effect of IMU to shank misalignment error on the estimation of knee flexion angle is numerically the largest, while the standard deviation (SD) is the smallest. This study can provide support for future research on the accuracy of 3-DOF knee angle estimation during different motions. Full article
(This article belongs to the Section Biomedical Sensors)
Show Figures

Figure 1

Figure 1
<p>Illustrations of joint constraints and I2S alignment. The schematic diagram of each vector in the joint constraint model is shown with colored arrows; the coordinates of IMU and segment are shown with different colors; the black dashed arrows indicate the alignment of IMU to segment.</p>
Full article ">Figure 2
<p>Flow chart of introduced optimization algorithm.</p>
Full article ">Figure 3
<p>The experimental scenario mainly includes an optical motion capture system, IMUs and a treadmill; six markers and two IMUs are set on each leg, three markers on each segment are not co-planar and the direction of IMUs is not required.</p>
Full article ">Figure 4
<p>Convergence curves of three algorithms under different combinations of particle swarm size and iteration number, N denotes the particle swarm size and D denotes the maximum number of iterations, the red, blue and green lines represent the convergence curves of the classical PSO, DPSO and our introduced algorithm, respectively.</p>
Full article ">Figure 5
<p>Effects of IMU to thigh misalignment error on 3-DOF knee angle estimation during different motions. The blue, green, purple and orange lines indicate the knee angle estimates after adding errors of −5°, +5°, −10° and +10° to the I2S alignment parameters, and the red line indicates the knee angle estimates without artificially added errors. (<b>a</b>–<b>c</b>), (<b>d</b>–<b>f</b>), (<b>g</b>–<b>i</b>) are the abduction, internal rotation and flexion angles of the knee during walking (3 km/h), jogging (6 km/h) and ordinary running (9 km/h) on treadmill. For example, in (<b>a</b>), 9.8 ± 5.2° denote that the abduction mean error is 9.8° and the SD is 5.2° when the introduced IMU to thigh misalignment error is +10° during the walking trial.</p>
Full article ">Figure 6
<p>Effects of IMU to shank misalignment error on 3-DOF knee angle estimation during different motions; (<b>a</b>–<b>c</b>), (<b>d</b>–<b>f</b>), (<b>g</b>–<b>i</b>) indicate the abduction, internal rotation and flexion angle of the knee during walking (3 km/h), jogging (6 km/h) and ordinary running (9 km/h), when added IMU to shank misalignment error from −10° to +10° in steps of 5°.</p>
Full article ">Figure 7
<p>Effects of both IMU to thigh and shank misalignment error on 3-DOF knee angle estimation during different motions; (<b>a</b>–<b>c</b>), (<b>d</b>–<b>f</b>), (<b>g</b>–<b>i</b>) indicate the abduction, internal rotation and flexion angles of the knee during walking (3 km/h), jogging (6 km/h) and ordinary running (9 km/h), when added IMU to thigh and shank misalignment error from −10° to +10° in steps of 5°.</p>
Full article ">Figure 8
<p>IMU to segment misalignment of all subjects. The bar and error bar represent the mean and standard deviation (SD) of the misalignment root mean square errors (RMSEs). (<b>a</b>–<b>c</b>) indicate the misalignment RMSEs of IMU to thigh misalignment, IMU to shank misalignment and IMU to thigh and shank misalignment, during different motions.</p>
Full article ">Figure 9
<p>Comparison of IMU-based and optical motion capture system data-based 3-DOF knee angle estimations of RMSEs during different motions, the bar and error bar represent the mean and SD of the RMSEs.</p>
Full article ">
10 pages, 648 KiB  
Article
Sensor Location Matters When Estimating Player Workload for Baseball Pitching
by Cristine Agresta, Michael T. Freehill, Jessica Zendler, Georgia Giblin and Stephen Cain
Sensors 2022, 22(22), 9008; https://doi.org/10.3390/s22229008 - 21 Nov 2022
Cited by 2 | Viewed by 2345
Abstract
Estimating external workload in baseball pitchers is important for training and rehabilitation. Since current methods of estimating workload through pitch counts and rest days have only been marginally successful, clubs are looking for more sophisticated methods to quantify the mechanical loads experienced by [...] Read more.
Estimating external workload in baseball pitchers is important for training and rehabilitation. Since current methods of estimating workload through pitch counts and rest days have only been marginally successful, clubs are looking for more sophisticated methods to quantify the mechanical loads experienced by pitchers. Among these are the use of wearable systems. While wearables offer a promising solution, there remains a lack of standards or guidelines for how best to employ these devices. As a result, sensor location and workload calculation methods vary from system to system. This can influence workload estimates and blur their interpretation and utility when making decisions about training or returning to sport. The primary purpose of this study was to determine the extent to which sensor location influences workload estimate. A secondary purpose was to compare estimates using different workload calculations. Acceleration data from three sensor locations—trunk, throwing upper arm, and throwing forearm—were collected from ten collegiate pitchers as they threw a series of pitches during a single bullpen session. The effect of sensor location and pitch type was assessed in relation to four different workload estimates. Sensor location significantly influenced workload estimates. Workload estimates calculated from the forearm sensor were significantly different across pitch types. Whole-body workload measured from a trunk-mounted sensor may not adequately reflect the mechanical loads experienced at throwing arm segments. A sensor on the forearm was the most sensitive to differences in workloads across pitch types, regardless of the calculation method. Full article
Show Figures

Figure 1

Figure 1
<p>Illustration of sensor placement on the forearm, upper arm, trunk, and feet. The trunk sensor was placed on the sternum at the level of the xyphoid process, the upper arm sensor was placed at the posterolateral portion of the distal one-third of the segment, and the forearm sensor was placed at the mid-point of the segment. Placement location on each segment was determined by comfort for the pitcher and the location that appeared to produce minimal migration due to muscle contraction across multiple pitches. Arm and foot sensors were secured with flexible adhesive tape, while the trunk sensor was held in place by a snug chest strap.</p>
Full article ">Figure 2
<p>Mean workload estimates across pitch type and sensor location. Red indicates values from sternum sensor; Green indicates values from upper arm sensor; and Blue indicates values from forearm sensor. Workload estimates varied by calculation method and did not show the same pattern across pitch types.</p>
Full article ">
17 pages, 2128 KiB  
Article
A Novel Joint Adversarial Domain Adaptation Method for Rotary Machine Fault Diagnosis under Different Working Conditions
by Xiaoping Zhao, Fan Shao and Yonghong Zhang
Sensors 2022, 22(22), 9007; https://doi.org/10.3390/s22229007 - 21 Nov 2022
Cited by 6 | Viewed by 1897
Abstract
In real-world applications of detecting faults, many factors—such as changes in working conditions, equipment wear, and environmental causes—can cause a significant mismatch between the source domain on which classifiers are trained and the target domain to which those classifiers are applied. As such, [...] Read more.
In real-world applications of detecting faults, many factors—such as changes in working conditions, equipment wear, and environmental causes—can cause a significant mismatch between the source domain on which classifiers are trained and the target domain to which those classifiers are applied. As such, existing deep network algorithms perform poorly under different working conditions. To solve this problem, we propose a novel fault diagnosis method named Joint Adversarial Domain Adaptation (JADA) for fault detection under different working conditions. Our approach simultaneously aligns marginal distribution and conditional distribution across the source and target through a unified adversarial learning process. JADA aims to construct domain-invariant and category-discriminative feature representation that is effective and robust for substantial distribution difference caused by working conditions. We also introduce a supervision signal, namely center loss, that penalizes the distances between the deep features and their corresponding class centers. This makes the learned features better equipped with more discriminative structures and effectively prevents mode collapse. Twenty-four transfer fault diagnosis tasks based on two experimental platforms were conducted to evaluate the effectiveness of the proposed methods. Extensive experiments verified that the JADA can significantly outperform several popular methods under different transfer diagnosis tasks. Full article
Show Figures

Figure 1

Figure 1
<p>Fault diagnosis process of JADA.</p>
Full article ">Figure 2
<p>Feature extractor and classifier model.</p>
Full article ">Figure 3
<p>Drivetrain diagnostics simulator.</p>
Full article ">Figure 4
<p>Four bearing health conditions.</p>
Full article ">Figure 5
<p>Diagnosis accuracies for the DDS dataset, respectively, achieved by (<b>a</b>) models with different <math display="inline"><semantics> <mi>κ</mi> </semantics></math> and fixed <math display="inline"><semantics> <mrow> <mi>λ</mi> <mo>=</mo> <mn>0.5</mn> </mrow> </semantics></math>. (<b>b</b>) models with different <math display="inline"><semantics> <mi>λ</mi> </semantics></math> and fixed <math display="inline"><semantics> <mrow> <mi>K</mi> <mo>=</mo> <mn>5</mn> <mo>×</mo> <msup> <mn>10</mn> <mrow> <mo>−</mo> <mn>3</mn> </mrow> </msup> </mrow> </semantics></math>.</p>
Full article ">Figure 6
<p>Classification accuracies of the different methods for the DDS dataset.</p>
Full article ">Figure 7
<p>Confusion matrices of different methods (<b>a</b>) DANN, (<b>b</b>) ADDA, and (<b>c</b>) JADA.</p>
Full article ">Figure 8
<p>Feature visualization of the different methods for the DDS dataset: (<b>a</b>) DANN, (<b>b</b>) ADDA, and (<b>c</b>) JADA.</p>
Full article ">Figure 9
<p>Experimental setup of motor bearing.</p>
Full article ">Figure 10
<p>Classification accuracies of the different methods for the CWRU dataset.</p>
Full article ">Figure 11
<p>Feature visualization of different methods for the CWRU dataset: (<b>a</b>) DANN, (<b>b</b>) ADDA, and (<b>c</b>) JADA.</p>
Full article ">
14 pages, 5819 KiB  
Article
Acoustic and Thermal Characterization of Therapeutic Ultrasonic Langevin Transducers under Continuous- and Pulsed Wave Excitations
by Jinhyuk Kim and Jungwoo Lee
Sensors 2022, 22(22), 9006; https://doi.org/10.3390/s22229006 - 21 Nov 2022
Cited by 4 | Viewed by 2086
Abstract
We previously conducted an empirical study on Langevin type transducers in medical use by examining the heat effect on porcine tissue. For maximum acoustic output, the transducer was activated by a continuous sinusoidal wave. In this work, pulsed waves with various duty factors [...] Read more.
We previously conducted an empirical study on Langevin type transducers in medical use by examining the heat effect on porcine tissue. For maximum acoustic output, the transducer was activated by a continuous sinusoidal wave. In this work, pulsed waves with various duty factors were applied to our transducer model in order to examine their effect on functionality. Acoustic power, electro-acoustic conversion efficiency, acoustic pressure, thermal effect on porcine tissue and bovine muscle, and heat generation in the transducer were investigated under various input conditions. For example, the results of applying a continuous wave of 200 VPP and a pulse wave of 70% duty factor with the same amplitude to the transducer were compared. It was found that continuous waves generated 9.79 W of acoustic power, 6.40% energy efficiency, and 24.84 kPa acoustic pressure. In pulsed excitation, the corresponding values were 9.04 W, 8.44%, and 24.7 kPa, respectively. The maximum temperature increases in bovine muscle are reported to be 83.0 °C and 89.5 °C for each waveform, whereas these values were 102.5 °C and 84.5 °C in fatty porcine tissue. Moreover, the heat generation around the transducer was monitored under continuous and pulsed modes and was found to be 51.3 °C and 50.4 °C. This shows that pulsed excitation gives rise to less thermal influence on the transducer. As a result, it is demonstrated that a transducer triggered by pulsed waves improves the energy efficiency and provides sufficient thermal impact on biological tissues by selecting proper electrical excitation types. Full article
(This article belongs to the Special Issue Ultrasound-Based Sensors for Physical Therapy Applications)
Show Figures

Figure 1

Figure 1
<p>Transducer layout (<b>a</b>) cross-sectional view (<b>b</b>) actual image of the transducer.</p>
Full article ">Figure 2
<p>Description of waves (<b>a</b>) Continuous wave (<b>b</b>) Pulsed wave.</p>
Full article ">Figure 3
<p>Acoustic power measurement: (<b>a</b>) whole experimental setup; and (<b>b</b>) zero-point setup of power meter.</p>
Full article ">Figure 4
<p>Hydrophone system: (<b>a</b>) water bath equipped with a motorized stage; and (<b>b</b>) transducer rod and hydrophone.</p>
Full article ">Figure 5
<p>Measurement of temperature rise in tissue: (<b>a</b>) porcine fat; and (<b>b</b>) bovine muscle.</p>
Full article ">Figure 6
<p>Comparison of frequency response between the experiment and the simulation.</p>
Full article ">Figure 7
<p>Acoustic power depending on CW and PW conditions: (<b>a</b>) 100 V<sub>PP</sub>; and (<b>b</b>) 200 V<sub>PP</sub>.</p>
Full article ">Figure 8
<p>Emitted power from the transducer by CW and PW with 60% duty factor: (<b>a</b>) 125 V<sub>PP</sub>; and (<b>b</b>) 150 V<sub>PP</sub>.</p>
Full article ">Figure 9
<p>Calibration data of acoustic pressure under CW and PW with a 50% duty factor: (<b>a</b>) 100 V<sub>PP</sub>; (<b>b</b>) 200 V<sub>PP</sub>; and (<b>c</b>) comparison of peak-to-peak pressure with standard deviation.</p>
Full article ">Figure 10
<p>Acoustic pressure under various PW conditions.</p>
Full article ">Figure 11
<p>Comparison of temperature increases induced by different electrical conditions: (<b>a</b>) 150 V<sub>PP</sub>; and (<b>b</b>) 175 V<sub>PP</sub>.</p>
Full article ">Figure 12
<p>Temperature changes in bovine muscle as a function of various electrical driving conditions. (<b>a</b>) CW (<b>b</b>) PW.</p>
Full article ">Figure 13
<p>Temperature rises in porcine tissue over time; (<b>a</b>) CW; and (<b>b</b>) PW.</p>
Full article ">Figure 14
<p>Heat generation in the piezoelectric element under CW and PW excitations. Top row: bovine muscle (<b>a</b>) with CW 175 VPP and (<b>b</b>) with PW 175 VPP. Bottom row: porcine fat (<b>c</b>) with CW 200 VPP and (<b>d</b>) PW 200 VPP.</p>
Full article ">
10 pages, 2893 KiB  
Communication
Ion-Imprinted Chitosan-Based Localized Surface Plasmon Resonance Sensor for Ni2+ Detection
by Xiujuan Zhong, Li Ma and Guolu Yin
Sensors 2022, 22(22), 9005; https://doi.org/10.3390/s22229005 - 21 Nov 2022
Cited by 5 | Viewed by 1947
Abstract
Heavy metals are important sources of environmental pollution and cause disease in organisms throughout the food chain. A localized surface plasmon resonance sensor was proposed and demonstrated to realize Ni2+ detection by using ion-imprinted chitosan. Au nanoparticles were coated on the multimode [...] Read more.
Heavy metals are important sources of environmental pollution and cause disease in organisms throughout the food chain. A localized surface plasmon resonance sensor was proposed and demonstrated to realize Ni2+ detection by using ion-imprinted chitosan. Au nanoparticles were coated on the multimode fiber to excite the local surface plasmon resonance, and Ni2+-imprinted chitosan was then functionalized by using the dip coating technique. Ethylene diamine tetra-acetic acid was used to release the Ni2+ ions and hence form countless voids. Ni2+ was refilled into the voids to increase the refractive index of the sensing material, thus realizing the measurement of Ni2+ by monitoring the wavelength shift in the localized surface plasmon resonant peak. The coating thickness of the Ni2+–chitosan gel was optimized to obtain greater sensitivity. Experimental results show that the proposed Ni2+ sensor has a sensitivity of 185 pm/μM, and the limit of detection is 0.512 μM. The comparison experiments indicated that the ion-imprinted chitosan has better selectivity than pure chitosan. Full article
(This article belongs to the Special Issue Optical Fiber Sensors for Chemical and Biomedical Applications)
Show Figures

Figure 1

Figure 1
<p>Functionalization of the proposed LSPR sensor based on Ni<sup>2+</sup>-imprinted chitosan.</p>
Full article ">Figure 2
<p>Experimental setup of the proposed Ni<sup>2+</sup> sensor based on LSPR. MMF: Multi-mode fiber; AuNPs: Au nanoparticles; Ni<sup>2+</sup>-imp-CS: Ni<sup>2+</sup> imprinted chitosan.</p>
Full article ">Figure 3
<p>SEM micrographs of (<b>a</b>) Au-nanoparticle-coated optical fiber and (<b>b</b>) four layers of chitosan-gel-coated optical fiber.</p>
Full article ">Figure 4
<p>Absorption spectra of the LSPR sensor (<b>a</b>) in the preparing process and (<b>b</b>) with different layer chitosan gels.</p>
Full article ">Figure 5
<p>SEM micrographs of (<b>a</b>) the Ni<sup>2+</sup>–chitosan film and (<b>b</b>) the film after EDTA treatment.</p>
Full article ">Figure 6
<p>Refractive index sensing of the LSPR sensor with difference thickness of the Ni<sup>2+</sup>–CS gel. Absorption spectra of the LSPR sensor with (<b>a</b>) one layer of Ni<sup>2+</sup>–CS gel and (<b>b</b>) four layers of Ni<sup>2+</sup>–CS gel. (<b>c</b>) Comparison of the wavelength shift.</p>
Full article ">Figure 7
<p>Ni<sup>2+</sup> sensing based on the proposed LSPR sensor. (<b>a</b>) Absorption spectra with different concentrations of Ni<sup>2+</sup> ions. (<b>b</b>) Wavelength shift as a function of the concentration of Ni<sup>2+</sup> ions.</p>
Full article ">Figure 8
<p>Selectivity of the LSPR sensor with (<b>a</b>) CS gel and (<b>b</b>) Ni<sup>2+</sup>–CS gel when the ion concentration is 100 μM.</p>
Full article ">Figure 9
<p>Absorption spectra of the LSPR sensor when an EDTA solution is used for releasing Ni<sup>2+</sup> ions for six cycles.</p>
Full article ">
11 pages, 3232 KiB  
Article
Wideband Versatile Receiver for CubeSat Microwave Front-Ends
by Emanuele Cardillo, Renato Cananzi and Paolo Vita
Sensors 2022, 22(22), 9004; https://doi.org/10.3390/s22229004 - 21 Nov 2022
Cited by 4 | Viewed by 2028
Abstract
One of the main features of CubeSats is represented by their extreme versatility, e.g., maintaining the same overall structure for different purposes. This requires high technological flexibility achievable in a cost-effective way while maintaining compact sizes. In this contribution, a microwave receiver specifically [...] Read more.
One of the main features of CubeSats is represented by their extreme versatility, e.g., maintaining the same overall structure for different purposes. This requires high technological flexibility achievable in a cost-effective way while maintaining compact sizes. In this contribution, a microwave receiver specifically designed for CubeSat applications is proposed. Due to the wide input operating bandwidth, i.e., 2 GHz–18 GHz, it can be exploited for different purposes, e.g., satellite communication, radars, and electronic warfare systems. This is beneficial for CubeSat systems, whereby the possibility to share the same front-end circuit for different purposes is a key feature in reducing the overall size and weight. The downconverter was designed to minimize the spurious contributions at low frequency by taking advantage, at the same time, of commercial off-the-shelf components due to their cost-effectiveness. The idea behind this work is to add flexibility to the CubeSat communication systems in order to be reusable in different contexts. This feature enables new applications but also provides the largest bandwidth if required from the ground system. An accurate experimental characterization was performed to validate the downconverter performance with the aim of allowing easy system integration for the new frontier of CubeSat technologies. This paves the way for the most effective implementation of the Internet of Things (IoT), machine-to-machine (M2M) communications, and smart-everything services. Full article
(This article belongs to the Special Issue Sensors and Satellite Network Systems)
Show Figures

Figure 1

Figure 1
<p>Schematic of the proposed microwave receiver.</p>
Full article ">Figure 2
<p>Frequency spectrum at the output of the frequency doubler MMD 1030HS. The input signal was centered at 5.5 GHz.</p>
Full article ">Figure 3
<p>Measured S<sub>21</sub> (black solid line), S<sub>11</sub> (blue solid line), and S<sub>22</sub> (red solid line) of the APM-6849 amplifier.</p>
Full article ">Figure 4
<p>Measured S<sub>21</sub> (black solid line), S<sub>11</sub> (blue solid line), and S<sub>22</sub> (red solid line) of the APM-6702 amplifier.</p>
Full article ">Figure 5
<p>(<b>a</b>) Schematic and (<b>b</b>) insertion loss of the measured (red solid line) and simulated (blue dashed line) S<sub>21</sub> for the PB1182WB filter.</p>
Full article ">Figure 6
<p>Measured S<sub>21</sub> (black solid line), S<sub>11</sub> (blue solid line), and S<sub>22</sub> (red solid line) of the (<b>a</b>) ZFL-1500VH+ amplifier and (<b>b</b>) VLF-1500+ filter.</p>
Full article ">Figure 7
<p>Power budget of the receiver. The input power is equal to −20 dBm.</p>
Full article ">Figure 8
<p>Picture of the microwave receiver.</p>
Full article ">Figure 9
<p>(<b>a</b>) Simulated and (<b>b</b>) measured frequency spectrum at the output of the receiver for an input signal with a center frequency and power equal to 2 GHz and −30 dBm, respectively.</p>
Full article ">Figure 10
<p>The measured conversion gain of the receiver. The input power is equal to −30 dBm.</p>
Full article ">
20 pages, 4054 KiB  
Article
A Reference Architecture for Cloud–Edge Meta-Operating Systems Enabling Cross-Domain, Data-Intensive, ML-Assisted Applications: Architectural Overview and Key Concepts
by Panagiotis Trakadas, Xavi Masip-Bruin, Federico M. Facca, Sotirios T. Spantideas, Anastasios E. Giannopoulos, Nikolaos C. Kapsalis, Rui Martins, Enrica Bosani, Joan Ramon, Raül González Prats, George Ntroulias and Dimitrios V. Lyridis
Sensors 2022, 22(22), 9003; https://doi.org/10.3390/s22229003 - 21 Nov 2022
Cited by 18 | Viewed by 3624
Abstract
Future data-intensive intelligent applications are required to traverse across the cloud-to-edge-to-IoT continuum, where cloud and edge resources elegantly coordinate, alongside sensor networks and data. However, current technical solutions can only partially handle the data outburst associated with the IoT proliferation experienced in recent [...] Read more.
Future data-intensive intelligent applications are required to traverse across the cloud-to-edge-to-IoT continuum, where cloud and edge resources elegantly coordinate, alongside sensor networks and data. However, current technical solutions can only partially handle the data outburst associated with the IoT proliferation experienced in recent years, mainly due to their hierarchical architectures. In this context, this paper presents a reference architecture of a meta-operating system (RAMOS), targeted to enable a dynamic, distributed and trusted continuum which will be capable of facilitating the next-generation smart applications at the edge. RAMOS is domain-agnostic, capable of supporting heterogeneous devices in various network environments. Furthermore, the proposed architecture possesses the ability to place the data at the origin in a secure and trusted manner. Based on a layered structure, the building blocks of RAMOS are thoroughly described, and the interconnection and coordination between them is fully presented. Furthermore, illustration of how the proposed reference architecture and its characteristics could fit in potential key industrial and societal applications, which in the future will require more power at the edge, is provided in five practical scenarios, focusing on the distributed intelligence and privacy preservation principles promoted by RAMOS, as well as the concept of environmental footprint minimization. Finally, the business potential of an open edge ecosystem and the societal impacts of climate net neutrality are also illustrated. Full article
(This article belongs to the Section Intelligent Sensors)
Show Figures

Figure 1

Figure 1
<p>Constellation of RAMOS nodes towards a peer-to-peer continuum.</p>
Full article ">Figure 2
<p>High-level meta-operating system architecture.</p>
Full article ">Figure 3
<p>Detailed layered architecture of Atoms and Molecules at component-level.</p>
Full article ">Figure 4
<p>RAMOS components in diverse domain scenarios, illustrating the different orchestration, communication and data sharing patterns.</p>
Full article ">Figure 5
<p>Resources and model sharing in the RAMOS architecture amongst the individual vehicles’ OBUs in the green driving scenario.</p>
Full article ">Figure 6
<p>The interconnection of NILM ML model and data sharing in the RAMOS framework, targeting to provide suggestions towards energy footprint reduction at households.</p>
Full article ">Figure 7
<p>Centralized ML port model and distributed ML vessel models coordinating in the RAMOS platform towards increasing the port efficiency and port call optimization, while also reducing greenhouse gas emissions.</p>
Full article ">Figure 8
<p>Implementation of the RAMOS layered architecture in data and ML model sharing amongst several factory sites towards carbon-neutral manufacturing.</p>
Full article ">Figure 9
<p>High-level overview of an EV charging system, where RES are integrated with the conventional power grid.</p>
Full article ">
16 pages, 410 KiB  
Article
Treatment of Extended Kalman Filter Implementations for the Gyroless Star Tracker
by Joshua J. R. Critchley-Marrows, Xiaofeng Wu and Iver H. Cairns
Sensors 2022, 22(22), 9002; https://doi.org/10.3390/s22229002 - 21 Nov 2022
Cited by 6 | Viewed by 2521
Abstract
The literature since Apollo contains exhaustive material on attitude filtering, usually treating the problem of two sensors, a combination of state measuring and inertial devices. More recently, it has become popular for a sole attitude determination device to be considered. This is especially [...] Read more.
The literature since Apollo contains exhaustive material on attitude filtering, usually treating the problem of two sensors, a combination of state measuring and inertial devices. More recently, it has become popular for a sole attitude determination device to be considered. This is especially the case for a star tracker given its unbiased stellar measurement and recent improvements in optical sensor performance. The state device indirectly estimates the attitude rate using a known dynamic model. In estimation theory, two main attitude filtering approaches are classified, the additive and the multiplicative. Each refers to the nature of the quaternion update in the filter. In this article, these two techniques are implemented for the case of a sole star tracker, using simulated and real night sky image data. Both sets of results are presented and compared with each other, with a baseline established through a basic linear least square estimate. The state approach is more accurate and precise for measuring angular velocity than using the error-based filter. However, no discernible difference is observed between each technique for determining pointing. These results are important not only for sole device attitude determination systems, but also for space situational awareness object localisation, where attitude and rate estimate accuracy are highly important. Full article
(This article belongs to the Section Physical Sensors)
Show Figures

Figure 1

Figure 1
<p>Rotation of an arbitrary vector <math display="inline"><semantics> <mi mathvariant="bold">x</mi> </semantics></math> about the rotation vector [<a href="#B15-sensors-22-09002" class="html-bibr">15</a>].</p>
Full article ">Figure 2
<p>Attitude measurement performance for the simulation operating at a constant speed of 36 arc-sec/s.</p>
Full article ">Figure 3
<p>Illustration of zenith facing camera for static image testing using Earth rotation. Angular movement of stars through camera field of view caused solely by Earth rotation.</p>
Full article ">Figure 4
<p>Attitude measurement performance for the standard time step of 1 s, considering each approach.</p>
Full article ">
12 pages, 2663 KiB  
Article
Feedback Regimes of LFI Sensors: Experimental Investigations
by Karl Bertling, Xiaoqiong Qi, Thomas Taimre, Yah Leng Lim and Aleksandar D. Rakić
Sensors 2022, 22(22), 9001; https://doi.org/10.3390/s22229001 - 21 Nov 2022
Cited by 10 | Viewed by 2072
Abstract
In this article, we revisit the concept of optical feedback regimes in diode lasers and explore each regime experimentally from a somewhat unconventional point of view by relating the feedback regimes to the laser bias current and its optical feedback level. The results [...] Read more.
In this article, we revisit the concept of optical feedback regimes in diode lasers and explore each regime experimentally from a somewhat unconventional point of view by relating the feedback regimes to the laser bias current and its optical feedback level. The results enable setting the operating conditions of the diode laser in different applications requiring operation in different feedback regimes. We experimentally explored and theoretically supported this relationship from the standard Lang and Kobayashi rate equation model for a laser diode under optical feedback. All five regimes were explored for two major types of laser diodes: inplane lasers and vertical-cavity surface emitting lasers. For both lasers, we mapped the self-mixing strength vs. drive current and feedback level, observed the differences in the shape of the self-mixing fringes between the two laser architectures and a general simulation, and monitored other parameters of the lasers with changing optical feedback. Full article
(This article belongs to the Special Issue Laser Optical Feedback Turns 60: Results, Frontiers and Perspectives)
Show Figures

Figure 1

Figure 1
<p>Five classical regimes of optical feedback in a typical laser diode.</p>
Full article ">Figure 2
<p>Experimental setup used to evaluate SM in the tested lasers.</p>
Full article ">Figure 3
<p>Bias current vs. attenuation vs. amplitude (peak-to-peak of SM displacement signal). (dashed line = Ith, dotted line = bias current, where representative waveforms and spectra were measured). (<b>a</b>) 850 nm VCSEL; (<b>b</b>) 852 nm DFB.</p>
Full article ">Figure 4
<p>Self-mixing signal for different feedback levels (attenuation in dB) for harmonic displacement target. (<b>a</b>) VCSEL at a bias current of 7 mA and laser temperature of 35 <math display="inline"><semantics> <msup> <mrow/> <mo>∘</mo> </msup> </semantics></math>C; (<b>b</b>) DFB at a bias current of 60 mA and laser temperature of 25 <math display="inline"><semantics> <msup> <mrow/> <mo>∘</mo> </msup> </semantics></math>C; (<b>c</b>) simulated DFB.</p>
Full article ">Figure 5
<p>SM signal strength vs. attenuation. Red curves are the experimental result (circles—DFB, squares—VCSEL), and the blue curve is the simulation result.</p>
Full article ">Figure 6
<p>VCSEL (Dashed) and DFB (solid) SM signal strength vs. attenuation. Circles—displacement signal, squares—optically chopped signal.</p>
Full article ">Figure 7
<p>Laser spectra for different feedback levels (attenuation) for a static target. (<b>a</b>) VCSEL spectrum at a bias current of 7 mA and laser temperature of 35 <math display="inline"><semantics> <msup> <mrow/> <mo>∘</mo> </msup> </semantics></math>C. (<b>b</b>) DFB spectrum at a bias current of 70 mA and laser temperature of 25 <math display="inline"><semantics> <msup> <mrow/> <mo>∘</mo> </msup> </semantics></math>C.</p>
Full article ">
29 pages, 7047 KiB  
Article
Secure Vehicular Platoon Management against Sybil Attacks
by Danial Ritzuan Junaidi, Maode Ma and Rong Su
Sensors 2022, 22(22), 9000; https://doi.org/10.3390/s22229000 - 21 Nov 2022
Cited by 4 | Viewed by 2208
Abstract
The capacity of highways has been an ever-present constraint in the 21st century, bringing about the issue of safety with greater likelihoods of traffic accidents occurring. Furthermore, recent global oil prices have inflated to record levels. A potential solution lies in vehicular platooning, [...] Read more.
The capacity of highways has been an ever-present constraint in the 21st century, bringing about the issue of safety with greater likelihoods of traffic accidents occurring. Furthermore, recent global oil prices have inflated to record levels. A potential solution lies in vehicular platooning, which has been garnering attention, but its deployment is uncommon due to cyber security concerns. One particular concern is a Sybil attack, by which the admission of fake virtual vehicles into the platoon allows malicious actors to wreak havoc on the platoon itself. In this paper, we propose a secure management scheme for platoons that can protect major events that occur in the platoon operations against Sybil attacks. Both vehicle identity and message exchanged are authenticated by adopting key exchange, digital signature and encryption schemes based on elliptic curve cryptography (ECC). Noteworthy features of the scheme include providing perfect forward secrecy and both group forward and backward secrecy to preserve the privacy of vehicles and platoons. Typical malicious attacks such as replay and man-in-the-middle attacks for example can also be resisted. A formal evaluation of the security functionality of the scheme by the Canetti–Krawczyk (CK) adversary and the random oracle model as well as a brief computational verification by CryptoVerif were conducted. Finally, the performance of the proposed scheme was evaluated to show its time and space efficiency. Full article
(This article belongs to the Special Issue Feature Papers in Communications Section 2022)
Show Figures

Figure 1

Figure 1
<p>System architecture.</p>
Full article ">Figure 2
<p>Identity Authentication Phase of the SPMSA.</p>
Full article ">Figure 3
<p>Message Authentication Phase of the SPMSA Scheme.</p>
Full article ">Figure 4
<p>Platoon Key Update Phase of the SPMSA Scheme when a Vehicle Enters the Platoon.</p>
Full article ">Figure 5
<p>Platoon Communication Event of the SPMSA Scheme.</p>
Full article ">Figure 6
<p>Exit Request Phase of the SPMSA Scheme.</p>
Full article ">Figure 7
<p>Platoon Key Update Phase of the SPMSA Scheme when a Vehicle Leaves the Platoon.</p>
Full article ">Figure 8
<p>Security Verification by the Game-Hopping Process of CryptoVerif.</p>
Full article ">Figure 9
<p>Game-Hopping to Verify IND-CPA Property of Platoon Key Update Phases.</p>
Full article ">Figure 10
<p>First Game of Verifying the IND-CPA Property of Platoon Key Update Phases.</p>
Full article ">Figure 11
<p>Final Output of IND-CPA Verification of Platoon Key Update Phases.</p>
Full article ">Figure 12
<p>Game-Hopping to Verify INT_CTXT Property of Platoon Key Update Phases.</p>
Full article ">Figure 13
<p>First Game of Verifying the INT_CTXT Property of Platoon Key Update Phases.</p>
Full article ">Figure 14
<p>Output of INT_CTXT Verification of Platoon Key Update Phases.</p>
Full article ">Figure 15
<p>Performance of the Two Schemes against Unknown Attacks.</p>
Full article ">Figure 16
<p>Performance of the Two Schemes in the Presence of Sybil Nodes.</p>
Full article ">
21 pages, 2158 KiB  
Article
Deep Learning-Based Computer-Aided Diagnosis (CAD): Applications for Medical Image Datasets
by Yezi Ali Kadhim, Muhammad Umer Khan and Alok Mishra
Sensors 2022, 22(22), 8999; https://doi.org/10.3390/s22228999 - 21 Nov 2022
Cited by 17 | Viewed by 5649
Abstract
Computer-aided diagnosis (CAD) has proved to be an effective and accurate method for diagnostic prediction over the years. This article focuses on the development of an automated CAD system with the intent to perform diagnosis as accurately as possible. Deep learning methods have [...] Read more.
Computer-aided diagnosis (CAD) has proved to be an effective and accurate method for diagnostic prediction over the years. This article focuses on the development of an automated CAD system with the intent to perform diagnosis as accurately as possible. Deep learning methods have been able to produce impressive results on medical image datasets. This study employs deep learning methods in conjunction with meta-heuristic algorithms and supervised machine-learning algorithms to perform an accurate diagnosis. Pre-trained convolutional neural networks (CNNs) or auto-encoder are used for feature extraction, whereas feature selection is performed using an ant colony optimization (ACO) algorithm. Ant colony optimization helps to search for the best optimal features while reducing the amount of data. Lastly, diagnosis prediction (classification) is achieved using learnable classifiers. The novel framework for the extraction and selection of features is based on deep learning, auto-encoder, and ACO. The performance of the proposed approach is evaluated using two medical image datasets: chest X-ray (CXR) and magnetic resonance imaging (MRI) for the prediction of the existence of COVID-19 and brain tumors. Accuracy is used as the main measure to compare the performance of the proposed approach with existing state-of-the-art methods. The proposed system achieves an average accuracy of 99.61% and 99.18%, outperforming all other methods in diagnosing the presence of COVID-19 and brain tumors, respectively. Based on the achieved results, it can be claimed that physicians or radiologists can confidently utilize the proposed approach for diagnosing COVID-19 patients and patients with specific brain tumors. Full article
Show Figures

Figure 1

Figure 1
<p>Three classes of COVID-19 dataset: (<b>left</b>) COVID-19, (<b>middle</b>) Pneumonia, and (<b>right</b>) Normal.</p>
Full article ">Figure 2
<p>Three classes of brain tumor dataset: (<b>right</b>) Meningioma, (<b>middle</b>) Glioma, and (<b>left</b>) Pituitary Brain Tumors.</p>
Full article ">Figure 3
<p>Working method of the proposed feature selection model.</p>
Full article ">Figure 4
<p>Relationship between information coefficient and entropy.</p>
Full article ">Figure 5
<p>Image dataset pre-processing steps.</p>
Full article ">Figure 6
<p>Overall framework of the proposed classification system.</p>
Full article ">
11 pages, 1259 KiB  
Article
A Pilot Study of Heart Rate Variability Synchrony as a Marker of Intraoperative Surgical Teamwork and Its Correlation to the Length of Procedure
by Katarzyna Powezka, Allan Pettipher, Apit Hemakom, Tricia Adjei, Pasha Normahani, Danilo P. Mandic and Usman Jaffer
Sensors 2022, 22(22), 8998; https://doi.org/10.3390/s22228998 - 21 Nov 2022
Cited by 5 | Viewed by 1906
Abstract
Objective: Quality of intraoperative teamwork may have a direct impact on patient outcomes. Heart rate variability (HRV) synchrony may be useful for objective assessment of team cohesion and good teamwork. The primary aim of this study was to investigate the feasibility of using [...] Read more.
Objective: Quality of intraoperative teamwork may have a direct impact on patient outcomes. Heart rate variability (HRV) synchrony may be useful for objective assessment of team cohesion and good teamwork. The primary aim of this study was to investigate the feasibility of using HRV synchrony in surgical teams. Secondary aims were to investigate the association of HRV synchrony with length of procedure (LOP), complications, number of intraoperative glitches and length of stay (LOS). We also investigated the correlation between HRV synchrony and team familiarity, pre- and intraoperative stress levels (STAI questionnaire), NOTECHS score and experience of team members. Methods: Ear, nose and throat (ENT) and vascular surgeons (consultant and registrar team members) were recruited into the study. Baseline demographics including level of team members’ experience were gathered before each procedure. For each procedure, continuous electrocardiogram (ECG) recording was performed and questionnaires regarding pre- and intraoperative stress levels and non-technical skills (NOTECHS) scores were collected for each team member. An independent observer documented the time of each intraoperative glitch. Statistical analysis was conducted using stepwise multiple linear regression. Results: Four HRV synchrony metrics which may be markers of efficient surgical collaboration were identified from the data: 1. number of HRV synchronies per hour of procedure, 2. number of HRV synchrony trends per hour of procedure, 3. length of HRV synchrony trends per hour of procedure, 4. area under the HRV synchrony trend curve per hour of procedure. LOP was inversely correlated with number of HRV synchrony trends per hour of procedure (p < 0.0001), area under HRV synchrony trend curve per hour of procedure (p = 0.001), length of HRV synchrony trends per hour of procedure (p = 0.002) and number of HRV synchronies per hour of procedure (p < 0.0001). LOP was positively correlated with: FS (p = 0.043; R = 0.358) and intraoperative STAI score of the whole team (p = 0.007; R = 0.493). Conclusions: HRV synchrony metrics within operating teams may be used as an objective marker to quantify surgical teamwork. We have shown that LOP is shorter when the intraoperative surgical teams’ HRV is more synchronised. Full article
Show Figures

Figure 1

Figure 1
<p>Example of synchrony analysis in a vascular surgical pair. (<b>A</b>) shows detailed HRV synchrony (magenta) for the duration of the procedure, black arrows show HRV synchronies; (<b>B</b>) shows the trend of HRV synchrony for that procedure (magenta), the baseline synchrony is also shown (black, dotted), HRV synchrony trends are marked with black arrows, length of the HRV synchrony trend is marked with a black bracket and area under the peak of the HRV synchrony trend is marked in black.</p>
Full article ">Figure 2
<p>Heart rate variability (HRV) synchronies and synchrony trends within vascular pair in response to glitches from a representative case. Detailed HRV synchrony (<b>A</b>) and its trend (<b>B</b>) (magenta) are shown. Black dotted horizontal line shows baseline of HRV synchrony. Glitches are shown as vertical black solid lines with the number of a recorded glitch at the top (black arrows showing glitches). Red vertical dotted line shows the time of starting the procedure.</p>
Full article ">Figure 3
<p>Correlations between LOP and number of HRV synchrony trends per hour (<b>A</b>), area under peak of HRV synchrony trends per hour (<b>B</b>), length of HRV synchrony per hour (<b>C</b>) and number of HRV synchronies per hour (<b>D</b>).</p>
Full article ">
18 pages, 4243 KiB  
Article
Digital Mapping of Soil Organic Carbon Based on Machine Learning and Regression Kriging
by Changda Zhu, Yuchen Wei, Fubin Zhu, Wenhao Lu, Zihan Fang, Zhaofu Li and Jianjun Pan
Sensors 2022, 22(22), 8997; https://doi.org/10.3390/s22228997 - 21 Nov 2022
Cited by 17 | Viewed by 3254
Abstract
In the last two decades, machine learning (ML) methods have been widely used in digital soil mapping (DSM), but the regression kriging (RK) model which combines the advantages of the ML and kriging methods has rarely been used in DSM. In addition, due [...] Read more.
In the last two decades, machine learning (ML) methods have been widely used in digital soil mapping (DSM), but the regression kriging (RK) model which combines the advantages of the ML and kriging methods has rarely been used in DSM. In addition, due to the limitation of a single-model structure, many ML methods have poor prediction accuracy in undulating terrain areas. In this study, we collected the SOC content of 115 soil samples in a hilly farming area with continuous undulating terrain. According to the theory of soil-forming factors in pedogenesis, we selected 10 topographic indices, 7 vegetation indices, and 2 soil indices as environmental covariates, and according to the law of geographical similarity, we used ML and RK methods to mine the relationship between SOC and environmental covariates to predict the SOC content. Four ensemble models—random forest (RF), Cubist, stochastic gradient boosting (SGB), and Bayesian regularized neural networks (BRNNs)—were used to fit the trend of SOC content, and the simple kriging (SK) method was used to interpolate the residuals of the ensemble models, and then the SOC and residual were superimposed to obtain the RK prediction result. Moreover, the 115 samples were divided into calibration and validation sets at a ratio of 80%, and the tenfold cross-validation method was used to fit the optimal parameters of the model. From the results of four ensemble models: RF performed best in the calibration set (R2c = 0.834) but poorly in the validation set (R2v = 0.362); Cubist had good accuracy and stability in both the calibration and validation sets (R2c = 0.693 and R2v = 0.445); SGB performed poorly (R2c = 0.430 and R2v = 0.336); and BRNN had the lowest accuracy (R2c = 0.323 and R2v = 0.282). The results showed that the R2 of the four RK models in the validation set were 0.718, 0.674, 0.724, and 0.625, respectively. Compared with the ensemble models without superimposed residuals, the prediction accuracy was improved by 0.356, 0.229, 0.388, and 0.343, respectively. In conclusion, Cubist has high prediction accuracy and generalization ability in areas with complex topography, and the RK model can make full use of trends and spatial structural factors that are not easy to mine by ML models, which can effectively improve the prediction accuracy. This provides a reference for soil survey and digital mapping in complex terrain areas. Full article
(This article belongs to the Special Issue Multi-Sensor Fusion for Soil Monitoring)
Show Figures

Figure 1

Figure 1
<p>Distribution of study area and sampling points. Note: the sample point size represents the amount of SOC content (4.29~19.79 g/kg).</p>
Full article ">Figure 2
<p>Variable Pearson correlation coefficient matrix. Notes: The diagonal line indicates the variable names and distribution density curves, the upper diagonal panel is the correlation coefficient matrix, and the lower diagonal panel is the scatter plot matrix between the variables. SCI, soil color index; SRI, soil red index. RVI, ratio vegetation index; DVI, difference vegetation index; NDVI, normalized difference vegetation index; EVI, enhanced vegetation index; SAVI, soil-adjusted vegetation index; MSAVI, modified soil-adjusted vegetation index; SATVI, soil-adjusted total vegetation index. CNBL, channel network base level; CND, channel network distance; CI, convergence index; PlC, plan curvature; PrC, profile curvature; TWI, topographic wetness index; VD, valley depth; RSP, relative slope position; LSF, length–slope factor.</p>
Full article ">Figure 3
<p>Absolute residuals and reverse cumulative distribution curves of absolute residuals for the ranger, Cubist, GBM, and BRNN models.</p>
Full article ">Figure 4
<p>Residual–predicted value fit plot of the ML models: ranger, Cubist, GBM, and BRNN. Note: the shaded area indicates the confidence interval of the fitted line at the 95-percent confidence level.</p>
Full article ">Figure 5
<p>Semi-variance function and main parameters of the residuals of the ML models: (<b>a</b>) rangerresidual error, (<b>b</b>) Cubist <span class="underline">residual</span> error, (<b>c</b>) GBM <span class="underline">residual</span> error, (<b>d</b>) BRNN <span class="underline">residual</span> error.</p>
Full article ">Figure 6
<p>Predicted–observed fitted line of the ML models and RK models. Notes: The shaded area indicates the confidence interval of the fitted line at the 95-percent confidence level. The r<sup>2</sup> evaluates the deviation from the 1:1 line, whereas the R<sup>2</sup> evaluates the deviation from the fitted linear regression line between measured and predicted values.</p>
Full article ">Figure 7
<p>Permutation base on variable importance measures for the explanatory variables. Notes: the box plot indicates the error values of 10 permutations, and the height of bars indicates the mean error values of 10 permutations.</p>
Full article ">Figure 8
<p>Spatial distribution of SOC predicted by trend models and RK models: (<b>a</b>) ranger, (<b>b</b>) Cubist, (<b>c</b>) GBM, (<b>d</b>) BRNN, (<b>e</b>) RK-ranger, (<b>f</b>) RK-Cubist, (<b>g</b>) RK-GBM, (<b>h</b>) RK-BRNN.</p>
Full article ">Figure 8 Cont.
<p>Spatial distribution of SOC predicted by trend models and RK models: (<b>a</b>) ranger, (<b>b</b>) Cubist, (<b>c</b>) GBM, (<b>d</b>) BRNN, (<b>e</b>) RK-ranger, (<b>f</b>) RK-Cubist, (<b>g</b>) RK-GBM, (<b>h</b>) RK-BRNN.</p>
Full article ">
Previous Issue
Back to TopTop