[go: up one dir, main page]
Next Issue
Volume 18, August
Previous Issue
Volume 18, June
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
7.3
3.4
Journals
Sensors
Volume 18
Issue 7
sensors-logo

Journal Menu

► Journal Menu

Journal Browser

► Journal Browser
share announcement

Sensors, Volume 18, Issue 7 (July 2018) – 422 articles

Cover Story (view full-size image): Radon is a noble gas originated from the radioactive decay chain of uranium or thorium. It emanates naturally from the soil and from some building materials, in particular in regions with soils containing granite or slate. A correlation has been established between the presence of high radon gas concentrations and the incidence of lung cancer. For instance, since February 2018, all EU countries under Directive 2013/59/Euratom, shall ensure radon levels lower than the reference concentration of 300 Bq/m–3 in workplaces. Although commercial radon detectors are available, most of them either are expensive or provide limited monitoring capabilities. This article presents a cost-effective IoT radon gas remote monitoring system able to measure radon concentration, trigger alerts to prevent dangerous situations, warn users about them, and activate mitigation devices (e.g., forced ventilation). View the paper [...] Read more.
  • Issues are regarded as officially published after their release is announced to the table of contents alert mailing list.
  • You may sign up for e-mail alerts to receive table of contents of newly released issues.
  • PDF is the official format for papers published in both, html and pdf forms. To view the papers in pdf format, click on the "PDF Full-text" link, and use the free Adobe Reader to open them.
Order results
Result details
Section
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
9 pages, 2692 KiB  
Article
Characteristics of Surface Acoustic Wave Sensors with Nanoparticles Embedded in Polymer Sensitive Layers for VOC Detection
by Cristian Viespe and Dana Miu
Sensors 2018, 18(7), 2401; https://doi.org/10.3390/s18072401 - 23 Jul 2018
Cited by 27 | Viewed by 4637
Abstract
Surface Acoustic Wave (SAW) sensors with several types of polymer sensing films, containing embedded Fe3O4 nanoparticles (NPs) with various dimensions and concentrations, were studied. A sensor with a sensing film consisting of the polymer alone was used for comparison. NPs [...] Read more.
Surface Acoustic Wave (SAW) sensors with several types of polymer sensing films, containing embedded Fe3O4 nanoparticles (NPs) with various dimensions and concentrations, were studied. A sensor with a sensing film consisting of the polymer alone was used for comparison. NPs with a mean diameter of 7 nm were produced by laser ablation with 5 ns pulse durations, and NPs with 13 nm diameters were obtained with a laser having 10 ps pulse durations. The properties of the Surface Acoustic Wave sensors with such sensing films were analyzed. Their response (frequency shift, sensitivity, noise and response time) to three different volatile organic components (VOCs) at various concentrations were compared with one another. The frequency shift and sensitivity increased with increasing NP concentration in the polymer for a given NP dimension and with decreasing NP diameter for a given concentration. The best results were obtained for the smallest NPs used. The SAW sensor containing 7 nm NPs had a limit of detection (LOD) of 65 ppm (almost five times better than the sensor with polymer alone), and a response time of about 9 s for ethanol. Full article
(This article belongs to the Section Chemical Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Transmission Electronic Microscopy (TEM) images of the nanoparticles obtained using (<b>a</b>) a picosecond laser and (<b>b</b>) a nanosecond laser, which are used in the nanocomposite Surface Acoustic Wave (SAW) sensitive layer (S2 and S3 respectively).</p> Full article ">Figure 2
<p>Distribution of nanoparticles (NP) diameters as obtained using TEM images in the case of (<b>a</b>) ps laser ablation and (<b>b</b>) ns laser ablation. Line is fit with a lognormal function.</p> Full article ">Figure 3
<p>TEM image and SAED of a nanoparticle.</p> Full article ">Figure 4
<p>Experimental set-up used for measurements of sensor response to various Volatile Organic Compounds (VOCs).</p> Full article ">Figure 5
<p>Frequency shift of the sensors for different concentrations of ethanol, methanol and toluene VOCs. (<b>a</b>,<b>c</b>,<b>e</b>) compare sensors all having 50 nm NPs and different NP concentrations, and sensor S1 with no NPs. (<b>b</b>,<b>d</b>,<b>f</b>) compare sensors with the same NP concentration of 0.4 mg/mL and different NP diameters, and sensor S1 with no NPs. See <a href="#sensors-18-02401-t001" class="html-table">Table 1</a> for sensor identification.</p> Full article ">Figure 6
<p>Frequency shift of sensors at 1600 ppm concentration of ethanol, methanol and toluene.</p> Full article ">Figure 7
<p>Response time of the devices at a concentration of 1600 ppm of ethanol.</p> Full article ">
21 pages, 47301 KiB  
Article
Integrity and Collaboration in Dynamic Sensor Networks
by Steffen Schön, Claus Brenner, Hamza Alkhatib, Max Coenen, Hani Dbouk, Nicolas Garcia-Fernandez, Colin Fischer, Christian Heipke, Katja Lohmann, Ingo Neumann, Uyen Nguyen, Jens-André Paffenholz, Torben Peters, Franz Rottensteiner, Julia Schachtschneider, Monika Sester, Ligang Sun, Sören Vogel, Raphael Voges and Bernardo Wagner
Sensors 2018, 18(7), 2400; https://doi.org/10.3390/s18072400 - 23 Jul 2018
Cited by 31 | Viewed by 9056
Abstract
Global Navigation Satellite Systems (GNSS) deliver absolute position and velocity, as well as time information (P, V, T). However, in urban areas, the GNSS navigation performance is restricted due to signal obstructions and multipath. This is especially true for applications dealing with highly [...] Read more.
Global Navigation Satellite Systems (GNSS) deliver absolute position and velocity, as well as time information (P, V, T). However, in urban areas, the GNSS navigation performance is restricted due to signal obstructions and multipath. This is especially true for applications dealing with highly automatic or even autonomous driving. Subsequently, multi-sensor platforms including laser scanners and cameras, as well as map data are used to enhance the navigation performance, namely in accuracy, integrity, continuity and availability. Although well-established procedures for integrity monitoring exist for aircraft navigation, for sensors and fusion algorithms used in automotive navigation, these concepts are still lacking. The research training group i.c.sens, integrity and collaboration in dynamic sensor networks, aims to fill this gap and to contribute to relevant topics. This includes the definition of alternative integrity concepts for space and time based on set theory and interval mathematics, establishing new types of maps that report on the trustworthiness of the represented information, as well as taking advantage of collaboration by improved filters incorporating person and object tracking. In this paper, we describe our approach and summarize the preliminary results. Full article
(This article belongs to the Special Issue GNSS and Fusion with Other Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Collaborative positioning by vehicle-to-vehicle, vehicle-to-infrastructure and vehicle-to-pedestrian schemes. In addition, GNSS signals are received. The inter-nodal measurements are carried out using, e.g., laser scanners, stereo cameras and relative GNSS. Vehicle-to-infrastructure measurements are based on laser scanners and stereo cameras with dedicated processing and storage.</p> Full article ">Figure 2
<p>The participating vans with mounted sensor platforms during the mapathon: (<b>a</b>) GIH van, (<b>b</b>) IfE van and (<b>c</b>) IKG van.</p> Full article ">Figure 3
<p>Sketches of the different multi-sensor platforms (top view) for the GIH van (<b>a</b>), IfE van (<b>b</b>) and IKG van (<b>c</b>), with their respective sensors (black: GNSS antenna; cyan: tactical-grade IMU; green: laser scanner; blue: tactical-grade IMU; red: camera).</p> Full article ">Figure 4
<p>Multi-sensor calibration of one platform (red ellipse) by a laser tracker (blue ellipse) with control points for the stereo cameras (green circles) and planes for laser scanner (yellow surfaces), carried out in the 3D laboratory of the Geodetic Institute Hannover.</p> Full article ">Figure 5
<p>Sketch of three different collaborative scenarios (<b>a</b>–<b>c</b>), carried out during the Mapathon.</p> Full article ">Figure 6
<p>Photograph of a typical situation during the mapathon: in the first Meet &amp; Greet scenario, all three cars meet at a junction (from left to right: IfE, IKG, GIH van).</p> Full article ">Figure 7
<p>Reference trajectory for IfE van during the first Meet &amp; Greet scenario as obtained by a JAVAD Sigma sensor.</p> Full article ">Figure 8
<p>Schematic overview of the first Meet &amp; Greet scenario, combining the measured GNSS data with a Level Of Detail 2 (LOD 2) city model as done in <a href="#sec3dot3-sensors-18-02400" class="html-sec">Section 3.3</a>. i.c.sens, integrity and collaboration in dynamic sensor networks.</p> Full article ">Figure 9
<p>Example for the data preview on the i.c.sens website, showing a plot of the mobile mapping system point cloud from the same scenario, superimposed on a topographic base layer. Trajectories of the IKG van from all scenarios (green) are added as a separate layer. In a similar way, layers for all datasets can be combined to produce map mashups.</p> Full article ">Figure 10
<p>Comparison of different bounding methods: (<b>a</b>) 2D simulated solution sets of all methods, (<b>b</b>) Set Inversion Via Interval Analysis (SIVIA) and Linear Programming (LP) inconsistency zones from the scenario 3 test drive. LSA-IA, Least Squares Adjustment based on Interval Analysis.</p> Full article ">Figure 11
<p>Point positioning errors in horizontal and vertical directions for the trajectory of <a href="#sensors-18-02400-f010" class="html-fig">Figure 10</a>b. Top panel: DOPvalues, Middle panel: horizontal error in (m), Bottom panel: vertical position error in (m).</p> Full article ">Figure 12
<p>Qualitative evolution of the state estimation under the Zonotopic and Gaussian Kalman Filter (ZGKF) for a hybrid linear time-invariant system.</p> Full article ">Figure 13
<p>Exemplary illustration of a scenario for information-based georeferencing of a kinematic MSS (white vehicle) with classified building facades and road surface (green), as well as their geometric relation (black).</p> Full article ">Figure 14
<p>Results from the interval-based SLAM approach. (<b>a</b>) Image features color-coded by depth from close (red) via orange, yellow and green to distant (blue); (<b>b</b>) 2D position boxes (blue) that contain the vehicle’s true position (red).</p> Full article ">Figure 15
<p>Point cloud from seven measurement campaigns. Points are colored by campaign (blue = oldest, red = most recent).</p> Full article ">Figure 16
<p>Semantic segmentation of 3D point clouds, using labels transferred from image segmentation, without (<b>a</b>) and with (<b>b</b>) estimated label noise.</p> Full article ">Figure 17
<p>(<b>a</b>) Environmental model obtained from preprocessed point cloud. (<b>b</b>) Standard deviation results from the LKF estimation in the up direction (collaborative approach versus single vehicle).</p> Full article ">Figure 18
<p>Estimated vehicle models backprojected to the image. (<b>Left</b>): models assessed to moving cars; (<b>Right</b>): models assessed to parked cars.</p> Full article ">Figure 19
<p>Pedestrians are observed from two cameras in the mapathon.</p> Full article ">
22 pages, 8784 KiB  
Article
Voiceprint Identification for Limited Dataset Using the Deep Migration Hybrid Model Based on Transfer Learning
by Cunwei Sun, Yuxin Yang, Chang Wen, Kai Xie and Fangqing Wen
Sensors 2018, 18(7), 2399; https://doi.org/10.3390/s18072399 - 23 Jul 2018
Cited by 36 | Viewed by 7121
Abstract
The convolutional neural network (CNN) has made great strides in the area of voiceprint recognition; but it needs a huge number of data samples to train a deep neural network. In practice, it is too difficult to get a large number of training [...] Read more.
The convolutional neural network (CNN) has made great strides in the area of voiceprint recognition; but it needs a huge number of data samples to train a deep neural network. In practice, it is too difficult to get a large number of training samples, and it cannot achieve a better convergence state due to the limited dataset. In order to solve this question, a new method using a deep migration hybrid model is put forward, which makes it easier to realize voiceprint recognition for small samples. Firstly, it uses Transfer Learning to transfer the trained network from the big sample voiceprint dataset to our limited voiceprint dataset for the further training. Fully-connected layers of a pre-training model are replaced by restricted Boltzmann machine layers. Secondly, the approach of Data Augmentation is adopted to increase the number of voiceprint datasets. Finally, we introduce fast batch normalization algorithms to improve the speed of the network convergence and shorten the training time. Our new voiceprint recognition approach uses the TLCNN-RBM (convolutional neural network mixed restricted Boltzmann machine based on transfer learning) model, which is the deep migration hybrid model that is used to achieve an average accuracy of over 97%, which is higher than that when using either CNN or the TL-CNN network (convolutional neural network based on transfer learning). Thus, an effective method for a small sample of voiceprint recognition has been provided. Full article
(This article belongs to the Special Issue Artificial Intelligence and Machine Learning in Sensors Networks)
►▼ Show Figures

Figure 1

Figure 1
<p>Transfer of TLCNN-RBM model.</p> Full article ">Figure 2
<p>Speech signal—speech spectrogram conversion diagram. (<b>a</b>) Original speech signal and (<b>b</b>) speech spectrogram.</p> Full article ">Figure 3
<p>Structure of the CNN network based on the voiceprint identification.</p> Full article ">Figure 4
<p>Convex lens imaging schematic diagram.</p> Full article ">Figure 5
<p>Structure of the TLCNN-RBM hybrid model.</p> Full article ">Figure 6
<p>The speech contained in the NIST 2008 SRE dataset and the corresponding spectrogram.</p> Full article ">Figure 7
<p>The speech contained in the TIMIT dataset and the corresponding spectrogram.</p> Full article ">Figure 8
<p>Experimental flowchart.</p> Full article ">Figure 9
<p>Comparison of network pre-training time for the three models.</p> Full article ">Figure 10
<p>The number of epoch affects the loss value.</p> Full article ">Figure 11
<p>Comparison of the performance for five models. (<b>a</b>) Comparison of accuracy for five models based on female, (<b>b</b>) comparison of accuracy for five models based on male, (<b>c</b>) comparison of EER for five models based on female, and (<b>d</b>) comparison of EER for five models based on male.</p> Full article ">Figure 12
<p>Contrast the influence of RBM and FC on voiceprint identification. (<b>a</b>) Comparison of accuracy for FC and RBM based on female, (<b>b</b>) comparison of accuracy for FC and RBM based on male, (<b>c</b>) comparison of EER for FC and RBM based on female, and (<b>d</b>) comparison of ERR for FC and RBM based on male.</p> Full article ">Figure 13
<p>The accuracy for voiceprint recognition under different retraining sample sizes. (<b>a</b>) Comparison of accuracy based on female, (<b>b</b>) comparison of accuracy based on male.</p> Full article ">Figure 14
<p>Speech signal—speech spectrogram conversion diagram. (<b>a</b>) The speech collected in the laboratory and the corresponding spectrogram, and (<b>b</b>) the speech collected in the supermarket and the corresponding spectrogram.</p> Full article ">Figure 15
<p>Separate tests of the TLCNN-RBM model using the speech collected in two environments. (<b>a</b>) Speech collected in quiet environment (laboratory) was used for retraining and recognition results, and (<b>b</b>) speech collected in noisy environment (supermarket) was used for retraining and recognition results.</p> Full article ">Figure 16
<p>Software interface based on voiceprint identification.</p> Full article ">Figure 17
<p>The intelligent mailbox based on voiceprint identification.</p> Full article ">Figure 18
<p>A physical connection diagram of the mailbox.</p> Full article ">Figure 19
<p>A schematic diagram of the mailbox.</p> Full article ">
22 pages, 896 KiB  
Article
Impact of Node Speed on Energy-Constrained Opportunistic Internet-of-Things with Wireless Power Transfer
by Seung-Woo Ko and Seong-Lyun Kim
Sensors 2018, 18(7), 2398; https://doi.org/10.3390/s18072398 - 23 Jul 2018
Cited by 10 | Viewed by 4487
Abstract
Wireless power transfer (WPT) is a promising technology to realize the vision of Internet-of-Things (IoT) by powering energy-hungry IoT nodes by electromagnetic waves, overcoming the difficulty in battery recharging for massive numbers of nodes. Specifically, wireless charging stations (WCS) are deployed to transfer [...] Read more.
Wireless power transfer (WPT) is a promising technology to realize the vision of Internet-of-Things (IoT) by powering energy-hungry IoT nodes by electromagnetic waves, overcoming the difficulty in battery recharging for massive numbers of nodes. Specifically, wireless charging stations (WCS) are deployed to transfer energy wirelessly to IoT nodes in the charging coverage. However, the coverage is restricted due to the limited hardware capability and safety issue, making mobile nodes have different battery charging patterns depending on their moving speeds. For example, slow moving nodes outside the coverage resort to waiting for energy charging from WCSs for a long time while those inside the coverage consistently recharge their batteries. On the other hand, fast moving nodes are able to receive energy within a relatively short waiting time. This paper investigates the above impact of node speed on energy provision and the resultant throughput of energy-constrained opportunistic IoT networks when data exchange between nodes are constrained by their intermittent connections as well as the levels of remaining energy. To this end, we design a two-dimensional Markov chain of which the state dimensions represent remaining energy and distance to the nearest WCS normalized by node speed, respectively. Solving this enables providing the following three insights. First, faster node speed makes the inter-meeting time between a node and a WCS shorter, leading to more frequent energy supply and higher throughput. Second, the above effect of node speed becomes marginal as the battery capacity increases. Finally, as nodes are more densely deployed, the throughput becomes scaling with the density ratio between mobiles and WCSs but independent of node speed, meaning that the throughput improvement from node speed disappears in dense networks. The results provide useful guidelines for IoT network provisioning and planning to achieve the maximum throughput performance given mobile environments. Full article
(This article belongs to the Special Issue Green, Energy-Efficient and Sustainable Networks)
►▼ Show Figures

Figure 1

Figure 1
<p>The pattern of wireless charging when node speed is slow. During the period that a node is in the charging coverage of the WCS, it receives energy from WCS continuously. Once a node is out of the charging range, on the other hand, it takes a long time to receive energy from WCS again.</p> Full article ">Figure 2
<p>Two-dimensional Markov chain of which the horizontal and vertical state dimensions represent the number of remaining units of energy and the relative distance to the nearest WCS normalized by node speed, respectively.</p> Full article ">Figure 3
<p>(<b>a</b>) The CCDF of inter-meeting time under different node speed <span class="html-italic">v</span> (meter/slot); (<b>b</b>) expected throughput as a function of node speed <span class="html-italic">v</span> (meter/slot). Parameters: Network size <span class="html-italic">S</span> = 400 (in meter<sup>2</sup>), battery capacity <span class="html-italic">L</span> = 10 (units of energy), the maximum number of simultaneous transferable nodes <span class="html-italic">u</span> = 1, the maximum transferable energy per slot <span class="html-italic">E</span> = 3 (in units of energy), the number of nodes <span class="html-italic">n</span> = 10, and the number of WCSs <span class="html-italic">m</span> = 1.</p> Full article ">Figure 4
<p>Throughput vs. battery capacity <span class="html-italic">L</span>. The same parameter setting as in <a href="#sensors-18-02398-f003" class="html-fig">Figure 3</a> is used unless specified.</p> Full article ">Figure 5
<p>Throughput vs. node density <span class="html-italic">n</span>. The same parameter setting as in <a href="#sensors-18-02398-f003" class="html-fig">Figure 3</a> is used unless specified.</p> Full article ">Figure A1
<p>The CCDF of inter-meeting time of practical mobility models. The same parameter setting as in <a href="#sensors-18-02398-f003" class="html-fig">Figure 3</a> is used unless specified.</p> Full article ">
14 pages, 5115 KiB  
Article
Detection of Broken Strands of Transmission Line Conductors Using Fiber Bragg Grating Sensors
by Long Zhao, Xinbo Huang, Jianyuan Jia, Yongcan Zhu and Wen Cao
Sensors 2018, 18(7), 2397; https://doi.org/10.3390/s18072397 - 23 Jul 2018
Cited by 27 | Viewed by 6369
Abstract
Transmission lines are affected by Aeolian vibration, which causes strands to break and eventually causes an entire line to break. In this paper, a method for monitoring strand breaking based on modal identification is proposed. First, the natural frequency variation of a conductor [...] Read more.
Transmission lines are affected by Aeolian vibration, which causes strands to break and eventually causes an entire line to break. In this paper, a method for monitoring strand breaking based on modal identification is proposed. First, the natural frequency variation of a conductor caused by strand breakage is analyzed, and a modal experiment of the LGJ-95/15 conductor is conducted. The measurement results show that the natural frequencies of the conductor decrease with an increasing number of broken strands. Next, a monitoring system incorporating a fiber Bragg grating (FBG)-based accelerometer is designed in detail. The FBG sensor is mounted on the conductor to measure the vibration signal. A wind speed sensor is used to measure the wind speed signal and is installed on the tower. An analyzer is also installed on the tower to calculate the natural frequencies, and the data are sent to the monitoring center via 3G. Finally, a monitoring system is tested on a 110 kV experimental transmission line, and the short-time Fourier transform (STFT) method and stochastic subspace identification (SSI) method are used to identify the natural frequencies of the conductor vibration. The experimental results show that SSI analysis provides a higher precision than does STFT and can extract the natural frequency under various wind speeds as an effective basis for discriminating between broken strands. Full article
(This article belongs to the Special Issue Optical Waveguide Based Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Model of the transverse vibration of a transmission line.</p> Full article ">Figure 2
<p>Cross-sectional view of a transmission line: (<b>a</b>) Before strands are broken; (<b>b</b>) After strands are broken.</p> Full article ">Figure 3
<p>Experiment platform: (<b>a</b>) Schematic of the experimental platform; (<b>b</b>) Photograph of the experimental platform. ACSR = aluminum conductor steel reinforced.</p> Full article ">Figure 4
<p>Broken strand location.</p> Full article ">Figure 5
<p>Time-domain waveforms of accelerations: (<b>a</b>) acceleration of the vibrator before any strands were broken; (<b>b</b>) acceleration of the vibrator after one strand was broken; (<b>c</b>) acceleration of the ACSR before any strands were broken; (<b>d</b>) acceleration of the ACSR after one strand was broken.</p> Full article ">Figure 6
<p>Frequency response function waveform of the intact wire.</p> Full article ">Figure 7
<p>Frequency response function for a different number of broken strands.</p> Full article ">Figure 8
<p>Overall diagram of the monitoring system. FBG = fiber Bragg grating.</p> Full article ">Figure 9
<p>Block diagram of the analyzer.</p> Full article ">Figure 10
<p>Field test of the FBG-based monitoring system. (<b>a</b>) Tower; (<b>b</b>) FBG-based acceleration sensor.</p> Full article ">Figure 11
<p>Acceleration responses of the conductor. (<b>a</b>) Vertical acceleration of the conductor; (<b>b</b>) root mean square (RMS) of the acceleration.</p> Full article ">Figure 12
<p>Short-time Fourier transform (STFT) analysis of the conductor’s vibration responses at a wind velocity of 1.5 m/s. (<b>a</b>) 1st Mode; (<b>b</b>) 2nd Mode; (<b>c</b>) 3rd Mode; (<b>d</b>) 4th Mode.</p> Full article ">Figure 13
<p>STFT analysis of the conductor’s vibration responses at a wind velocity of 2.2 m/s. (<b>a</b>) 1st Mode; (<b>b</b>) 2nd Mode; (<b>c</b>) 3rd Mode; (<b>d</b>) 4th Mode.</p> Full article ">Figure 14
<p>STFT analysis of the conductor’s vibration responses at a wind velocity of 3.7 m/s. (<b>a</b>) 1st Mode; (<b>b</b>) 2nd Mode; (<b>c</b>) 3rd Mode; (<b>d</b>) 4th Mode.</p> Full article ">Figure 15
<p>STFT analysis of the conductor’s vibration responses at a wind velocity of 4.5 m/s. (<b>a</b>) 1st Mode; (<b>b</b>) 2nd Mode; (<b>c</b>) 3rd Mode; (<b>d</b>) 4th Mode.</p> Full article ">Figure 16
<p>Stability chart and power spectrum density function (PSD) response of stochastic subspace identification (SSI) for modal identification of the conductor at wind velocities of (<b>a</b>) 1.5 m/s; (<b>b</b>) 2.2 m/s; (<b>c</b>) 3.7 m/s; (<b>d</b>) 4.5 m/s.</p> Full article ">
9 pages, 2839 KiB  
Article
Intensity Demodulated Refractive Index Sensor Based on Front-Tapered Single-Mode-Multimode-Single-Mode Fiber Structure
by Jing Kang, Jiuru Yang, Xudong Zhang, Chunyu Liu and Lu Wang
Sensors 2018, 18(7), 2396; https://doi.org/10.3390/s18072396 - 23 Jul 2018
Cited by 22 | Viewed by 4104
Abstract
A novel intensity demodulated refractive index (RI) sensor is theoretically and experimentally demonstrated based on the front-tapered single-mode-multimode-single-mode (FT-SMS) fiber structure. The front taper is fabricated in a section of multimode fiber by flame-heated drawing technique. The intensity feature in the taper area [...] Read more.
A novel intensity demodulated refractive index (RI) sensor is theoretically and experimentally demonstrated based on the front-tapered single-mode-multimode-single-mode (FT-SMS) fiber structure. The front taper is fabricated in a section of multimode fiber by flame-heated drawing technique. The intensity feature in the taper area is analyzed through the beam propagation method and the comprehensive tests are then conducted in terms of RI and temperature. The experimental results show that, in FT-SMS, the relative sensitivity is −342.815 dB/RIU in the range of 1.33~1.37. The corresponding resolution reaches 2.92 × 10−5 RIU, which is more than four times higher than that in wavelength demodulation. The temperature sensitivity is 0.307 dB/°C and the measurement error from cross-sensitivity is less than 2 × 10−4. In addition, fabricated RI sensor presents high stability in terms of wavelength (±0.045 nm) and intensity (±0.386 dB) within 2 h of continuous operation. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Scheme diagram of the front-tapered single-mode-multimode-single-mode (FT-SMS) fiber structure.</p> Full article ">Figure 2
<p>Interference patterns in (<b>a</b>) front and (<b>b</b>) middle tapers, and the normalized intensity of tapered fiber structures with varied (<b>c</b>) transitional area length and (<b>d</b>) external refractive index (RI).</p> Full article ">Figure 3
<p>(<b>a</b>) Schematic diagram of fabrication of FT-SMS; (<b>b</b>) the CCD image of taper waist; (<b>c</b>) the transmission spectrum of FT-SMS.</p> Full article ">Figure 4
<p>Experimental setup for RI measurement.</p> Full article ">Figure 5
<p>(<b>a</b>) Shift of interference fringe with different concentration and (<b>b</b>) the sensitivity and linearity of wavelength and intensity with varied RI.</p> Full article ">Figure 6
<p>(<b>a</b>) Shift of interfered fringe with varied temperature and (<b>b</b>) the corresponding sensitivity and linearity of wavelength and intensity.</p> Full article ">Figure 7
<p>Stabilities in terms of wavelength and intensity within 2 h.</p> Full article ">
10 pages, 2164 KiB  
Article
A Flexible Capacitive Pressure Sensor Based on Ionic Liquid
by Xiaofeng Yang, Yishou Wang and Xinlin Qing
Sensors 2018, 18(7), 2395; https://doi.org/10.3390/s18072395 - 23 Jul 2018
Cited by 47 | Viewed by 9365
Abstract
A flexible microfluidic super-capacitive pressure sensor is developed to measure the surface pressure of a complex structure. The innovative sensor contains a filter paper filled with ionic liquid, and coated with two indium tin oxide polyethylene terephthalate (ITO-PET) films on the top and [...] Read more.
A flexible microfluidic super-capacitive pressure sensor is developed to measure the surface pressure of a complex structure. The innovative sensor contains a filter paper filled with ionic liquid, and coated with two indium tin oxide polyethylene terephthalate (ITO-PET) films on the top and bottom, respectively. When external pressure is applied on the top ITO-PET film of the sensor mounted on the surface of an aircraft, the capacitance between the two ITO-PET films will change because of the deformation of the top ITO-PET film. The external pressure will be determined based on the change of the capacitance. Compared to the traditional pressure sensor, the developed sensor provides a high sensitivity of up to 178.5 nF/KPa and rapid dynamic responses for pressure measurement. Meanwhile, experiments are also conducted to study the influence of the thickness of the sensing film, sensing area, temperature, and humidity. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Sensing principle of the flexible ionic liquid super-capacitive pressure sensor. (<b>a</b>) No external pressure applied on the sensor. (<b>b</b>) Low external pressure applied on the sensor. (<b>c</b>) High external pressure applied on the sensor. (<b>d</b>) Equivalent circuit model of the sensor.</p> Full article ">Figure 2
<p>Flexible ionic liquid super-capacitive pressure sensor. (<b>a</b>) Chemical structure of ionic liquid. (<b>b</b>) Schematic diagram of the flexible sensor. (<b>c</b>) Actual fabricated flexible ionic super-capacitive pressure sensor.</p> Full article ">Figure 3
<p>Characterization of the pressure sensing performance of the flexible ionic liquid super-capacitive pressure sensor. (<b>a</b>) The capacitance changes with driving frequency sweep from 20 Hz to 1.5 KHz at 0 KPa and 10 KPa, respectively. (<b>b</b>) Relative rate of change in capacitance of the sensor as a function of different pressure. The inset shows that the relative rate of change in capacitance of low pressure region A. (<b>c</b>) Relationship of the change in capacitance of the sensor attached on a circular tube and pressure applied. The inset shows the actual sensor attached on the circular tube. (<b>d</b>) The relationship of pressure and capacitance of the sensor cyclically compressed by 10 KPa. (<b>e</b>) The capacitance variation of the sensor under a square shaped pressure applied.</p> Full article ">Figure 4
<p>Influence of geometrical parameters on the sensor sensitivity. (<b>a</b>) Different thickness of the sensing membrane, 50 μm, 125 μm, and 175 μm. (<b>b</b>) The enlargement of the low pressure region A in <a href="#sensors-18-02395-f004" class="html-fig">Figure 4</a>a. (<b>c</b>) Different areas of the sensing chamber, 10 mm × 10 mm, 15 mm × 15 mm, and 20 mm × 20 mm.</p> Full article ">Figure 5
<p>Effect of the environment factors. (<b>a</b>) Capacitance change with frequency at temperature range from 25 °C to 55 °C. (<b>b</b>) Capacitance follows the frequency at humidity range from 40% to 80%.</p> Full article ">
19 pages, 444 KiB  
Article
A Multi-Server Two-Factor Authentication Scheme with Un-Traceability Using Elliptic Curve Cryptography
by Guosheng Xu, Shuming Qiu, Haseeb Ahmad, Guoai Xu, Yanhui Guo, Miao Zhang and Hong Xu
Sensors 2018, 18(7), 2394; https://doi.org/10.3390/s18072394 - 23 Jul 2018
Cited by 37 | Viewed by 5640
Abstract
To provide secure communication, the authentication-and-key-agreement scheme plays a vital role in multi-server environments, Internet of Things (IoT), wireless sensor networks (WSNs), etc. This scheme enables users and servers to negotiate for a common session initiation key. Our proposal first analyzes Amin et [...] Read more.
To provide secure communication, the authentication-and-key-agreement scheme plays a vital role in multi-server environments, Internet of Things (IoT), wireless sensor networks (WSNs), etc. This scheme enables users and servers to negotiate for a common session initiation key. Our proposal first analyzes Amin et al.’s authentication scheme based on RSA and proves that it cannot provide perfect forward secrecy and user un-traceability, and is susceptible to offline password guessing attack and key-compromise user impersonation attack. Secondly, we provide that Srinivas et al.’s multi-server authentication scheme is not secured against offline password guessing attack and key-compromise user impersonation attack, and is unable to ensure user un-traceability. To remedy such limitations and improve computational efficiency, we present a multi-server two-factor authentication scheme using elliptic curve cryptography (ECC). Subsequently, employing heuristic analysis and Burrows–Abadi–Needham logic (BAN-Logic) proof, it is proven that the presented scheme provides security against all known attacks, and in particular provides user un-traceability and perfect forward security. Finally, appropriate comparisons with prevalent works demonstrate the robustness and feasibility of the presented solution in multi-server environments. Full article
(This article belongs to the Section Internet of Things)
►▼ Show Figures

Figure 1

Figure 1
<p>The architecture of the multi-server authentication system.</p> Full article ">Figure 2
<p>Server registration.</p> Full article ">Figure 3
<p>User registration.</p> Full article ">Figure 4
<p>Login and authentication.</p> Full article ">
10 pages, 2059 KiB  
Article
Cooperative Dynamic Game-Based Optimal Power Control in Wireless Sensor Network Powered by RF Energy
by Manxi Wang, Haitao Xu and Xianwei Zhou
Sensors 2018, 18(7), 2393; https://doi.org/10.3390/s18072393 - 23 Jul 2018
Cited by 6 | Viewed by 3730
Abstract
This paper focuses on optimal power control in wireless sensor networks powered by RF energy, under the simultaneous wireless information and power transfer (SWIFT) protocol, where the information and power can be transmitted at the same time. We aim to maximize the utility [...] Read more.
This paper focuses on optimal power control in wireless sensor networks powered by RF energy, under the simultaneous wireless information and power transfer (SWIFT) protocol, where the information and power can be transmitted at the same time. We aim to maximize the utility for each sensor through the optimal power control, considering the influences of both the SINR and the harvested energy. The utility maximization problem is formulated as a cooperative dynamic game of a given time duration. All the sensors cooperate together to control their transmission power to maximize the utility and agree to act cooperatively so that a team optimum can be achieved. As a result, a feedback Nash equilibrium solution for each sensor is given based on the dynamic programming theory. Simulation results verify the effectiveness of the proposed approach, by comparing the grand coalition solutions with the non-cooperative solutions. Full article
(This article belongs to the Special Issue Low Energy Wireless Sensor Networks: Protocols, Architectures and Solutions)
►▼ Show Figures

Figure 1

Figure 1
<p>Wireless sensor network powered by RF energy.</p> Full article ">Figure 2
<p>Time switching (TS) protocol.</p> Full article ">Figure 3
<p>Optimal power level for information transmission of each sensor.</p> Full article ">Figure 4
<p>Network profit.</p> Full article ">Figure 5
<p>Profit of each sensor.</p> Full article ">Figure 6
<p>Battery energy variation.</p> Full article ">
19 pages, 9095 KiB  
Article
Geometric Parameter Calibration for a Cable-Driven Parallel Robot Based on a Single One-Dimensional Laser Distance Sensor Measurement and Experimental Modeling
by XueJun Jin, Jinwoo Jung, Seong Young Ko, Eunpyo Choi, Jong-Oh Park and Chang-Sei Kim
Sensors 2018, 18(7), 2392; https://doi.org/10.3390/s18072392 - 23 Jul 2018
Cited by 33 | Viewed by 7648
Abstract
A cable-driven parallel robot has benefits of wide workspace, high payload, and high dynamic response owing to its light cable actuator utilization. For wide workspace applications, in particular, the body frame becomes large to cover the wide workspace that causes robot kinematic errors [...] Read more.
A cable-driven parallel robot has benefits of wide workspace, high payload, and high dynamic response owing to its light cable actuator utilization. For wide workspace applications, in particular, the body frame becomes large to cover the wide workspace that causes robot kinematic errors resulting from geometric uncertainty. However, appropriate sensors as well as inexpensive and easy calibration methods to measure the actual robot kinematic parameters are not currently available. Hence, we present a calibration sensor device and an auto-calibration methodology for the over-constrained cable-driven parallel robots using one-dimension laser distance sensors attached to the robot end-effector, to overcome the robot geometric uncertainty and to implement precise robot control. A novel calibration workflow with five phases—preparation, modeling, measuring, identification, and adjustment—is proposed. The proposed calibration algorithms cover the cable-driven parallel robot kinematics, as well as uncertainty modeling such as cable elongation and pulley kinematics. We performed extensive simulations and experiments to verify the performance of the suggested method using the MINI cable robot. The experimental results show that the kinematic parameters can be identified correctly with 0.92 mm accuracy, and the robot position control accuracy is increased by 58%. Finally, we verified that the developed calibration sensor devices and the calibration methodology are applicable to the massive-size cable-driven parallel robot system. Full article
(This article belongs to the Section Intelligent Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Structure schematics of a cable-driven parallel robot with eight-cable configuration; (<b>a</b>) a schematic of the eight-cable-driven parallel robot, (<b>b</b>) kinematic drawing of the cable robot.</p> Full article ">Figure 2
<p>The MINI cable robot; (<b>a</b>) a photo of the whole MINI cable robot system and (<b>b</b>) a winch-motor drawing and actual installation.</p> Full article ">Figure 3
<p>MINI cable robot controller schematics.</p> Full article ">Figure 4
<p>A schematics of the motor-winch component and course of the cable with cable guide pulley.</p> Full article ">Figure 5
<p>Principles for cable force measurement.</p> Full article ">Figure 6
<p>Force sensor calibration results.</p> Full article ">Figure 7
<p>Complete hardware setup to start the referencing procedure.</p> Full article ">Figure 8
<p>Reference platform and coordinate system for the home position reference platform; (<b>a</b>) attachment of the reference planes at the platform, and (<b>b</b>) reference platform coordinate.</p> Full article ">Figure 9
<p>Polynomial surface model of the cable elongation as a function of cable length and cable tension measurement; (<b>a</b>) experimental data for different cable lengths and tensions, and (<b>b</b>) surface-fitted model where <math display="inline"><semantics> <mi>x</mi> </semantics></math> is the cable length, <math display="inline"><semantics> <mi>y</mi> </semantics></math> is the cable tension and <math display="inline"><semantics> <mi>z</mi> </semantics></math> is the cable deformation.</p> Full article ">Figure 10
<p>Extended pulley kinematics to encounter the pulley geometry into the cable robot kinematics; (<b>a</b>) determination of the point <math display="inline"><semantics> <mrow> <msub> <mi>C</mi> <mi>i</mi> </msub> </mrow> </semantics></math> where the cable exit, and (<b>b</b>) pulley rotation angle <math display="inline"><semantics> <mrow> <msub> <mi>γ</mi> <mi>i</mi> </msub> </mrow> </semantics></math> out of the <math display="inline"><semantics> <mrow> <mi>x</mi> <mi>z</mi> </mrow> </semantics></math> -plane.</p> Full article ">Figure 11
<p>Laser distance sensor setup.</p> Full article ">Figure 12
<p>Data acquisition phase.</p> Full article ">Figure 13
<p>Flowchart of parameter estimation using optimization method in the identification stage.</p> Full article ">Figure 14
<p>A photo-shoot of the actual measurements of the end-effector position using Faro edge ScanArm equipment.</p> Full article ">Figure 15
<p>Experimental results.</p> Full article ">Figure 16
<p>Path accuracy experimental results; (<b>a</b>) circular trajectory and (<b>b</b>) linear trajectory.</p> Full article ">
21 pages, 1281 KiB  
Article
Staged Incentive and Punishment Mechanism for Mobile Crowd Sensing
by Dan Tao, Shan Zhong and Hong Luo
Sensors 2018, 18(7), 2391; https://doi.org/10.3390/s18072391 - 23 Jul 2018
Cited by 7 | Viewed by 4351
Abstract
Having an incentive mechanism is crucial for the recruitment of mobile users to participate in a sensing task and to ensure that participants provide high-quality sensing data. In this paper, we investigate a staged incentive and punishment mechanism for mobile crowd sensing. We [...] Read more.
Having an incentive mechanism is crucial for the recruitment of mobile users to participate in a sensing task and to ensure that participants provide high-quality sensing data. In this paper, we investigate a staged incentive and punishment mechanism for mobile crowd sensing. We first divide the incentive process into two stages: the recruiting stage and the sensing stage. In the recruiting stage, we introduce the payment incentive coefficient and design a Stackelberg-based game method. The participants can be recruited via game interaction. In the sensing stage, we propose a sensing data utility algorithm in the interaction. After the sensing task, the winners can be filtered out using data utility, which is affected by time–space correlation. In particular, the participants’ reputation accumulation can be carried out based on data utility, and a punishment mechanism is presented to reduce the waste of payment costs caused by malicious participants. Finally, we conduct an extensive study of our solution based on realistic data. Extensive experiments show that compared to the existing positive auction incentive mechanism (PAIM) and reverse auction incentive mechanism (RAIM), our proposed staged incentive mechanism (SIM) can effectively extend the incentive behavior from the recruiting stage to the sensing stage. It not only achieves being a real-time incentive in both the recruiting and sensing stages but also improves the utility of sensing data. Full article
(This article belongs to the Special Issue Recent Advances in Artificial Intelligence and Deep Learning for Sensor Information Fusion)
►▼ Show Figures

Figure 1

Figure 1
<p>Group relationship.</p> Full article ">Figure 2
<p>The incentive process of mobile crowd sensing (MCS).</p> Full article ">Figure 3
<p>Maps.</p> Full article ">Figure 4
<p>Spatial and temporal distribution of the crowd.</p> Full article ">Figure 5
<p>Staged incentive mechanism framework.</p> Full article ">Figure 6
<p>Data utility decided by different sensing times and distances.</p> Full article ">Figure 7
<p>Flow of staged incentive and punishment mechanism.</p> Full article ">Figure 8
<p>Some simulation parameters used.</p> Full article ">Figure 9
<p>Comparison of data quality distribution.</p> Full article ">Figure 10
<p>Comparison of data delay distribution.</p> Full article ">Figure 11
<p>Comparison of data distance distribution.</p> Full article ">Figure 12
<p>Comparison of data orientation distribution.</p> Full article ">Figure 13
<p>Comparison of data utility distribution.</p> Full article ">Figure 14
<p>The number of winners for all sensing tasks.</p> Full article ">Figure 15
<p>Reputation accumulation of participants.</p> Full article ">
22 pages, 22998 KiB  
Article
Watermarking Based on Compressive Sensing for Digital Speech Detection and Recovery †
by Wenhuan Lu, Zonglei Chen, Ling Li, Xiaochun Cao, Jianguo Wei, Naixue Xiong, Jian Li and Jianwu Dang
Sensors 2018, 18(7), 2390; https://doi.org/10.3390/s18072390 - 23 Jul 2018
Cited by 19 | Viewed by 4846
Abstract
In this paper, a novel imperceptible, fragile and blind watermark scheme is proposed for speech tampering detection and self-recovery. The embedded watermark data for content recovery is calculated from the original discrete cosine transform (DCT) coefficients of host speech. The watermark information is [...] Read more.
In this paper, a novel imperceptible, fragile and blind watermark scheme is proposed for speech tampering detection and self-recovery. The embedded watermark data for content recovery is calculated from the original discrete cosine transform (DCT) coefficients of host speech. The watermark information is shared in a frames-group instead of stored in one frame. The scheme trades off between the data waste problem and the tampering coincidence problem. When a part of a watermarked speech signal is tampered with, one can accurately localize the tampered area, the watermark data in the area without any modification still can be extracted. Then, a compressive sensing technique is employed to retrieve the coefficients by exploiting the sparseness in the DCT domain. The smaller the tampered the area, the better quality of the recovered signal is. Experimental results show that the watermarked signal is imperceptible, and the recovered signal is intelligible for high tampering rates of up to 47.6%. A deep learning-based enhancement method is also proposed and implemented to increase the SNR of recovered speech signal. Full article
(This article belongs to the Special Issue Advances on Resources Management for Multi-Platform Infrastructures)
►▼ Show Figures

Figure 1

Figure 1
<p>Sketch of watermark embedding procedure.</p> Full article ">Figure 2
<p>Sketch of watermark recovery procedure.</p> Full article ">Figure 3
<p>The SNR values of all the recovered speech signal.</p> Full article ">Figure 4
<p>Original speech signal (<b>a</b>) Waveform (<b>b</b>) Spectrogram.</p> Full article ">Figure 4 Cont.
<p>Original speech signal (<b>a</b>) Waveform (<b>b</b>) Spectrogram.</p> Full article ">Figure 5
<p>Watermarked speech signal (<b>a</b>) Waveform (<b>b</b>) Spectrogram.</p> Full article ">Figure 6
<p>Damaged speech signal, the tampering percentage is 19.5% (<b>a</b>) Waveform (<b>b</b>) Spectrogram.</p> Full article ">Figure 6 Cont.
<p>Damaged speech signal, the tampering percentage is 19.5% (<b>a</b>) Waveform (<b>b</b>) Spectrogram.</p> Full article ">Figure 7
<p>Recovered speech signal, the tampering percentage is 19.5% (<b>a</b>) Waveform (<b>b</b>) Spectrogram.</p> Full article ">Figure 8
<p>Damaged speech signal, the tampering percentage is 47.6% (<b>a</b>) Waveform (<b>b</b>) Spectrogram.</p> Full article ">Figure 8 Cont.
<p>Damaged speech signal, the tampering percentage is 47.6% (<b>a</b>) Waveform (<b>b</b>) Spectrogram.</p> Full article ">Figure 9
<p>Recovered speech signal, the tampering percentage is 47.6% (<b>a</b>) Waveform (<b>b</b>) Spectrogram.</p> Full article ">Figure 10
<p>Damaged speech signal, different parts of the signal are tampered separately (<b>a</b>) Waveform (<b>b</b>) Spectrogram.</p> Full article ">Figure 10 Cont.
<p>Damaged speech signal, different parts of the signal are tampered separately (<b>a</b>) Waveform (<b>b</b>) Spectrogram.</p> Full article ">Figure 11
<p>Recovered speech signal, different parts of the signal are tampered separately (<b>a</b>) Waveform (<b>b</b>) Spectrogram.</p> Full article ">Figure 12
<p>Spectrograms. (<b>a</b>) Watermarked speech signal; (<b>b</b>) Recovered speech signal without processing; (<b>c</b>) Four hidden layers, decrement of nodes; (<b>d</b>) Four hidden layers, same number of nodes.</p> Full article ">Figure 12 Cont.
<p>Spectrograms. (<b>a</b>) Watermarked speech signal; (<b>b</b>) Recovered speech signal without processing; (<b>c</b>) Four hidden layers, decrement of nodes; (<b>d</b>) Four hidden layers, same number of nodes.</p> Full article ">Figure 13
<p>Waveforms. (<b>a</b>) Watermarked speech signal; (<b>b</b>) Recovered speech signal without processing; (<b>c</b>) Four hidden layers, decrement of nodes; (<b>d</b>) Four hidden layers, same number of nodes.</p> Full article ">Figure 13 Cont.
<p>Waveforms. (<b>a</b>) Watermarked speech signal; (<b>b</b>) Recovered speech signal without processing; (<b>c</b>) Four hidden layers, decrement of nodes; (<b>d</b>) Four hidden layers, same number of nodes.</p> Full article ">Figure 14
<p>Spectrograms of recovered speech signal using different numbers of hidden layers of DNN (<b>a</b>) 4 hidden layers, (<b>b</b>) 7 hidden layers.</p> Full article ">Figure 14 Cont.
<p>Spectrograms of recovered speech signal using different numbers of hidden layers of DNN (<b>a</b>) 4 hidden layers, (<b>b</b>) 7 hidden layers.</p> Full article ">Figure 15
<p>Waveforms of recovered speech signal by using different numbers of hidden layers of DNN (<b>a</b>) 4 hidden layers (<b>b</b>) 7 hidden layers.</p> Full article ">Figure 16
<p>Spectrograms of recovered speech signal by using different numbers of iterations of DNN (<b>a</b>) 100 iterations (<b>b</b>) 200 iterations(<b>c</b>) 500 iterations.</p> Full article ">Figure 17
<p>Waveforms of recovered speech signal by using different iterations of DNN (<b>a</b>) 100 iterations (<b>b</b>) 200 iterations(<b>c</b>) 500 iterations.</p> Full article ">
19 pages, 2239 KiB  
Article
ED-FNN: A New Deep Learning Algorithm to Detect Percentage of the Gait Cycle for Powered Prostheses
by Huong Thi Thu Vu, Felipe Gomez, Pierre Cherelle, Dirk Lefeber, Ann Nowé and Bram Vanderborght
Sensors 2018, 18(7), 2389; https://doi.org/10.3390/s18072389 - 23 Jul 2018
Cited by 67 | Viewed by 16257
Abstract
Throughout the last decade, a whole new generation of powered transtibial prostheses and exoskeletons has been developed. However, these technologies are limited by a gait phase detection which controls the wearable device as a function of the activities of the wearer. Consequently, gait [...] Read more.
Throughout the last decade, a whole new generation of powered transtibial prostheses and exoskeletons has been developed. However, these technologies are limited by a gait phase detection which controls the wearable device as a function of the activities of the wearer. Consequently, gait phase detection is considered to be of great importance, as achieving high detection accuracy will produce a more precise, stable, and safe rehabilitation device. In this paper, we propose a novel gait percent detection algorithm that can predict a full gait cycle discretised within a 1% interval. We called this algorithm an exponentially delayed fully connected neural network (ED-FNN). A dataset was obtained from seven healthy subjects that performed daily walking activities on the flat ground and a 15-degree slope. The signals were taken from only one inertial measurement unit (IMU) attached to the lower shank. The dataset was divided into training and validation datasets for every subject, and the mean square error (MSE) error between the model prediction and the real percentage of the gait was computed. An average MSE of 0.00522 was obtained for every subject in both training and validation sets, and an average MSE of 0.006 for the training set and 0.0116 for the validation set was obtained when combining all subjects’ signals together. Although our experiments were conducted in an offline setting, due to the forecasting capabilities of the ED-FNN, our system provides an opportunity to eliminate detection delays for real-time applications. Full article
(This article belongs to the Special Issue Assistance Robotics and Biosensors)
►▼ Show Figures

Figure 1

Figure 1
<p>A gait cycle is described as a dynamic and continuous occurrence of eight phases from the heel-contact at 0% to the next heel-contact at 100% percent of the gait cycle. Phase 0 is initial double-limb support, which appears during the first 10% of the cycle. Phase 1 is mid-stance, which appears from 10% to approximately 30% of the gait cycle. The following 10% of the gait cycle is terminal-stance. The propulsion phase or toe-off occurs after foot flat from 40% of the gait. This stage pushes the body forwards and prepares for swing phase from approximately 60% of the gait cycle. Single-limb support occurs from foot flat until 50% of the gait-related opposite initial contact limb, typically at 50% of the gait cycle. The second double-limb support occurs from the opposite limb at 50% until the toe leaves the ground at 60% of the gait cycle. Then, the second single-limb support completes the cycle. The following phases are early swing at approximately 60% to 75% of the gait cycle, mid swing at approximately 75% to 85% of the gait cycle, and late swing at approximately 85% to 100% of the gait cycle. Adapted from [<a href="#B36-sensors-18-02389" class="html-bibr">36</a>].</p> Full article ">Figure 2
<p>This figure illustrates the information flow in a recurrent neural network (RNN). The left image shows an RNN as an infinite loop network where the model outputs are fed back as inputs. The right figure is an unfolded representation of an RNN [<a href="#B53-sensors-18-02389" class="html-bibr">53</a>].</p> Full article ">Figure 3
<p>This figure illustrates how the matrix <span class="html-italic">D</span> was created. Every sample in the inertial measurement unit (IMU) is delayed by <span class="html-italic">n</span> times (in this case five times). The output matrix is shifted <span class="html-italic">n</span> times into the future (in this case three times).</p> Full article ">Figure 4
<p>This figure illustrates the exponentially delayed fully connected neural network (ED-FNN) architecture. Initially, the network individually receives each sensor input from the matrix <span class="html-italic">X</span> in Equation (<a href="#FD5-sensors-18-02389" class="html-disp-formula">5</a>). Then, the network separately extracts the features of each sensor and concatenates them into a single feature vector. Finally, the output layer uses the feature vector to forecast the gait events of the cycle.</p> Full article ">Figure 5
<p>Sensor positions of the IMU and the FSR on the foot. Arrow (A) illustrates the position of the IMU, and arrow (B) the position of FSRs under the sole.</p> Full article ">Figure 6
<p>The figure shows one gait cycle discretised with a 1% interval. The division was based on measuring cycle latency, from an initial-contact (IC) at 0% to the next at 100%.</p> Full article ">Figure 7
<p>This figure shows the prediction and the results of the learning process for one subject. (<b>a</b>) The ground truth and mean prediction of the gait phase discretisation divided into 100 portions normalised between 0 and 1 (0 equals to 0 percent and 1 equals 100 percent of the gait cycle). The bottom figure shows the <span class="html-italic">y</span> and <span class="html-italic">z</span> signals of the gyroscope sensor; (<b>b</b>) The mean and variance of the mean square error (MSE) learning curve. The average of MSE reached a loss of <math display="inline"><semantics> <mrow> <mn>0.003</mn> </mrow> </semantics></math> in the training set and <math display="inline"><semantics> <mrow> <mn>0.0662</mn> </mrow> </semantics></math> in the validation set.</p> Full article ">Figure 8
<p>This figure shows the prediction and learning process results of the joined signal for several subjects. Similar to <a href="#sensors-18-02389-f007" class="html-fig">Figure 7</a>, (<b>a</b>) shows the comparison between the prediction and the ground truth and (<b>b</b>) illustrates the learning curve of the MSE. The average MSE reached a loss of 0.006 in the training set and an average of 0.0115 in the test set.</p> Full article ">Figure 9
<p>This figure illustrates the MSE for every subject in the experiments. Each number in the <span class="html-italic">x</span>-axis from 1 to 7 represents a subject in the experiments. Violin plots 7, 8, and 9 were of two 15 degree incline walks and to all subjects’ signals combined, respectively. Every violin plot consists of two distributions (i.e., Train—blue and Test—orange ) and the mean of the MSE. The distributions illustrate the MSE variance over 100 runs. This figure shows that the ED-FNN managed to accurately predict the gait cycle over several subjects.</p> Full article ">
14 pages, 21753 KiB  
Article
The Influence of Radial Undersampling Schemes on Compressed Sensing in Cardiac DTI
by Jianping Huang, Wenlong Song, Lihui Wang and Yuemin Zhu
Sensors 2018, 18(7), 2388; https://doi.org/10.3390/s18072388 - 23 Jul 2018
Cited by 3 | Viewed by 4044
Abstract
Diffusion tensor imaging (DTI) is known to suffer from long acquisition time, which greatly limits its practical and clinical use. Undersampling of k-space data provides an effective way to reduce the amount of data to acquire while maintaining image quality. Radial undersampling is [...] Read more.
Diffusion tensor imaging (DTI) is known to suffer from long acquisition time, which greatly limits its practical and clinical use. Undersampling of k-space data provides an effective way to reduce the amount of data to acquire while maintaining image quality. Radial undersampling is one of the most popular non-Cartesian k-space sampling schemes, since it has relatively lower sensitivity to motion than Cartesian trajectories, and artifacts from linear reconstruction are more noise-like. Therefore, radial imaging is a promising strategy of undersampling to accelerate acquisitions. The purpose of this study is to investigate various radial sampling schemes as well as reconstructions using compressed sensing (CS). In particular, we propose two randomly perturbed radial undersampling schemes: golden-angle and random angle. The proposed methods are compared with existing radial undersampling methods, including uniformity-angle, randomly perturbed uniformity-angle, golden-angle, and random angle. The results on both simulated and real human cardiac diffusion weighted (DW) images show that, for the same amount of k-space data, randomly sampling around a random radial line results in better reconstruction quality for DTI indices, such as fractional anisotropy (FA), mean diffusivities (MD), and that the randomly perturbed golden-angle undersampling yields the best results for cardiac CS-DTI image reconstruction. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Illustration of randomly perturbed radial lines.</p> Full article ">Figure 2
<p>k-space sampling masks in one diffusion direction with sampling rate of 20%. (<b>a</b>) Uniform-angle; (<b>b</b>) golden-angle; (<b>c</b>) random-angle; (<b>d</b>–<b>f</b>) randomly perturbed sampling schemes corresponding to (<b>a</b>–<b>c</b>), respectively.</p> Full article ">Figure 3
<p>Reference human cardiac DW images. (<b>a</b>) Simulated data; (<b>b</b>) real data (the first acquisition dataset, the 67th slice); (<b>c</b>) real data (the second acquisition dataset, the fourth slice).</p> Full article ">Figure 4
<p>Top row: Fractional anisotropy (FA) maps of the simulated human heart data reconstructed from undersampled k-space data with a 20% sampling rate. Bottom row: FA error maps. (<b>a</b>) Reconstructed from the complete k-space data. Reconstructed from undersampled k-space using (<b>b</b>) uniform-angle radial; (<b>c</b>) golden-angle radial; (<b>d</b>) random-angle radial; and (<b>e</b>–<b>g</b>) the corresponding reconstructions with randomly perturbed radial sampling of (<b>b</b>–<b>d</b>).</p> Full article ">Figure 5
<p>Top row: FA maps of the first acquisition dataset reconstructed from undersampled k-space data with a 20% sampling rate. Bottom row: FA error maps. (<b>a</b>) Reconstruction from the complete k-space data. Reconstructions from undersampled k-space using (<b>b</b>) uniform-angle radial; (<b>c</b>) golden-angle radial; (<b>d</b>) random-angle radial; and (<b>e</b>–<b>g</b>) the corresponding randomly perturbed radial sampling of (<b>b</b>–<b>d</b>).</p> Full article ">Figure 6
<p>Top row: FA maps of the second acquisition dataset reconstructed from undersampled k-space data with a 20% sampling rate. Bottom row: FA error maps. (<b>a</b>) Reconstructed from the complete k-space data. Reconstructed from undersampled k-space using (<b>b</b>) uniform-angle radial; (<b>c</b>) golden-angle radial; (<b>d</b>) random-angle radial; and (<b>e</b>–<b>g</b>) the corresponding reconstructions with randomly perturbed radial sampling of (<b>b</b>–<b>d</b>).</p> Full article ">Figure 7
<p>Mean error bars of FA, Mean diffusivity (MD), Elevation angle (EA) and Azimuth angle (AA) for all the datasets. The sampling rate is 20%. (<b>a</b>) Mean error bars for the simulation dataset; (<b>b</b>) Mean error bars for the first acquisition dataset; (<b>c</b>) Mean error bars for the second acquisition dataset.</p> Full article ">Figure 7 Cont.
<p>Mean error bars of FA, Mean diffusivity (MD), Elevation angle (EA) and Azimuth angle (AA) for all the datasets. The sampling rate is 20%. (<b>a</b>) Mean error bars for the simulation dataset; (<b>b</b>) Mean error bars for the first acquisition dataset; (<b>c</b>) Mean error bars for the second acquisition dataset.</p> Full article ">Figure 8
<p>Performance comparisons on simulated data with different sampling rates. (<b>a</b>) RMSE of FA; (<b>b</b>) RMSE of MD; (<b>c</b>) RMSE of EA; (<b>d</b>) RMSE of AA.</p> Full article ">Figure 9
<p>Performance comparisons on the first acquisition dataset with different sampling rates. (<b>a</b>) mRMSE of FA; (<b>b</b>) mRMSE of MD; (<b>c</b>) mRMSE of EA; (<b>d</b>) mRMSE of AA.</p> Full article ">Figure 10
<p>Performance comparisons on the second acquisition dataset with different sampling rates. (<b>a</b>) mRMSE of FA; (<b>b</b>) mRMSE of MD; (<b>c</b>) mRMSE of EA; (<b>d</b>) mRMSE of AA.</p> Full article ">
15 pages, 2695 KiB  
Article
Determination of Optimal Heart Rate Variability Features Based on SVM-Recursive Feature Elimination for Cumulative Stress Monitoring Using ECG Sensor
by Dajeong Park, Miran Lee, Sunghee E. Park, Joon-Kyung Seong and Inchan Youn
Sensors 2018, 18(7), 2387; https://doi.org/10.3390/s18072387 - 23 Jul 2018
Cited by 26 | Viewed by 5549
Abstract
Routine stress monitoring in daily life can predict potentially serious health impacts. Effective stress monitoring in medical and healthcare fields is dependent upon accurate determination of stress-related features. In this study, we determined the optimal stress-related features for effective monitoring of cumulative stress. [...] Read more.
Routine stress monitoring in daily life can predict potentially serious health impacts. Effective stress monitoring in medical and healthcare fields is dependent upon accurate determination of stress-related features. In this study, we determined the optimal stress-related features for effective monitoring of cumulative stress. We first investigated the effects of short- and long-term stress on various heart rate variability (HRV) features using a rodent model. Subsequently, we determined an optimal HRV feature set using support vector machine-recursive feature elimination (SVM-RFE). Experimental results indicate that the HRV time domain features generally decrease under long-term stress, and the HRV frequency domain features have substantially significant differences under short-term stress. Further, an SVM classifier with a radial basis function kernel proved most accurate (93.11%) when using an optimal HRV feature set comprising the mean of R-R intervals (mRR), the standard deviation of R-R intervals (SDRR), and the coefficient of variance of R-R intervals (CVRR) as time domain features, and the normalized low frequency (nLF) and the normalized high frequency (nHF) as frequency domain features. Our findings indicate that the optimal HRV features identified in this study can effectively and efficiently detect stress. This knowledge facilitates development of in-facility and mobile healthcare system designs to support stress monitoring in daily life. Full article
(This article belongs to the Section Biosensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Experimental design used in this study. Based on the timeline of the experimental protocols, after 10 days of recovery from surgery, rats in the short-term stress (SS) or long-term stress (LS) groups were exposed to two or four weeks of stress, respectively, while rats in the control group rested. (<b>A</b>) At the baseline and the end of the two- and four-week stress periods, electrocardiograms (ECGs) were recorded for each group. (<b>B</b>) Based on the schedule of stress procedures, rats in the two stress groups (SS and LS) were exposed to unpredictable mild stress every day using seven randomly selected stressors.</p> Full article ">Figure 2
<p>Schematics of radio-telemetry system for the ECG recording. The ECGs, which were measured from the transmitter ECG sensor, were wirelessly transferred to the receiver through radio communication. The data stored in the receiver can be monitored and processed on the PC via the USB serial communication.</p> Full article ">Figure 3
<p>Rat model with surgically implanted transmitter to record electrocardiograms. (<b>A</b>) Implantable transmitter placement in a lead II configuration and (<b>B</b>) transmitter implant surgical process in which the transmitter body was positioned in the rat’s abdominal cavity with the positive and negative electrodes fixed to the left caudal rib and right pectoral muscle, respectively.</p> Full article ">Figure 4
<p>Schematic outline of the support vector machine-recursive feature elimination (SVM-RFE) process, which includes generating a ranking criterion for heart rate variability (HRV) features (<b>top</b>) and using it to determine the optimal feature set based on the results of six classifiers (<b>bottom</b>).</p> Full article ">Figure 5
<p>Effects of short- and long-term stress on (<b>A</b>–<b>F</b>) HRV time domain features and (<b>G</b>–<b>K</b>) frequency domain features. The LS group had significantly lower values across all HRV time domain features when compared with the control group and their pre-test. The SS group had significantly higher log-transformed low frequency power (<span class="html-italic">ln LF</span>), normalized low frequency (<span class="html-italic">nLF</span>), and log-transformed ratio of low frequency and high frequency powers (<span class="html-italic">ln (LF/HF)</span>) values and a significantly lower normalized high frequency (<span class="html-italic">nHF</span>) value when compared with the control group and their pre-test. (One-way analysis of variance followed by Tukey’s post-hoc test; * <span class="html-italic">p</span> &lt; 0.01 and ** <span class="html-italic">p</span> &lt; 0.001).</p> Full article ">Figure 6
<p>Colormap of weight value from the multiple SVM, Control versus all other groups (C-SVM), SS group versus all other groups (SS-SVM), and LS group versus all other groups (LS-SVM) for various HRV features. (<b>A</b>) The Multiple SVM included features determined by the C-SVM, SS-SVM, and LS-SVM. (<b>B</b>) Optimal features determined included the mean of R-R intervals (<span class="html-italic">mRR</span>), <span class="html-italic">ln (LF/HF</span>), and <span class="html-italic">ln HF</span> by the C-SVM; (<b>C</b>) the <span class="html-italic">nLF</span>, <span class="html-italic">nHF</span>, <span class="html-italic">ln LF</span>, and square root of the mean squared difference between adjacent R-R intervals (<span class="html-italic">RMSSD</span>) by the SS-SVM; and (<b>D</b>) the standard deviation of R-R intervals (<span class="html-italic">SDRR</span>), coefficient of variance of R-R intervals (<span class="html-italic">CVRR</span>), and mean of heart rates (<span class="html-italic">mHR</span>) by the LS-SVM. The results of the ranking process indicate the following descending order for the HRV features: <span class="html-italic">CVRR</span>, <span class="html-italic">nLF</span>, <span class="html-italic">nHF</span>, <span class="html-italic">SDRR</span>, <span class="html-italic">mRR</span>, <span class="html-italic">ln (LF/HF)</span>, <span class="html-italic">ln HF</span>, <span class="html-italic">mHR</span>, <span class="html-italic">ln LF,</span> proportion of N-N intervals that differ by more than 5 ms (<span class="html-italic">pNN5</span>), and <span class="html-italic">RMSSD</span>.</p> Full article ">Figure 7
<p>Comparison of optimal feature set on six different classifiers based on results of accuracies. Classifiers included (<b>A</b>) linear function SVM, (<b>B</b>) polynomial function SVM, (<b>C</b>) radial basis function (RBF) kernel function SVM, (<b>D</b>) <span class="html-italic">k</span>-nearest neighbor (<span class="html-italic">K</span>-NN), (<b>E</b>) linear discriminant analysis (LDA), and (<b>F</b>) the quadratic analysis (QDA) algorithm.</p> Full article ">Figure 8
<p>Comparison of six different classifiers based on overall accuracy for all HRV features and an optimal HRV feature set. The accuracy of each classifier increased when considering the optimal HRV feature set rather than all HRV features. The RBF kernel SVM classifier achieved the highest accuracy (93.11%) when considering the optimal HRV feature set.</p> Full article ">
19 pages, 1334 KiB  
Article
Supplemental Boosting and Cascaded ConvNet Based Transfer Learning Structure for Fast Traffic Sign Detection in Unknown Application Scenes
by Chunsheng Liu, Shuang Li, Faliang Chang and Wenhui Dong
Sensors 2018, 18(7), 2386; https://doi.org/10.3390/s18072386 - 22 Jul 2018
Cited by 10 | Viewed by 3821
Abstract
With rapid calculation speed and relatively high accuracy, the AdaBoost-based detection framework has been successfully applied in some real applications of machine vision-based intelligent systems. The main shortcoming of the AdaBoost-based detection framework is that the off-line trained detector cannot be transfer retrained [...] Read more.
With rapid calculation speed and relatively high accuracy, the AdaBoost-based detection framework has been successfully applied in some real applications of machine vision-based intelligent systems. The main shortcoming of the AdaBoost-based detection framework is that the off-line trained detector cannot be transfer retrained to adapt to unknown application scenes. In this paper, a new transfer learning structure based on two novel methods of supplemental boosting and cascaded ConvNet is proposed to address this shortcoming. The supplemental boosting method is proposed to supplementally retrain an AdaBoost-based detector for the purpose of transferring a detector to adapt to unknown application scenes. The cascaded ConvNet is designed and attached to the end of the AdaBoost-based detector for improving the detection rate and collecting supplemental training samples. With the added supplemental training samples provided by the cascaded ConvNet, the AdaBoost-based detector can be retrained with the supplemental boosting method. The detector combined with the retrained boosted detector and cascaded ConvNet detector can achieve high accuracy and a short detection time. As a representative object detection problem in intelligent transportation systems, the traffic sign detection problem is chosen to show our method. Through experiments with the public datasets from different countries, we show that the proposed framework can quickly detect objects in unknown application scenes. Full article
(This article belongs to the Section Intelligent Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>The structure of the proposed detection method.</p> Full article ">Figure 2
<p>The structure difference between the traditional cascade and our AdaBoost-based cascade for supplemental training.</p> Full article ">Figure 3
<p>The supplemental training process.</p> Full article ">Figure 4
<p>The process of training a supplemental trained stage.</p> Full article ">Figure 5
<p>The program pseudo code of the supplemental training method.</p> Full article ">Figure 6
<p>The structure of the cascaded ConvNet-based detector.</p> Full article ">Figure 7
<p>The signs from GTSRB and CVL.</p> Full article ">Figure 8
<p>The curves of the feature number and the test error rate on circular sign detection. The curve in (<b>a</b>) is the performance of the off-line-trained AdaBoost-based detector. The curve in (<b>b</b>) is the performance of the supplemental trained AdaBoost-based detector. The three red circle markers in the curves denote the test error rates of the basic cascaded stages, the stage of supplemental training and the supplemental trained detector, respectively.</p> Full article ">Figure 9
<p>The curves of the feature number and the test error rate on triangular sign detection. The curve in (<b>a</b>) is the performance of the off-line-trained AdaBoost-based detector. The curve in (<b>b</b>) is the performance of the supplemental trained AdaBoost-based detector.</p> Full article ">Figure 10
<p>The ROC curves of AdaBoost, faster faster regions with CNN (CNN) and the proposed method: (<b>a</b>) shows the ROC curves of these three methods tested on circular signs; (<b>b</b>) shows the ROC curves of these three methods tested on triangular signs.</p> Full article ">Figure 11
<p>Part of our detection results. The images from (<b>a</b>–<b>d</b>) are the detected results from CVL dataset.</p> Full article ">
25 pages, 6746 KiB  
Article
Passive RFID-Based Inventory of Traffic Signs on Roads and Urban Environments
by José Ramón García Oya, Rubén Martín Clemente, Eduardo Hidalgo Fort, Ramón González Carvajal and Fernando Muñoz Chavero
Sensors 2018, 18(7), 2385; https://doi.org/10.3390/s18072385 - 22 Jul 2018
Cited by 28 | Viewed by 7260
Abstract
This paper presents a system with location functionalities for the inventory of traffic signs based on passive RFID technology. The proposed system simplifies the current video-based techniques, whose requirements regarding visibility are difficult to meet in some scenarios, such as dense urban areas. [...] Read more.
This paper presents a system with location functionalities for the inventory of traffic signs based on passive RFID technology. The proposed system simplifies the current video-based techniques, whose requirements regarding visibility are difficult to meet in some scenarios, such as dense urban areas. In addition, the system can be easily extended to consider any other street facilities, such as dumpsters or traffic lights. Furthermore, the system can perform the inventory process at night and at a vehicle’s usual speed, thus avoiding interfering with the normal traffic flow of the road. Moreover, the proposed system exploits the benefits of the passive RFID technologies over active RFID, which are typically employed on inventory and vehicular routing applications. Since the performance of passive RFID is not obvious for the required distance ranges on these in-motion scenarios, this paper, as its main contribution, addresses the problem in two different ways, on the one hand theoretically, presenting a radio wave propagation model at theoretical and simulation level for these scenarios; and on the other hand experimentally, comparing passive and active RFID alternatives regarding costs, power consumption, distance ranges, collision problems, and ease of reconfiguration. Finally, the performance of the proposed on-board system is experimentally validated, testing its capabilities for inventory purposes. Full article
(This article belongs to the Section Sensor Networks)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>System overview.</p> Full article ">Figure 2
<p>Spatial location of the detected signs.</p> Full article ">Figure 3
<p>Map generated from the detected signs information.</p> Full article ">Figure 4
<p>Propagation paths.</p> Full article ">Figure 5
<p>Test configuration.</p> Full article ">Figure 6
<p>Maximum distance range for a non-faced passive tag.</p> Full article ">Figure 7
<p>Simulation results: <span class="html-italic">v</span> = 50 Km/h, <span class="html-italic">d</span> = 2 m, <span class="html-italic">f</span> = 866.5 MHz.</p> Full article ">Figure 8
<p>Simulation results: <span class="html-italic">v</span> = 120 Km/h, <span class="html-italic">d</span> = 2 m, <span class="html-italic">f</span> = 866.5 MHz.</p> Full article ">Figure 9
<p>Simulation results: <span class="html-italic">v</span> = 50 Km/h, <span class="html-italic">d</span> = 4 m, <span class="html-italic">f</span> = 866.5 MHz.</p> Full article ">Figure 10
<p>Simulation results: <span class="html-italic">v</span> = 120 Km/h, <span class="html-italic">d</span> = 4 m, <span class="html-italic">f</span> = 866.5 MHz.</p> Full article ">Figure 11
<p>Simulation results: <span class="html-italic">v</span> = 50 Km/h, <span class="html-italic">d</span> = 2 m, <span class="html-italic">f</span> = 866.5 MHz in a severe multi-path environment.</p> Full article ">Figure 12
<p>Simulation results: <span class="html-italic">v</span> = 50 Km/h, <span class="html-italic">d</span> = 2 m, <span class="html-italic">f</span> = 2450 MHz.</p> Full article ">Figure 13
<p>Tags over the traffic sign (<b>left</b>) and configuration setup (<b>right</b>).</p> Full article ">Figure 14
<p>Number of detections for each studied case.</p> Full article ">
10 pages, 2277 KiB  
Article
Self-Power Dynamic Sensor Based on Triboelectrification for Tilt of Direction and Angle
by Hyeonhee Roh, Inkyum Kim, Jinsoo Yu and Daewon Kim
Sensors 2018, 18(7), 2384; https://doi.org/10.3390/s18072384 - 22 Jul 2018
Cited by 8 | Viewed by 3479
Abstract
With the great development of the Internet of Things (IoT), the use of sensors have increased rapidly because of the importance in the connection between machines and people. A huge number of IoT sensors consume vast amounts of electrical power for stable operation [...] Read more.
With the great development of the Internet of Things (IoT), the use of sensors have increased rapidly because of the importance in the connection between machines and people. A huge number of IoT sensors consume vast amounts of electrical power for stable operation and they are also used for a wide range of applications. Therefore, sensors need to operate independently, sustainably, and wirelessly to improve their capabilities. In this paper, we propose an orientation and the tilt triboelectric sensor (OT-TES) as a self-powered active sensor, which can simultaneously sense the tilting direction and angle by using the two classical principles of triboelectrification and electrostatic induction. The OT-TES device consists of a rectangular acrylic box containing polytetrafluoroethylene (PTFE) balls moved by gravity. The output voltage and current were 2 V and 20 nA, respectively, with a PTFE ball and Al electrode. The multi-channel system was adopted for measuring the degree and direction of tilt by integrating the results of measured electrical signals from the eight electrodes. This OT-TES can be attached on the equipment for drones or divers to measure their stability. As a result, this proposed device is expected to expand the field of TES, as a sensor for sky and the underwater. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>(<b>a</b>) Overall schematics of the orientation and tilt triboelectric sensor (OT-TES); (<b>b</b>) Cross-sectional view of the OT-TES; (<b>c</b>) Working mechanism of the OT-TES.</p> Full article ">Figure 2
<p>Electrical signals of a polytetrafluoroethylene (PTFE) ball from one Al electrode: (<b>a</b>) Open-circuit voltage; (<b>b</b>) Short-circuit current.</p> Full article ">Figure 3
<p>(<b>a</b>) Simple circuit diagram of a multi-channel system, including the eight electrodes located on the OT-TES. The data acquisition board (DAQ) conducts the data acquisition, and the personal computer (PC) analyzes the data; (<b>b</b>) Cross-sectional view of the OT-TES during tilting to the left and right side; (<b>c</b>–<b>e</b>) The DAQ output voltage measurements of a PTFE ball moving between two Al electrodes: (<b>c</b>) Tilt OT-TES at 10°; (<b>d</b>) tilt OT-TES at 20°; and (<b>e</b>) tilt OT-TES at 30°.</p> Full article ">Figure 4
<p>(<b>a</b>–<b>c</b>) The DAQ output voltage according to the number of PTFE balls occupying the bottom surface. The tilting angle and frequency are 20° and 0.5 Hz, respectively: (<b>a</b>) The amount of balls filled 12% of the bottom surface; (<b>b</b>) The amount of balls filled 25% of the bottom surface; (<b>c</b>) The amount of balls filled 50% of the bottom surface; (<b>d</b>) The open-circuit voltage of the bottom side with 25% amount of balls according to the tilting angle; (<b>e</b>) The open-circuit voltage of the bottom side and the top side.</p> Full article ">Figure 5
<p>(<b>a</b>) The schematic of the OT-TES and each electrode number. (<b>b</b>–<b>e</b>) Voltage measurements of the OT-TES from eight Al electrodes when the tilt is 45 degrees to the left and right side and around to the center of the OT-TES: (<b>b</b>) The DAQ output voltage of electrode 1 and 2, and (<b>c</b>) the DAQ output voltage of electrode 5 and 6. (<b>d</b>) The DAQ output voltage of electrode 3 and 4, and (<b>e</b>) the DAQ output voltage of electrode 7 and 8.</p> Full article ">
9 pages, 19501 KiB  
Article
Polymer Based Whispering Gallery Mode Humidity Sensor
by Ann Britt Petermann, Thomas Hildebrandt, Uwe Morgner, Bernhard Wilhelm Roth and Merve Meinhardt-Wollweber
Sensors 2018, 18(7), 2383; https://doi.org/10.3390/s18072383 - 22 Jul 2018
Cited by 22 | Viewed by 4561
Abstract
Whispering gallery mode (WGM) resonators are versatile high sensitivity sensors, but applications regularly suffer from elaborate and expensive manufacturing and read-out. We have realized a simple and inexpensive concept for an all-polymer WGM sensor. Here, we evaluate its performance for relative humidity measurements [...] Read more.
Whispering gallery mode (WGM) resonators are versatile high sensitivity sensors, but applications regularly suffer from elaborate and expensive manufacturing and read-out. We have realized a simple and inexpensive concept for an all-polymer WGM sensor. Here, we evaluate its performance for relative humidity measurements demonstrating a sensitivity of 47 pm/% RH. Our results show the sensor concepts’ promising potential for use in real-life applications and environments. Full article
(This article belongs to the Special Issue State-of-the-Art Sensors Technology in Germany)
►▼ Show Figures

Figure 1

Figure 1
<p>Experimental setup: Ten percent of the laser intensity is detected with a photodiode for calibration purposes. The remaining ninety percent of the laser intensity is collimated and coupled under 45° in a PMMA plate and guided based on total internal reflection. PMMA spheres supporting the whispering gallery modes (WGMs) are placed in the evanescent field present at the plate surface. The light distribution is captured by a CMOS camera equipped with a microscope objective.</p> Full article ">Figure 2
<p>The WGM sensor, the commercially available relative humidity sensor and the saturated salt solution for humidity adjustment, placed under an acrylic box.</p> Full article ">Figure 3
<p>Mode map for an array of twenty spheres with a mean diameter of <math display="inline"><semantics> <mrow> <mn>74.44</mn> </mrow> </semantics></math> <math display="inline"><semantics> <mi mathvariant="sans-serif">μ</mi> </semantics></math><math display="inline"><semantics> <mi mathvariant="normal">m</mi> </semantics></math> (the frame number relates to wavelength).</p> Full article ">Figure 4
<p>Comparison of scanning wavelengths and determined wavelengths, if the relative humidity changes (circles) or remains constant (crosses).</p> Full article ">Figure 5
<p>Determined wavelength compared to the laser scanning wavelength for different relative humidity levels. The initial relative humidity was <math display="inline"><semantics> <mrow> <mn>40.1</mn> </mrow> </semantics></math>%, and the array consisted of twenty 74 <math display="inline"><semantics> <mi mathvariant="sans-serif">μ</mi> </semantics></math><math display="inline"><semantics> <mi mathvariant="normal">m</mi> </semantics></math> spheres.</p> Full article ">Figure 6
<p>Wavelength shift <math display="inline"><semantics> <mrow> <mo>Δ</mo> <mi>λ</mi> </mrow> </semantics></math> as a function of relative humidity for arrays consisting of 74 <math display="inline"><semantics> <mi mathvariant="sans-serif">μ</mi> </semantics></math><math display="inline"><semantics> <mi mathvariant="normal">m</mi> </semantics></math> spheres and 165 <math display="inline"><semantics> <mi mathvariant="sans-serif">μ</mi> </semantics></math><math display="inline"><semantics> <mi mathvariant="normal">m</mi> </semantics></math> spheres, respectively. The slope of the curves is equal to the sensitivity.</p> Full article ">Figure 7
<p>Determined wavelength compared to the laser scanning wavelength for a fixed sensor array at different relative humidity levels. The initial relative humidity was <math display="inline"><semantics> <mrow> <mn>44.46</mn> </mrow> </semantics></math>%, and the array consisted of twenty fixed spheres with diameter of 74 <math display="inline"><semantics> <mi mathvariant="sans-serif">μ</mi> </semantics></math><math display="inline"><semantics> <mi mathvariant="normal">m</mi> </semantics></math>.</p> Full article ">
20 pages, 7169 KiB  
Article
Design and Development of a 5-Channel Arduino-Based Data Acquisition System (ABDAS) for Experimental Aerodynamics Research
by Antonio Vidal-Pardo and Santiago Pindado
Sensors 2018, 18(7), 2382; https://doi.org/10.3390/s18072382 - 22 Jul 2018
Cited by 20 | Viewed by 7535
Abstract
In this work, a new and low-cost Arduino-Based Data Acquisition System (ABDAS) for use in an aerodynamics lab is developed. Its design is simple and reliable. The accuracy of the system has been checked by being directly compared with a commercial and high [...] Read more.
In this work, a new and low-cost Arduino-Based Data Acquisition System (ABDAS) for use in an aerodynamics lab is developed. Its design is simple and reliable. The accuracy of the system has been checked by being directly compared with a commercial and high accuracy level hardware from National Instruments. Furthermore, ABDAS has been compared to the accredited calibration system in the IDR/UPM Institute, its measurements during this testing campaign being used to analyzed two different cup anemometer frequency determination procedures: counting pulses and the Fourier transform. The results indicate a more accurate transfer function of the cup anemometers when counting pulses procedure is used. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Diagram of the Arduino-Based Data Acquisition System (ABDAS) described in the present work.</p> Full article ">Figure 2
<p>Sketch of the voltage dividers used to enlarge the measuring rang of ABDAS.</p> Full article ">Figure 3
<p>Transfer function (Equation (2)) of ABDAS voltage dividers. The Equation (linear fitting) correspondent to Channel-1 voltage divider (VD-1) is included in the Figure. See also <a href="#sensors-18-02382-t002" class="html-table">Table 2</a>.</p> Full article ">Figure 4
<p>Pictures of ABDAS along its development/manufacturing process. (<b>a</b>) Connectors, lightning and switching panel. (<b>b</b>) ABDAS distribution inside the enclosure. (<b>c</b>) Front view. (<b>d</b>) Rear view.</p> Full article ">Figure 5
<p>ABDAS software flowchart. Program steps and measurement loop.</p> Full article ">Figure 6
<p>Image capture with part of the ABDAS software control program code.</p> Full article ">Figure 7
<p>Serial monitor window of ABDAS.</p> Full article ">Figure 8
<p>Results from a measurement carried out with ABDAS. Text file example. Two channels measured (two first column). The third column indicates the time of each measurement point (expressed in seconds).</p> Full article ">Figure 9
<p>Arduino-based acquisition system (ABDAS) and the National Instruments NI USB-6210 data acquisition system (NIDAS), used to check its performances.</p> Full article ">Figure 10
<p>One wave period of the 10 Hz frequency reference signal (Equation (3)). The data measured with NIDAS and ABDAS have been included for comparison purposes.</p> Full article ">Figure 11
<p>Sketch of the S4 wind tunnel at the IDR/UPM Institute used for anemometer calibration. The different parts of the wind tunnel are indicated in the figure as follows: 1. Fans; 2. Plenum chamber; 3. Honeycomb and grids; 4. Contraction; 5. Test chamber; 6. Diffuser. [<a href="#B11-sensors-18-02382" class="html-bibr">11</a>].</p> Full article ">Figure 12
<p>Offset, <span class="html-italic">y</span><sub>0</sub>, and the first harmonic term, <span class="html-italic">y</span><sub>1</sub>, extracted from the sampled data of same signal (Equation (3)) measured with NIDAS and ABDAS.</p> Full article ">Figure 13
<p>Harmonic terms, <span class="html-italic">y</span><sub>2</sub> to <span class="html-italic">y</span><sub>10</sub>, extracted from the sampled data of same signal (Equation (3)) measured with NIDAS and ABDAS.</p> Full article ">Figure 14
<p>Estimated accuracy ABDAS and NIDAS in relation to the frequency of the signal measured (Equation (3)).</p> Full article ">Figure 15
<p>Number sample points (open circles) measured with ABDAS into three 2-period brackets of the three tested cup anemometers’ output signal.</p> Full article ">Figure 16
<p>Fourier transforms of the recording data with ABDAS during the calibration process of Anemometer-2, at 4 and 16 m·s<sup>−1</sup>.</p> Full article ">Figure 17
<p>Calibration points of Anemometer-1, obtained from the IDR/UPM calibration system (the transfer function resulting from these data is included in the graph), and from the data measured with ABDAS and post-processed counting pulses (ABDAS-CP; see Equation (6)) and using the Fourier transform (ABDAS-FR; see Equation (7)).</p> Full article ">Figure 18
<p>Percentage difference of the frequencies obtained, ∆<span class="html-italic">f</span> (Equation (8)), from the ABDAS data measurements using the pulse-counting (CP; Equation (6)) and the Fourier transform (FR; Equation (7)) procedures, related to the ones measured with the IDR/UPM system, <span class="html-italic">f<sub>IDR</sub></span>.</p> Full article ">
13 pages, 3080 KiB  
Article
Estimation of Cough Peak Flow Using Cough Sounds
by Yasutaka Umayahara, Zu Soh, Kiyokazu Sekikawa, Toshihiro Kawae, Akira Otsuka and Toshio Tsuji
Sensors 2018, 18(7), 2381; https://doi.org/10.3390/s18072381 - 22 Jul 2018
Cited by 18 | Viewed by 6783
Abstract
Cough peak flow (CPF) is a measurement for evaluating the risk of cough dysfunction and can be measured using various devices, such as spirometers. However, complex device setup and the face mask required to be firmly attached to the mouth impose burdens on [...] Read more.
Cough peak flow (CPF) is a measurement for evaluating the risk of cough dysfunction and can be measured using various devices, such as spirometers. However, complex device setup and the face mask required to be firmly attached to the mouth impose burdens on both patients and their caregivers. Therefore, this study develops a novel cough strength evaluation method using cough sounds. This paper presents an exponential model to estimate CPF from the cough peak sound pressure level (CPSL). We investigated the relationship between cough sounds and cough flows and the effects of a measurement condition of cough sound, microphone type and participant’s height and gender on CPF estimation accuracy. The results confirmed that the proposed model estimated CPF with a high accuracy. The absolute error between CPFs and estimated CPFs were significantly lower when the microphone distance from the participant’s mouth was within 30 cm than when the distance exceeded 30 cm. Analysis of the model parameters showed that the estimation accuracy was not affected by participant’s height or gender. These results indicate that the proposed model has the potential to improve the feasibility of measuring and assessing CPF. Full article
(This article belongs to the Special Issue Wearable Sensors and Devices for Healthcare Applications)
►▼ Show Figures

Figure 1

Figure 1
<p>Experimental methods. (<b>a</b>) Experiment 1 method. The cough flow measurement is performed with the participants in a sitting position. The participants wear a face mask with an attached flow sensor. Two microphones are installed 30 cm from the point of face mask contact with the face. Microphone 1 is attached to the flow sensor and microphone 2 is fixed to the microphone stand; (<b>b</b>) Experiment 2 method. Microphones are installed 20 cm, 30 cm, 40 cm, 50 cm and 60 cm away from the point of face mask contact with the face; (<b>c</b>) In-ear microphone. The in-ear microphone was used in experiment 3 and fixed at the right external auditory canal; (<b>d</b>) Mini speech microphone. The mini speech microphone was used in experiment 3; (<b>e</b>) Smartphone microphone: The smartphone microphone was used in experiment 3.</p> Full article ">Figure 2
<p>Examples of cough flow and cough sounds measured by microphones 1 and 2. (<b>a</b>) Experimental data of cough flow signals in experiment 1. <span class="html-italic">CPF</span>, cough peak flow; (<b>b</b>) Experimental data of cough sound signals measured by microphone 1 attached to the flow sensor in experiment 1. (<b>c</b>) Experimental data of cough sound signals measured by microphone 2 fixed to the microphone stand in experiment 1; (<b>d</b>) Absolute values of cough sound measured by microphone 1; (<b>e</b>) Envelope of cough sound signals calculated from the absolute cough sound values measured by microphone 1. The cough peak sound pressure level (<span class="html-italic">CPSL</span>) is defined as the maximum value of the envelope.</p> Full article ">Figure 3
<p>Estimation accuracy of Equation (7) calculated from the experimental data measured by microphone 1 attached to the flow sensor. (<b>a</b>) Relationship of <span class="html-italic">CPF</span> and <span class="html-italic">CPSL<sub>microphone1</sub></span>. The solid lines represent the regression curves derived by fitting the coefficients in the proposed model using the Levenberg-Marquardt method based on <span class="html-italic">CPF</span> and <span class="html-italic">CPSL<sub>microphone1</sub></span>. The dotted lines represent 95% confidence bands. <span class="html-italic">CPF</span>, cough peak flow; <span class="html-italic">CPSL<sub>microphone1</sub></span>, cough peak sound pressure level by microphone 1; (<b>b</b>) Relationship between <span class="html-italic">CPF</span> and <span class="html-italic">CPS<sub>microphone1</sub></span>. <span class="html-italic">CPS<sub>microphone1</sub></span>, estimated cough peak flow calculated from <span class="html-italic">CPSL<sub>microphone1</sub></span>.</p> Full article ">Figure 4
<p>Estimation accuracy in each measurement condition. (<b>a</b>–<b>e</b>): Scatter plots of the measured data and the regression results of the proposed model: <span class="html-italic">CPSL</span><sub>20 cm</sub><sub>–60 cm</sub>, cough peak sound pressure level measured by each microphone installed 20 cm, 30 cm, 40 cm, 50 cm and 60 cm from the point of face mask contact with the face; <span class="html-italic">CPS</span><sub>20 cm</sub><sub>–60 cm</sub>, estimated cough peak flow calculated by <span class="html-italic">CPSL</span><sub>20 cm</sub><sub>–60 cm</sub>. Solid lines represent regression lines derived from <span class="html-italic">CPF</span> and <span class="html-italic">CPSL</span>. The variation around the identity line is visibly reduced in graphs (<b>a</b>,<b>b</b>).</p> Full article ">Figure 5
<p>Comparison of the absolute error between each measurement condition. <span class="html-italic">CPF</span>, cough peak flow; <span class="html-italic">CPS</span>, estimated cough peak flow; * <span class="html-italic">p</span> &lt; 0.001.</p> Full article ">Figure 6
<p>The results of experiment 3. <span class="html-italic">CPF</span>, cough peak flow; <span class="html-italic">CPS<sub>in-ear</sub></span>, estimated <span class="html-italic">CPF</span> calculated from cough sound measured using the in-ear microphone; <span class="html-italic">CPS<sub>mini-speech</sub></span>, estimated <span class="html-italic">CPF</span> calculated from cough sound measured using the mini speech microphone; <span class="html-italic">CPS<sub>smartphone</sub></span>, estimated <span class="html-italic">CPF</span> calculated from cough sound measured using the smartphone microphone; (<b>a</b>) Relationship between <span class="html-italic">CPF</span> and <span class="html-italic">CPS<sub>in-ear</sub></span>. (<b>b</b>) Relationship between <span class="html-italic">CPF</span> and <span class="html-italic">CPS<sub>mini-speech</sub></span>. (<b>c</b>) Relationship between <span class="html-italic">CPF</span> and <span class="html-italic">CPS<sub>smart</sub></span>.</p> Full article ">Figure 7
<p>Bland-Altman plot of the measured and estimated cough peak flow. <span class="html-italic">CPF</span>, cough peak flow; <span class="html-italic">CPS</span>, estimated cough peak flow; (<b>a</b>) <span class="html-italic">CPS</span> estimated using our proposed model of Equation (1); (<b>b</b>) <span class="html-italic">CPS</span> estimated using Equation (4). Blue and black dots represent the difference between <span class="html-italic">CPF</span> and <span class="html-italic">CPS</span>. Bold black solid lines represent the mean difference between <span class="html-italic">CPF</span> and <span class="html-italic">CPS</span>. Green dotted lines represent the mean difference ± 2 standard deviation bands. Red lines represent the approximate straight line of the difference between <span class="html-italic">CPF</span> and <span class="html-italic">CPS</span> and the mean of <span class="html-italic">CPF</span> and <span class="html-italic">CPS</span>.</p> Full article ">
18 pages, 5318 KiB  
Article
A Portable Quantum Cascade Laser Spectrometer for Atmospheric Measurements of Carbon Monoxide
by Silvia Viciani, Alessio Montori, Antonio Chiarugi and Francesco D’Amato
Sensors 2018, 18(7), 2380; https://doi.org/10.3390/s18072380 - 21 Jul 2018
Cited by 22 | Viewed by 4730
Abstract
Trace gas concentration measurements in the stratosphere and troposphere are critically required as inputs to constrain climate models. For this purpose, measurement campaigns on stratospheric aircraft and balloons are being carried out all over the world, each one involving sensors which are tailored [...] Read more.
Trace gas concentration measurements in the stratosphere and troposphere are critically required as inputs to constrain climate models. For this purpose, measurement campaigns on stratospheric aircraft and balloons are being carried out all over the world, each one involving sensors which are tailored for the specific gas and environmental conditions. This paper describes an automated, portable, mid-infrared quantum cascade laser spectrometer, for in situ carbon monoxide mixing ratio measurements in the stratosphere and troposphere. The instrument was designed to be versatile, suitable for easy installation on different platforms and capable of operating completely unattended, without the presence of an operator, not only during one flight but for the whole period of a campaign. The spectrometer features a small size (80 × 25 × 41 cm3), light weight (23 kg) and low power consumption (85 W typical), without being pressurized and without the need of calibration on the ground or during in-flight operation. The device was tested in the laboratory and in-field during a research campaign carried out in Nepal in summer 2017, onboard the stratospheric aircraft M55 Geophysica. The instrument worked extremely well, without external maintenance during all flights, proving an in-flight sensitivity of 1–2 ppbV with a time resolution of 1 s. Full article
(This article belongs to the Special Issue Sensors for Emerging Environmental Markers and Contaminants)
►▼ Show Figures

Figure 1

Figure 1
<p>HITRAN simulation of the spectral region around the selected CO line. (<b>a</b>) Simulation of the tropospheric spectrum at 300 mbar (about 10 km altitude). The CO mixing ratio is 100 ppbV, the N<math display="inline"> <semantics> <msub> <mrow/> <mn>2</mn> </msub> </semantics> </math>O mixing ratio is 325 ppbV and we overestimated the water vapour mixing ratio to 1000 ppmV. (<b>b</b>) Simulation of the stratospheric spectrum at 80 mbar (about 20 km altitude). The CO mixing ratio is 20 ppbV, the N<math display="inline"> <semantics> <msub> <mrow/> <mn>2</mn> </msub> </semantics> </math>O mixing ratio is 260 ppbV and the water vapour (not visible) mixing ratio is 10 ppmV.</p> Full article ">Figure 2
<p>(<b>a</b>) Sketch and (<b>b</b>) picture of COLD2 instrument. QCL: Quantum Cascade Laser; BS: beam-splitter; ET: etalon; RC: reference cell; RD: reference detector; MPC: multipass cell; MD: main detector; AL: red laser for alignment; F: humidity filter; LD: QCL driver; TC: QCL temperature controller; cRIO: CompactRIO crate; PS: power supply; P: pump; AP: access point for wireless remote control.</p> Full article ">Figure 3
<p>Picture of the external appearance of COLD2 inserted in the rack, before installation into the dome of the M55 Geophysica stratospheric aircraft.</p> Full article ">Figure 4
<p>Block diagram of the electronics.</p> Full article ">Figure 5
<p>Example of a typical acquired signals. (<b>a</b>) main channel with ambient CO; (<b>b</b>) reference channel with reference cell containing CO at low pressure and etalon.</p> Full article ">Figure 6
<p>Normalized absorption spectra of a 125 ppbV CO mixture (scatter) and Voigt fit results (red line). The residual of the fitting procedure is shown in the bottom graph. The acquisition time is 1 s. The measurement was carried out at ambient pressure and at a temperature of 295 K. The effect of the PZT in the multipass cell is also shown: (<b>a</b>) PZT on; (<b>b</b>) PZT off.</p> Full article ">Figure 7
<p>Normalized absorption spectra of a 125 ppbV CO mixture (scatter) and Voigt fit results (red line). The residual of the fitting procedure is shown in the bottom graph. The acquisition time is 1 s. The measurement was carried out at a pressure of 70 mbar and at a temperature of 295 K. The effect of the PZT in the multipass cell is also shown: (<b>a</b>) PZT on; (<b>b</b>) PZT off.</p> Full article ">Figure 8
<p>Allan-Werle Variance (<math display="inline"> <semantics> <msub> <mi>σ</mi> <mrow> <mi>A</mi> <mi>V</mi> <mi>W</mi> </mrow> </msub> </semantics> </math>) plot of the CO mixing ratio as a function of the integration time, for 30 min measurement of a CO mixture (mixing ratio 125 ppbV) at a rate of 10 Hz, at a temperature of 295 K and for two different pressures: 980 mbar (red open circles) and 50 mbar (blue closed circles)</p> Full article ">Figure 9
<p>Example of a typical in-flight CO (36 ± 1 ppbV) normalized absorption spectrum (scatter) and Voigt fit results (red line). The residual of the fitting procedure is shown in the bottom graph. The acquisition time is 1 s. The measurement was carried out during the flight on the 29th July 2017 at an altitude of 19 km and at a temperature of 281.8 K.</p> Full article ">Figure 10
<p>(<b>a</b>) COLD2 results on board the M55 stratospheric aircraft during the flight on the 6th August 2017 of the StratoClim campaign from Katmandu (Nepal). CO mixing ratio (black line) and aircraft altitude (blue line) vs UTC time. (<b>b</b>) map of the M55 route.</p> Full article ">Figure 11
<p>(<b>a</b>) COLD2 results on board the M55 stratospheric aircraft during the flight on the 8th August 2017 of the StratoClim campaign from Katmandu (Nepal). CO mixing ratio (black line) and aircraft altitude (blue line) vs UTC time. (<b>b</b>) map of the M55 route.</p> Full article ">Figure 12
<p>CO vertical profiles during the whole StratoClim campaign in Nepal in summer 2017 (black scatter). The red line is the CO mean vertical profile.</p> Full article ">
14 pages, 9816 KiB  
Article
Fast Feature-Preserving Approach to Carpal Bone Surface Denoising
by Ibrahim Salim and A. Ben Hamza
Sensors 2018, 18(7), 2379; https://doi.org/10.3390/s18072379 - 21 Jul 2018
Viewed by 4039
Abstract
We present a geometric framework for surface denoising using graph signal processing, which is an emerging field that aims to develop new tools for processing and analyzing graph-structured data. The proposed approach is formulated as a constrained optimization problem whose objective function consists [...] Read more.
We present a geometric framework for surface denoising using graph signal processing, which is an emerging field that aims to develop new tools for processing and analyzing graph-structured data. The proposed approach is formulated as a constrained optimization problem whose objective function consists of a fidelity term specified by a noise model and a regularization term associated with prior data. Both terms are weighted by a normalized mesh Laplacian, which is defined in terms of a data-adaptive kernel similarity matrix in conjunction with matrix balancing. Minimizing the objective function reduces it to iteratively solve a sparse system of linear equations via the conjugate gradient method. Extensive experiments on noisy carpal bone surfaces demonstrate the effectiveness of our approach in comparison with existing methods. We perform both qualitative and quantitative comparisons using various evaluation metrics. Full article
(This article belongs to the Special Issue Sensor Signal and Information Processing)
►▼ Show Figures

Figure 1

Figure 1
<p>Hand model (<b>left</b>) and sparsity pattern plot of its weighted Laplacian matrix (<b>right</b>).</p> Full article ">Figure 2
<p>Flowchart of our proposed surface denoising method, where <math display="inline"><semantics> <mi mathvariant="bold">v</mi> </semantics></math> is the noisy graph signal, and <math display="inline"><semantics> <msup> <mrow> <mi mathvariant="bold">u</mi> </mrow> <mo>⋆</mo> </msup> </semantics></math> is the estimated signal.</p> Full article ">Figure 3
<p>Visualization of four eigenvectors of the normalized mesh Laplacian matrix. From <b>left</b> to <b>right</b>: a 3D hand model Gouraud shaded and color-coded by the values of the second, eighth, fifteenth and twentieth eigenvectors.</p> Full article ">Figure 4
<p>Carpal bone anatomy of a healthy male from a palmar view. The carpus consists of eight carpal bones which are arranged in proximal and distal rows. The proximal row contains the scaphoid (Sp), lunate (Ln), triquetrum (Tq), and pisiform (Pf), while the distal row contains the trapezium (Tm), trapezoid (Td), capitate (Cp), and hamate (Hm). The distal row adjoins the five metacarpals (Mc1-5) of the wrist. The radius (Rd) and ulna (Un) are also shown.</p> Full article ">Figure 5
<p>Carpal bone models.</p> Full article ">Figure 6
<p>Surface denoising results of the noisy right metacarpal model corrupted by Gaussian noise with <math display="inline"><semantics> <mrow> <mi>σ</mi> <mo>=</mo> <mn>0.5</mn> </mrow> </semantics></math>. The magnified views of denoised models show that our method outperformed the baselines in preserving the surface features.</p> Full article ">Figure 7
<p>Surface denoising results for the noisy scaphoid, left metacarpal, and left hamate models. The noise standard deviation was set to <math display="inline"><semantics> <mrow> <mi>σ</mi> <mo>=</mo> <mn>0.5</mn> </mrow> </semantics></math>.</p> Full article ">Figure 8
<p>Surface denoising results for the noisy scaphoid, lunate, and pisiform models. The noise standard deviation was set to <math display="inline"><semantics> <mrow> <mi>σ</mi> <mo>=</mo> <mn>0.7</mn> </mrow> </semantics></math>.</p> Full article ">Figure 9
<p><math display="inline"><semantics> <msup> <mi>L</mi> <mn>2</mn> </msup> </semantics></math> face-normal errors for the left metacarpal model.</p> Full article ">Figure 10
<p><math display="inline"><semantics> <msup> <mi>L</mi> <mn>2</mn> </msup> </semantics></math> face-normal position errors for the scaphoid model.</p> Full article ">Figure 11
<p><math display="inline"><semantics> <msup> <mi>L</mi> <mn>2</mn> </msup> </semantics></math> face-normal errors for the lunate model.</p> Full article ">Figure 12
<p><math display="inline"><semantics> <msup> <mi>L</mi> <mn>2</mn> </msup> </semantics></math> face-normal errors for the right metacarpal model.</p> Full article ">Figure 13
<p><math display="inline"><semantics> <msup> <mi>L</mi> <mn>2</mn> </msup> </semantics></math> face-normal errors for the left hamate model.</p> Full article ">
23 pages, 12853 KiB  
Article
Time-Frequency Energy Sensing of Communication Signals and Its Application in Co-Channel Interference Suppression
by Yue Li, Liang Ye and Xuejun Sha
Sensors 2018, 18(7), 2378; https://doi.org/10.3390/s18072378 - 21 Jul 2018
Cited by 2 | Viewed by 3767
Abstract
As the number of mobile users and video traffics grow explosively, the data rate demands increase tremendously. To improve the spectral efficiency, the spectrum are reused inter cell or intra cell, such as the ultra dense network with multi-cell or the cellular network [...] Read more.
As the number of mobile users and video traffics grow explosively, the data rate demands increase tremendously. To improve the spectral efficiency, the spectrum are reused inter cell or intra cell, such as the ultra dense network with multi-cell or the cellular network with Device-to-Device communications, where the co-channel interferences are brought and needs to be suppressed. According to the time-frequency energy sensing to the communication signals, the desired signal and the interference signal have different energy concentration areas on the time frequency plane, which provide opportunities to suppress the co-channel interference with time varying filter. This paper analyzes the time-frequency distributions of the Gaussian pulse shaping signals, discusses the effect of the analyzing window length on the time-frequency resolution, exploits the equivalence between the time frequency analysis at the baseband and at the radio front end, and finally reveals the advantages of the proposed masking threshold constrained time varying filter based co-channel interference mitigation method. The pass region for the linear time varying filter is generated according to the time-varying energy characteristics of the I/Q separated 4-QAM pulse shaping signals, where the optimum masking threshold is obtained by the optimum-BER criterion. The proposed co-channel interference suppression method is evaluated in aspect of BER performance, and simulation results show that the proposed method outperforms existing methods with low-pass or band-pass filters. Full article
(This article belongs to the Section Sensor Networks)
►▼ Show Figures

Figure 1

Figure 1
<p>The time frequency distribution of a 4-QAM pulse shaping signal. (<b>a</b>) WVD; (<b>b</b>) STFT; (<b>c</b>) CWD; (<b>d</b>) SPWVD.</p> Full article ">Figure 2
<p>The Gaussian pulse waveform with different BTs.</p> Full article ">Figure 3
<p>The baseband waveform and RF waveform of the pulse shaping signal. (<b>a</b>) Data stream <math display="inline"><semantics> <mrow> <mo>{</mo> <mn>1</mn> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mn>1</mn> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mn>1</mn> <mo>}</mo> </mrow> </semantics></math>; (<b>b</b>) data stream <math display="inline"><semantics> <mrow> <mo>{</mo> <mn>1</mn> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mn>1</mn> <mo>}</mo> </mrow> </semantics></math>.</p> Full article ">Figure 4
<p>The time frequency distribution for pulse shaping signals. (<b>a</b>) Data stream <math display="inline"><semantics> <mrow> <mo>{</mo> <mn>1</mn> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mn>1</mn> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mn>1</mn> <mo>}</mo> </mrow> </semantics></math>; (<b>b</b>) data stream <math display="inline"><semantics> <mrow> <mo>{</mo> <mn>1</mn> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mn>1</mn> <mo>}</mo> </mrow> </semantics></math>.</p> Full article ">Figure 5
<p>Time-frequency distribution and frequency deviation of 4-QAM modulated pulse shaping signals. (<b>a</b>) The time frequency distribution of <math display="inline"><semantics> <mrow> <mo>{</mo> <mo>-</mo> <mn>1</mn> <mo>-</mo> <mi>i</mi> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>+</mo> <mi>i</mi> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>-</mo> <mi>i</mi> <mo>}</mo> </mrow> </semantics></math>; (<b>b</b>) The frequency deviation of <math display="inline"><semantics> <mrow> <mo>{</mo> <mo>-</mo> <mn>1</mn> <mo>-</mo> <mi>i</mi> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>+</mo> <mi>i</mi> <mo>,</mo> <mo>-</mo> <mn>1</mn> <mo>-</mo> <mi>i</mi> <mo>}</mo> </mrow> </semantics></math>; (<b>c</b>) The time frequency distribution of <math display="inline"><semantics> <mrow> <mo>{</mo> <mn>1</mn> <mo>-</mo> <mi>i</mi> <mo>,</mo> <mn>1</mn> <mo>-</mo> <mi>i</mi> <mo>,</mo> <mn>1</mn> <mo>+</mo> <mi>i</mi> <mo>}</mo> </mrow> </semantics></math>; (<b>d</b>) The frequency deviation of <math display="inline"><semantics> <mrow> <mo>{</mo> <mn>1</mn> <mo>-</mo> <mi>i</mi> <mo>,</mo> <mn>1</mn> <mo>-</mo> <mi>i</mi> <mo>,</mo> <mn>1</mn> <mo>+</mo> <mi>i</mi> <mo>}</mo> </mrow> </semantics></math>.</p> Full article ">Figure 6
<p>The time frequency distribution of the SC-FDMA signal.</p> Full article ">Figure 7
<p>The effect of carrier leakage. (<b>a</b>) Baseband complex signal, <math display="inline"><semantics> <msub> <mi>L</mi> <mi>h</mi> </msub> </semantics></math> = 40; (<b>b</b>) RF signal, <math display="inline"><semantics> <msub> <mi>L</mi> <mi>h</mi> </msub> </semantics></math> = 40; (<b>c</b>) Baseband complex signal, <math display="inline"><semantics> <msub> <mi>L</mi> <mi>h</mi> </msub> </semantics></math> = 50; (<b>d</b>) RF signal, <math display="inline"><semantics> <msub> <mi>L</mi> <mi>h</mi> </msub> </semantics></math> = 50.</p> Full article ">Figure 8
<p>Realizing time-frequency equivalent analysis of baseband and RF by controlling the window length. (<b>a</b>) Baseband complex signal, <math display="inline"><semantics> <msub> <mi>L</mi> <mi>h</mi> </msub> </semantics></math> = 30; (<b>b</b>) RF signal, <math display="inline"><semantics> <msub> <mi>L</mi> <mi>h</mi> </msub> </semantics></math> = 30; (<b>c</b>) Baseband complex signal, <math display="inline"><semantics> <msub> <mi>L</mi> <mi>h</mi> </msub> </semantics></math> = 60; (<b>d</b>) RF signal, <math display="inline"><semantics> <msub> <mi>L</mi> <mi>h</mi> </msub> </semantics></math> = 60.</p> Full article ">Figure 9
<p>Realizing time-frequency equivalent analysis of baseband and RF by adding Hamming window and zero padding. (<b>a</b>) baseband complex signal, with hamming analysis window and zero padding, STFT, <math display="inline"><semantics> <msub> <mi>L</mi> <mi>h</mi> </msub> </semantics></math> = 40; (<b>b</b>) RF signal, with hamming analysis window and zero padding, STFT, <math display="inline"><semantics> <msub> <mi>L</mi> <mi>h</mi> </msub> </semantics></math> = 40; (<b>c</b>) baseband complex signal, with hamming analysis window and zero padding, CW, <math display="inline"><semantics> <msub> <mi>L</mi> <mi>h</mi> </msub> </semantics></math> = 40; (<b>d</b>) RF signal, with hamming analysis window and zero padding, CW, <math display="inline"><semantics> <msub> <mi>L</mi> <mi>h</mi> </msub> </semantics></math> = 40.</p> Full article ">Figure 10
<p>The pass regions with different masking thresholds. (<b>a</b>) The pass region with <math display="inline"><semantics> <mi>γ</mi> </semantics></math> = 0.04; (<b>b</b>) The pass region with <math display="inline"><semantics> <mi>γ</mi> </semantics></math> = 0.1; (<b>c</b>) The pass region with <math display="inline"><semantics> <mi>γ</mi> </semantics></math> = 0.2; (<b>d</b>) The pass region with <math display="inline"><semantics> <mi>γ</mi> </semantics></math> = 0.26.</p> Full article ">Figure 11
<p>BER varies with different masking thresholds. (<b>a</b>) <math display="inline"><semantics> <mrow> <msub> <mi>E</mi> <mi>b</mi> </msub> <mo>/</mo> <msub> <mi>N</mi> <mn>0</mn> </msub> </mrow> </semantics></math> = 8 dB, SIR = 5 dB; (<b>b</b>) <math display="inline"><semantics> <mrow> <msub> <mi>E</mi> <mi>b</mi> </msub> <mo>/</mo> <msub> <mi>N</mi> <mn>0</mn> </msub> </mrow> </semantics></math> = 8 dB, SIR = 8 dB.</p> Full article ">Figure 12
<p>The pass regions with different analysis windows. (<b>a</b>) The pass region with the rectangular window; (<b>b</b>) The pass region with the Gaussian window; (<b>c</b>) The pass region with the Blackman window; (<b>d</b>) The pass region with the Hanning window.</p> Full article ">Figure 13
<p>The pass regions with different window lengths. (<b>a</b>) The pass region with <span class="html-italic">L<sub>h</sub></span> = 10; (<b>b</b>) The pass region with <span class="html-italic">L<sub>h</sub></span> = 20; (<b>c</b>) The pass region with <span class="html-italic">L<sub>h</sub></span> = 30; (<b>d</b>) The pass region with <span class="html-italic">L<sub>h</sub></span> = 40.</p> Full article ">Figure 13 Cont.
<p>The pass regions with different window lengths. (<b>a</b>) The pass region with <span class="html-italic">L<sub>h</sub></span> = 10; (<b>b</b>) The pass region with <span class="html-italic">L<sub>h</sub></span> = 20; (<b>c</b>) The pass region with <span class="html-italic">L<sub>h</sub></span> = 30; (<b>d</b>) The pass region with <span class="html-italic">L<sub>h</sub></span> = 40.</p> Full article ">Figure 14
<p>Time frequency distributions of the desired and co-channel interference signals.</p> Full article ">Figure 15
<p>The BER performance with the proposed co-channel interference suppression method. (<b>a</b>) BER varies with different SIRs; (<b>b</b>) BER varies with SIRs with different BTs.</p> Full article ">
27 pages, 1051 KiB  
Article
Distributed Compressed Sensing Based Ground Moving Target Indication for Dual-Channel SAR System
by Jing Liu, Xiaoqing Tian, Jiayuan Jiang and Kaiyu Huang
Sensors 2018, 18(7), 2377; https://doi.org/10.3390/s18072377 - 21 Jul 2018
Cited by 2 | Viewed by 4274
Abstract
The dual-channel synthetic aperture radar (SAR) system is widely applied in the field of ground moving-target indication (GMTI). With the increase of the imaging resolution, the resulting substantial raw data samples increase the transmission and storage burden. We tackle the problem by adopting [...] Read more.
The dual-channel synthetic aperture radar (SAR) system is widely applied in the field of ground moving-target indication (GMTI). With the increase of the imaging resolution, the resulting substantial raw data samples increase the transmission and storage burden. We tackle the problem by adopting the joint sparsity model 1 (JSM-1) in distributed compressed sensing (DCS) to exploit the correlation between the two channels of the dual-channel SAR system. We propose a novel algorithm, namely the hierarchical variational Bayesian based distributed compressed sensing (HVB-DCS) algorithm for the JSM-1 model, which decouples the common component from the innovation components by applying variational Bayesian approximation. Using the proposed HVB-DCS algorithm in the dual-channel SAR based GMTI (SAR-GMTI) system, we can jointly reconstruct the dual-channel signals, and simultaneously detect the moving targets and stationary clutter, which enables sampling at a further lower rate in azimuth as well as improves the reconstruction accuracy. The simulation and experimental results show that the proposed HVB-DCS algorithm is capable of detecting multiple moving targets while suppressing the clutter at a much lower data rate in azimuth compared with the compressed sensing (CS) and range-Doppler (RD) algorithms. Full article
(This article belongs to the Special Issue Synthetic Aperture Radar (SAR) Technology: New Perspectives Offered by New-Generation and Forthcoming SAR Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Geometry of the dual-channel SAR system.</p> Full article ">Figure 2
<p>Slant range history geometry of the moving target.</p> Full article ">Figure 3
<p>Random sampling mode in the azimuth direction.</p> Full article ">Figure 4
<p>Graphical model for the JSM-1 model. Doubly circled node represents measurements, while single circled nodes represent hidden variables. Nodes denoted with squares correspond to hyperparameters. The directed edges represent the dependencies among the variables.</p> Full article ">Figure 5
<p>The influence of cross-track velocity on SAR image.</p> Full article ">Figure 6
<p>The influence of along-track velocity on SAR image: (<b>a</b>) <math display="inline"><semantics> <mrow> <msub> <mi>v</mi> <mi>a</mi> </msub> <mo>=</mo> <mn>0</mn> </mrow> </semantics></math> m/s; (<b>b</b>) <math display="inline"><semantics> <mrow> <msub> <mi>v</mi> <mi>a</mi> </msub> <mo>=</mo> <mn>1</mn> </mrow> </semantics></math> m/s; (<b>c</b>) <math display="inline"><semantics> <mrow> <msub> <mi>v</mi> <mi>a</mi> </msub> <mo>=</mo> <mn>2</mn> </mrow> </semantics></math> m/s.</p> Full article ">Figure 7
<p>Relative reconstruction error versus along-track velocity.</p> Full article ">Figure 8
<p>Flow chart of the RD based GMTI system.</p> Full article ">Figure 9
<p>Flow chart of the CS-based GMTI system.</p> Full article ">Figure 10
<p>Flow chart of the HVB-DCS based GMTI system.</p> Full article ">Figure 11
<p>RD based reconstruction results: (<b>a</b>) channel 1 with oversampled raw data; (<b>b</b>) channel 2 with oversampled raw data; (<b>c</b>) DPCA with oversampled raw data.</p> Full article ">Figure 12
<p>CS-based reconstruction results: (<b>a</b>) channel 1 with data rate <math display="inline"><semantics> <mrow> <mn>50</mn> <mo>%</mo> </mrow> </semantics></math> in azimuth; (<b>b</b>) channel 2 with data rate <math display="inline"><semantics> <mrow> <mn>50</mn> <mo>%</mo> </mrow> </semantics></math> in azimuth; (<b>c</b>) DPCA with data rate <math display="inline"><semantics> <mrow> <mn>50</mn> <mo>%</mo> </mrow> </semantics></math> in azimuth.</p> Full article ">Figure 13
<p>HVB-DCS based reconstruction results with data rate <math display="inline"><semantics> <mrow> <mn>37.5</mn> <mo>%</mo> </mrow> </semantics></math> in azimuth: (<b>a</b>) static scattering centers; (<b>b</b>) moving target in channel one; (<b>c</b>) moving target in channel two.</p> Full article ">Figure 14
<p>IFs of the RD, CS and HVB-DCS algorithms, based GMTI system for the varying data rates in azimuth and different <math display="inline"><semantics> <msub> <mi>SCNR</mi> <mi>in</mi> </msub> </semantics></math> levels.</p> Full article ">Figure 15
<p>Reconstruction errors of three algorithms, i.e., RD, CS and HVB-DCS algorithms, based GMTI system for the varying data rates in azimuth and different <math display="inline"><semantics> <msub> <mi>SCNR</mi> <mi>in</mi> </msub> </semantics></math> levels.</p> Full article ">Figure 16
<p>Complexity comparisons of the RD, CS and HVB-DCS algorithms: (<b>a</b>) runtime versus sparse vector dimension <span class="html-italic">N</span>; (<b>b</b>) runtime versus data rate <math display="inline"><semantics> <mrow> <mi>M</mi> <mo>/</mo> <mi>N</mi> </mrow> </semantics></math>.</p> Full article ">Figure 17
<p>Ground truth scene containing the static clutter, five stationary targets and four targets with same cross-track velocity of <math display="inline"><semantics> <mrow> <msub> <mi>v</mi> <mi>r</mi> </msub> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math> m/s, and same along-track velocity of <math display="inline"><semantics> <mrow> <msub> <mi>v</mi> <mi>a</mi> </msub> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math> m/s. The red rectangle indicates the moving target.</p> Full article ">Figure 18
<p>GMTI results: (<b>a</b>) RD with oversampled raw data; (<b>b</b>) CS with <math display="inline"><semantics> <mrow> <mn>50</mn> <mo>%</mo> </mrow> </semantics></math> of the oversampled raw data; (<b>c</b>) HVB-DCS with <math display="inline"><semantics> <mrow> <mn>37.5</mn> <mo>%</mo> </mrow> </semantics></math> of the oversampled raw data. The red rectangle indicates the moving target. The partial enlarged result is shown in the upper right corner of the SAR image.</p> Full article ">
14 pages, 3737 KiB  
Article
Inductive Loop Axle Detector based on Resistance and Reactance Vehicle Magnetic Profiles
by Zbigniew Marszalek, Tadeusz Zeglen, Ryszard Sroka and Janusz Gajda
Sensors 2018, 18(7), 2376; https://doi.org/10.3390/s18072376 - 21 Jul 2018
Cited by 11 | Viewed by 7779
Abstract
The article presents a measurement system that captures two components of a motor vehicle’s magnetic profile, which are associated with the real and imaginary part of the impedance of a narrow inductive loop sensor. The proposed algorithm utilizes both components of the impedance [...] Read more.
The article presents a measurement system that captures two components of a motor vehicle’s magnetic profile, which are associated with the real and imaginary part of the impedance of a narrow inductive loop sensor. The proposed algorithm utilizes both components of the impedance magnetic profile to detect vehicle axles, including lifted axles. Accuracies of no less than 71.8% were achieved for vehicles travelling with a lifted axle, and no less than 98.8% for other vehicles. The axle detection accuracy was determined during a series of experiments carried out under normal traffic conditions, using profile analysis, video footage and reference signals from an axle load detector on a total of 4000 vehicles. Full article
(This article belongs to the Special Issue Intelligent Sensor Systems for Environmental Monitoring)
►▼ Show Figures

Figure 1

Figure 1
<p>Sensor amplifier, where QD—Quadrature Demodulator (details shown in <a href="#sensors-18-02376-f002" class="html-fig">Figure 2</a>).</p> Full article ">Figure 2
<p>Block diagram of the conditioning system cooperating with the IL sensor comprised of a sensor amplifier (<a href="#sensors-18-02376-f001" class="html-fig">Figure 1</a>), where: the C-S block provides two signals shifted by 90 degrees in relation to each other; D1, D2—synchronous demodulators; LPF1, LPF2—low-pass filters.</p> Full article ">Figure 3
<p>Characteristics of inductive sensor resistance measurement channel: (<b>a</b>) static; (<b>b</b>) relative nonlinearity error.</p> Full article ">Figure 4
<p>Characteristics of inductive sensor inductance measurement channel: (<b>a</b>) static; (<b>b</b>) relative nonlinearity error.</p> Full article ">Figure 5
<p>Multichannel measurement system: (<b>a</b>) multichannel conditioning system; (<b>b</b>) single card of the system; (<b>c</b>) graphical user interface of the system; (<b>d</b>) research stand diagram, where: IL1—sensor for vehicle detection; IL2 and IL3—narrow IL sensors for axle detection; P1 and P2—piezoelectric load sensors for axle detection; MC—multichannel conditioning system (shown in <a href="#sensors-18-02376-f005" class="html-fig">Figure 5</a>a) equipped with dedicated cards for individual sensors (VD—vehicle detector card; AD1 and AD2—axle detection cards, PD—card for piezoelectric sensors).</p> Full article ">Figure 5 Cont.
<p>Multichannel measurement system: (<b>a</b>) multichannel conditioning system; (<b>b</b>) single card of the system; (<b>c</b>) graphical user interface of the system; (<b>d</b>) research stand diagram, where: IL1—sensor for vehicle detection; IL2 and IL3—narrow IL sensors for axle detection; P1 and P2—piezoelectric load sensors for axle detection; MC—multichannel conditioning system (shown in <a href="#sensors-18-02376-f005" class="html-fig">Figure 5</a>a) equipped with dedicated cards for individual sensors (VD—vehicle detector card; AD1 and AD2—axle detection cards, PD—card for piezoelectric sensors).</p> Full article ">Figure 6
<p>Examples of R- and X-profiles from narrow IL sensor: (<b>a</b>) passenger vehicle (low suspension); (<b>b</b>) truck vehicle (high suspension); where: AA—refers to the areas where there are artifacts originating from the vehicle axles.</p> Full article ">Figure 7
<p>Algorithm-core of vehicle axle detection: (<b>a</b>) functional diagram; (<b>b</b>) example illustrating the principle of operation and signal processing, where: R and X-profile are input signals, gain, level, and hist are the setting parameters, A and K<sub>N</sub> are the output signals.</p> Full article ">Figure 8
<p>Vehicle axle detection algorithm based on R- and X-profiles.</p> Full article ">
16 pages, 3457 KiB  
Article
Three-Dimensional Registration of Freehand-Tracked Ultrasound to CT Images of the Talocrural Joint
by Nazlı Tümer, Aimee C. Kok, Frans M. Vos, Geert J. Streekstra, Christian Askeland, Gabrielle J. M. Tuijthof and Amir A. Zadpoor
Sensors 2018, 18(7), 2375; https://doi.org/10.3390/s18072375 - 21 Jul 2018
Cited by 4 | Viewed by 4588
Abstract
A rigid surface–volume registration scheme is presented in this study to register computed tomography (CT) and free-hand tracked ultrasound (US) images of the talocrural joint. Prior to registration, bone surfaces expected to be visible in US are extracted from the CT volume and [...] Read more.
A rigid surface–volume registration scheme is presented in this study to register computed tomography (CT) and free-hand tracked ultrasound (US) images of the talocrural joint. Prior to registration, bone surfaces expected to be visible in US are extracted from the CT volume and bone contours in 2D US data are enhanced based on monogenic signal representation of 2D US images. A 3D monogenic signal data is reconstructed from the 2D data using the position of the US probe recorded with an optical tracking system. When registering the surface extracted from the CT scan to the monogenic signal feature volume, six transformation parameters are estimated so as to optimize the sum of monogenic signal features over the transformed surface. The robustness of the registration algorithm was tested on a dataset collected from 12 cadaveric ankles. The proposed method was used in a clinical case study to investigate the potential of US imaging for pre-operative planning of arthroscopic access to talar (osteo)chondral defects (OCDs). The results suggest that registrations with a registration error of 2 mm and less is achievable, and US has the potential to be used in assessment of an OCD’ arthroscopic accessibility, given the fact that 51% of the talar surface could be visualized. Full article
(This article belongs to the Section Biosensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Overview of the rigid surface–volume registration scheme to match computed tomography (CT) and ultrasound (US) images of the talocrural joints. CT and US images are pre-processed before the registration. During the pre-processing step, surfaces of bones (i.e., tibia and talus) that can be visualized with US in maximal-plantar flexed ankle are extracted from the CT image and bone contours in freehand tracked 2D US images are enhanced using intensity invariant local-phase based approach and bone shadow information. The 3D bone response data is reconstructed from 2D enhanced US images using the position of the US probe that had been recorded with an optical tracking system. Registration is initialized at a location defined using the position of the six fiducials in the US and CT spaces.</p> Full article ">Figure 2
<p>Experimental settings in the ultrasound room. (<b>a</b>) Prior to CT imaging and US sweeping, each cadaveric ankle was placed on a foot plate in maximal plantar flexion and was tightened using straps. The US probe on which the position sensor mounted was slowly swept over the cadaveric ankles and the position of the probe was recorded using (<b>b</b>) the optical tracking camera.</p> Full article ">Figure 3
<p>Bone contour enhancement in a 2D US image. (<b>a</b>) An original 2D US image (<b>b</b>) Phase Symmetry (<span class="html-italic">PS</span>) map calculated based on monogenic signal representation of the 2D US image. (<b>c</b>) Bone response map obtained based on the product of <span class="html-italic">PS</span> and the shadow values (<span class="html-italic">SH</span>).</p> Full article ">Figure 4
<p>Sagittal view of one of the US volumes registered to the CT scan. The visible cartilage and cartilage covered by the tibia are represented, respectively, with <span class="html-italic">α</span> and <span class="html-italic">β.</span> The percentage of the visible cartilage was defined as the ratio of <span class="html-italic">α</span> to <span class="html-italic">α + β</span> (i.e., the total cartilage surface).</p> Full article ">Figure 5
<p>A visualization of one of the successful registrations achieved in Test I. (<b>a</b>) the 3D view (<b>b</b>) Axial view (<b>c</b>) Coronal view and (<b>d</b>) Sagittal view. US volume in purple color represents the data at initialization position prior to registration. US volume in green depicts the result of the registration. Dashed lines in yellow highlight the contours of the tibia and talus observed in both axial and sagittal views. Arrows in red shows that US volume goes from its initial position (i.e., US volume in purple) to its final position (i.e., US volume in green).</p> Full article ">Figure 6
<p>Plots of the <math display="inline"><semantics> <mrow> <mi>d</mi> <msub> <mi>f</mi> <mrow> <mi>R</mi> <mi>M</mi> <mi>S</mi> </mrow> </msub> </mrow> </semantics></math> values vs. the <math display="inline"><semantics> <mrow> <mi>d</mi> <msub> <mi>i</mi> <mrow> <mi>R</mi> <mi>M</mi> <mi>S</mi> </mrow> </msub> </mrow> </semantics></math> values calculated over 100 registrations are given for (<b>a</b>) the ‘Sweep Type I’ US sweeps (<b>b</b>) the ‘Sweep Type II’ US sweeps showing the best and worst success rate in Test I. Black (<span class="html-italic">n</span> = 100) and red (<span class="html-italic">n</span> = 100) dots in (<b>a</b>) are samples retrieved from a set of registrations performed for the ‘Sweep Type I’ US sweeps with the best and worst success rate, respectively. In a similar manner, those (<math display="inline"><semantics> <mrow> <msub> <mi>n</mi> <mrow> <mi>b</mi> <mi>l</mi> <mi>a</mi> <mi>c</mi> <mi>k</mi> </mrow> </msub> </mrow> </semantics></math> <span class="html-italic">=</span> 100 and <math display="inline"><semantics> <mrow> <msub> <mi>n</mi> <mrow> <mi>r</mi> <mi>e</mi> <mi>d</mi> </mrow> </msub> </mrow> </semantics></math> = 100) in (<b>b</b>) are related to a set of registrations run for the ‘Sweep Type II’ US sweeps with the best and worst success rate. Visualizations of 3D bone response volumes of the ‘Sweep Type I’ US sweeps (<b>c</b>) with the best and (<b>d</b>) the worst success rate, and of the ‘Sweep Type II’ US sweeps (<b>e</b>) with the best and (<b>f</b>) the worst success rate in the Test I are shown. Original 2D slices taken from the US sweeps (<b>c</b>–<b>f</b>) are presented in a corresponding manner (<b>g</b>–<b>j</b>). The intensity profiles of the two vertical lines (i.e., intensity vs. distance along profile graphs shown in red and blue correspond to the lines drawn in red and blue, respectively) displayed in (<b>h</b>) and (<b>j</b>) indicate that the ridge-like edge feature corresponding to the expected bone boundary can be weaker than those related to the soft tissue interface. Thin and thick arrows shown in green and purple point to the ridge edge features linked, respectively, to soft tissue interface and bone boundary.</p> Full article ">
14 pages, 4658 KiB  
Article
Design and Fabrication Technology of Low Profile Tactile Sensor with Digital Interface for Whole Body Robot Skin
by Mitsutoshi Makihata, Masanori Muroyama, Shuji Tanaka, Takahiro Nakayama, Yutaka Nonomura and Masayoshi Esashi
Sensors 2018, 18(7), 2374; https://doi.org/10.3390/s18072374 - 21 Jul 2018
Cited by 24 | Viewed by 6232
Abstract
Covering a whole surface of a robot with tiny sensors which can measure local pressure and transmit the data through a network is an ideal solution to give an artificial skin to robots to improve a capability of action and safety. The crucial [...] Read more.
Covering a whole surface of a robot with tiny sensors which can measure local pressure and transmit the data through a network is an ideal solution to give an artificial skin to robots to improve a capability of action and safety. The crucial technological barrier is to package force sensor and communication function in a small volume. In this paper, we propose the novel device structure based on a wafer bonding technology to integrate and package capacitive force sensor using silicon diaphragm and an integrated circuit separately manufactured. Unique fabrication processes are developed, such as the feed-through forming using a dicing process, a planarization of the Benzocyclobutene (BCB) polymer filled in the feed-through and a wafer bonding to stack silicon diaphragm onto ASIC (application specific integrated circuit) wafer. The ASIC used in this paper has a capacitance measurement circuit and a digital communication interface mimicking a tactile receptor of a human. We successfully integrated the force sensor and the ASIC into a 2.5×2.5×0.3 mm die and confirmed autonomously transmitted packets which contain digital sensing data with the linear force sensitivity of 57,640 Hz/N and 10 mN of data fluctuation. A small stray capacitance of 1.33 pF is achieved by use of 10 μm thick BCB isolation layer and this minimum package structure. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Figure 1

Figure 1
<p>A tactile sensor network with small and smart sensor dies: (<b>a</b>) one-dimensional sensor array on a flexible cable; (<b>b</b>) tactile sensors with a force sensor and bus network interface on a chip.</p> Full article ">Figure 2
<p>Device structure: (<b>a</b>) cross section of the chip; (<b>b</b>) top side; (<b>c</b>) back side.</p> Full article ">Figure 3
<p>Mechanical and circuit simulation: (<b>a</b>) mechanical simulation of the force sensing element; (<b>b</b>) capacitor–frequency simulation extracted from the layout of Schmitt trigger oscillator in the ASIC; (<b>c</b>) calculated change of digital counter under an external force on a diaphragm; (<b>d</b>) influence of parasitic capacitance on sensitivity.</p> Full article ">Figure 4
<p>Procedure of the through silicon groove (TSG) technology: (<b>a</b>) groove forming; (<b>b</b>) insulation and rewiring of I/O pad into groove; (<b>c</b>) groove filling with polymer and planarization; (<b>d</b>) back grinding.</p> Full article ">Figure 5
<p>Filling of grooves with BCB and anti-swelling layer: (<b>a</b>) procedure of TSG forming; (<b>b</b>) swelling at groove due to swelling of BCB; (<b>c</b>) suppression of swelling by anti-swelling layer.</p> Full article ">Figure 6
<p>The full process chart of wafer level integration and packaging for tactile sensors.</p> Full article ">Figure 7
<p>Prototyped tactile sensor: (<b>a</b>) front; (<b>b</b>) back; and (<b>c</b>) cross sectional view.</p> Full article ">Figure 8
<p>Experimental data with prototyped sensors: (<b>a</b>) a 45 MHz digital packet; (<b>b</b>) linear correlation between encoded counter value and external force; (<b>c</b>) sensitivity and data fluctuation with various sampling times.</p> Full article ">Figure 9
<p>Response to an external force with threshold operation.</p> Full article ">
16 pages, 4641 KiB  
Article
A 2D Magneto-Acousto-Electrical Tomography Method to Detect Conductivity Variation Using Multifocus Image Method
by Ming Dai, Xin Chen, Tong Sun, Lingyao Yu, Mian Chen, Haoming Lin and Siping Chen
Sensors 2018, 18(7), 2373; https://doi.org/10.3390/s18072373 - 21 Jul 2018
Cited by 19 | Viewed by 4283
Abstract
As magneto-acoustic-electrical tomography (MAET) combines the merits of high contrast and high imaging resolution, and is extremely useful for electrical conductivity measurement, so it is expected to be a promising medical imaging modalities for diagnosis of early-stage cancer. Based on the Verasonics system [...] Read more.
As magneto-acoustic-electrical tomography (MAET) combines the merits of high contrast and high imaging resolution, and is extremely useful for electrical conductivity measurement, so it is expected to be a promising medical imaging modalities for diagnosis of early-stage cancer. Based on the Verasonics system and the MC600 displacement platform, we designed and implemented a MAET system with a chirp pulse stimulation (MAET-CPS) method and a focal probe was utilized for stepscan focus excitation to enhance the imaging resolution. The relevant experiments were conducted to explore the influence of excitation positions of the single-focus point, and the effect of the excitation position on the amplitudes of the conductivity variation was clearly demonstrated. In order to take advantage of the merits of multifocus imaging, we firstly proposed a single focus MAET system with a chirp pulse stimulation (sfMAET-CPS) method and a multifocus MAET system with a chirp pulse stimulation (mfMAET-CPS) method for high-resolution conductivity imaging, and a homogenous gelatin phantom with a cuboid-shaped hole was used to investigate the accuracy of mfMAET-CPS. Comparative experiments were carried out on the same uniform phantom by the sfMAET-CPS and the mfMAET-CPS, respectively. The results showed that: (1) the electrical conductivity distributions of the homogenous phantom with a cuboid-shaped hole were detected by the sfMAET-CPS but were easily affected by the focal point, which demonstrated that the sfMAET-CPS had a low imaging resolution. (2) Compared with the sfMAET-CPS, the imaging effect of the mfMAET-CPS was much better than that of the sfMAET-CPS. (3) A linear interpolation algorithm was used to process the 2D conductivity distribution; it increased the smoothness of the conductivity distribution and improved the imaging effect. The stepscan focus excitation and the linearly frequency-modulated theory provide an alternative scheme for the clinical application of MAET. Full article
(This article belongs to the Section Biosensors)
►▼ Show Figures

Figure 1

Figure 1
<p>Magneto-acoustic-electrical tomography (MAET) imaging process using linearly frequency-modulated theory (MAET-chirp pulse stimulation (CPS)): (<b>a</b>) principle of the MAET-CPS; (<b>b</b>) frequency characteristics of the stimulating signal and the received signal.</p> Full article ">Figure 2
<p>System composition of the MAET-CPS detection system: (<b>a</b>) physical diagram; (<b>b</b>) connection diagram; and (<b>c</b>) the detection front end.</p> Full article ">Figure 3
<p>Algorithm flow chart of the single-focus B-scan imaging MAET with chirp pulse stimulation (sfMAET-CPS) and the multifocus imaging MAET with CPS (mfMAET-CPS).</p> Full article ">Figure 4
<p>The influence of single focus on the experiment: (<b>a</b>) the movement track of the focal point position, which passed through a cuboid-shaped hole; and (<b>b</b>) the brightness curves with the interface information.</p> Full article ">Figure 5
<p>The movement track of the focal point position and the amplitude variation of conductivity curves of 10 step-scan movements in the depth direction.</p> Full article ">Figure 6
<p>The imaging process of the sfMAET-CPS.</p> Full article ">Figure 7
<p>The imaging process of mfMAET-CPS.</p> Full article ">Figure 8
<p>The motion control and image synthesis processes: (<b>a</b>) the sfMAET-CPS; and (<b>b</b>) the mfMAET-CPS.</p> Full article ">Figure 9
<p>The stimulating positions of two experiments.</p> Full article ">Figure 10
<p>Results of accuracy measurement of mfMAET-CPS: (<b>a</b>) two interfaces detected by mfMAET-CPS; and (<b>b</b>) four interfaces detected by mfMAET-CPS.</p> Full article ">Figure 11
<p>The uniform phantom for testing.</p> Full article ">Figure 12
<p>Imaging results: (<b>a</b>) conductivity distribution of the sfMAET-CPS; (<b>b</b>) conductivity distribution of the sfMAET-CPS after being processed by the linear interpolation algorithm; (<b>c</b>) conductivity distribution of the mfMAET-CPS; and (<b>d</b>) conductivity distribution of the mfMAET-CPS after being processed by the linear interpolation algorithm.</p> Full article ">
13 pages, 3641 KiB  
Article
A Colorimetric Probe Based on Functionalized Gold Nanorods for Sensitive and Selective Detection of As(III) Ions
by Kun Ge, Jingmin Liu, Guozhen Fang, Peihua Wang, Dongdong Zhang and Shuo Wang
Sensors 2018, 18(7), 2372; https://doi.org/10.3390/s18072372 - 21 Jul 2018
Cited by 18 | Viewed by 5286
Abstract
A colorimetric probe for determination of As(III) ions in aqueous solutions on basis of localized surface plasmon resonance (LSPR) was synthesized. The dithiothreitol molecules with two end thiols covalently combined with Au Nanorods (AuNRs) with an aspect ratio of 2.9 by Au-S bond [...] Read more.
A colorimetric probe for determination of As(III) ions in aqueous solutions on basis of localized surface plasmon resonance (LSPR) was synthesized. The dithiothreitol molecules with two end thiols covalently combined with Au Nanorods (AuNRs) with an aspect ratio of 2.9 by Au-S bond to form dithiothreitol coated Au Nanorods (DTT-AuNRs), acting as colorimetric probe for the determination of As(III) ions. With the adding of As(III) ions, the AuNRs will be aggregated and leading the longitudinal SPR absorption band of DTT-AuNRs decrease due to the As(III) ions can bind with three DTT molecules through an As-S linkage. The potential factors affect the response of DTT-AuNRs to As(III) ions including the concentration of DTT, pH values of DTT-AuNRs, reaction time and NaCl concentration were optimized. Under optimum assay conditions, the DTT-AuNRs colorimetric probe has high sensitivity towards As(III) ions with low detection limit of 38 nM by rules of 3σ/k and excellent linear range of 0.13–10.01 μM. The developed colorimetric probe shows high selectivity for As(III) ions sensing and has applied to determine of As(III) in environmental water samples with quantitative spike-recoveries range from 95.2% to 100.4% with low relative standard deviation of less than 4.4% (n = 3). Full article
(This article belongs to the Special Issue Colorimetric Nanosensors)
►▼ Show Figures

Figure 1

Figure 1
<p>The schematic mechanism of determination of As(III) by DTT-AuNRs colorimetric probe.</p> Full article ">Figure 2
<p>(<b>a</b>) TEM image of AuNRs; (<b>b</b>) Size distribution of AuNRs and DTT-AuNRs exposed to As(III); (<b>c</b>) UV–vis absorption spectra of CTAB-coating AuNRs; (<b>d</b>) UV–vis absorption spectra of AuNRs and DTT-AuNRs.</p> Full article ">Figure 3
<p>(<b>a</b>) UV–vis absorption spectra of DTT-AuNRs (0.8 nM) upon adding increasing concentration of As(III) ions (from 0.13 to 10.01 μM); (<b>b</b>) Effects of concentration of DTT on UV-vis spectra of AuNRs.</p> Full article ">Figure 4
<p>Evaluation of stability of DTT-AuNRs colorimetric probe: (<b>a</b>) Effect of the pH value of NaAc-HAc buffer (10 mM); (<b>b</b>) Effect of the reaction time; (<b>c</b>) Effect of the concentration of NaCl.</p> Full article ">Figure 5
<p>Effects of pH and buffer media on UV-vis spectra of AuNRs: (<b>a</b>) 10 mM PBS; (<b>b</b>) 10 mM Tris-HCl; (<b>c</b>) BR buffer; (<b>d</b>) 10 mM NaAc-HAc.</p> Full article ">Figure 6
<p>Optimization of developed AuNR probe for the colorimetric detection of As(III): (<b>a</b>) Effect of the concentration of DTT; (<b>b</b>) Effect of the pH value of NaAc-HAc buffer (10 mM); (<b>c</b>) Effect of the reaction time; (<b>d</b>) Effect of the concentration of NaCl.</p> Full article ">Figure 7
<p>Selectivity test of the developed colorimetric probe for As(III) ions (6.67 μM) over other metal ions (50 μM, except for Hg<sup>2+</sup>, 25 μM). Black bars denote the responses of individual metal ions, while red bars show the responses of As(III) (6.67 μM) in the presence of other metal ions.</p> Full article ">Figure 8
<p>Plot of decreased longitudinal SPR absorption intensity (ΔA) against As(III) concentration over the linear range of 0.13–10.01 μM.</p> Full article ">
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-422
Previous Issue
Next Issue
Back to TopTop