Sensors

Taste is a key sense that helps identify different food types and most of this work was carried out on primary tastes rather than generating different flavors. In this work, we proposed a plan to create other flavors rather than primary tastes, adjusted the electrical (40–180 µA) and thermal stimulation (20–38 °C and 38–20 °C), and revealed the digital coding for multi-flavors. Our results showed that different combinations of digital coding could generate different flavors and that tastes related to different stimuli are easy to develop. The novelty of this work is to design other types of flavors and primary tastes. The experimental results demonstrated that the novel method proposed for digital taste coding could realize primary tastes (sweet, sour, salty, spicy, and mint) and mixed flavors. Furthermore, some innovative sensations have been realized, which are sprite, soda water, sweet-sour, salty-sweet, and salty-mint sensations. We presume that this innovation could digitally enhance various flavors. Full article

There are currently no standard methods for evaluating gait and balance performance at home. Smartphones include acceleration sensors and may represent a promising and easily accessible tool for this purpose. We performed an interventional feasibility study and compared a smartphone-based approach with two standard gait analysis systems (force plate and motion capturing systems). Healthy adults (n = 25, 44.1 ± 18.4 years) completed two laboratory evaluations before and after a three-week gait and balance training at home. There was an excellent agreement between all systems for stride time and cadence during normal, tandem and backward gait, whereas correlations for gait velocity were lower. Balance variables of both standard systems were moderately intercorrelated across all stance tasks, but only few correlated with the corresponding smartphone measures. Significant differences over time were found for several force plate and mocap system-obtained gait variables of normal, backward and tandem gait. Changes in balance variables over time were more heterogeneous and not significant for any system. The smartphone seems to be a suitable method to measure cadence and stride time of different gait, but not balance, tasks in healthy adults. Additional optimizations in data evaluation and processing may further improve the agreement between the analysis systems. Full article

Current methods of measuring heart rate (HR) and oxygen levels ($SP{O}_2$) require physical contact, are individualised, and for accurate oxygen levels may also require a blood test. No-touch or non-invasive technologies are not currently commercially available for use in healthcare settings. To date, there has been no assessment of a system that measures HR and $SP{O}_2$ using commercial off-the-shelf camera technology that utilises R, G, B, and IR data. Moreover, no formal remote photoplethysmography studies have been performed in real-life scenarios with participants at home with different demographic characteristics. This novel study addresses all these objectives by developing, optimising, and evaluating a system that measures the HR and $SP{O}_2$ of 40 participants. HR and $SP{O}_2$ are determined by measuring the frequencies from different wavelength band regions using FFT and radiometric measurements after pre-processing face regions of interest (forehead, lips, and cheeks) from colour, IR, and depth data. Detrending, interpolating, hamming, and normalising the signal with FastICA produced the lowest RMSE of 7.8 for HR with the r-correlation value of 0.85 and RMSE 2.3 for $SP{O}_2$. This novel system could be used in several critical care settings, including in care homes and in hospitals and prompt clinical intervention as required. Full article

Balance and posture are two topics that have been extensively studied, although with some conflicting findings. Therefore, the aim of this work is to analyze the relationship between the postural angles of the spine in the sagittal plane and the stable static balance. A cross-sectional study was conducted with children and adolescents from schools in northern Portugal in 2019. An online questionnaire was used to characterize the sample and analyze back pain. Spinal postural angle assessment (pelvic, lumbar, and thoracic) was performed using the Spinal Mouse^®, while stabilometry assessment was performed using Namrol^® Podoprint^®. Statistical significance was set as α = 0.05. The results showed that girls have better balance variables. There is a weak correlation between the anthropometric variables with stabilometry variables and the postural angles. This correlation is mostly negative, except for the thoracic spine with anthropometric variables and the lumbar spine with BMI. The results showed that postural angles of the spine are poor predictors of the stabilometric variables. Concerning back pain, increasing the postural angle of the thoracic spine increases the odds ratio of manifestation of back pain by 3%. Full article

The classification of surface myoelectric signals (sEMG) remains a great challenge when focused on its implementation in an electromechanical hand prosthesis, due to its nonlinear and stochastic nature, as well as the great difference between models applied offline and online. In this work, the selection of the set of the features that allowed us to obtain the best results for the classification of this type of signals is presented. In order to compare the results obtained, the Nina PRO DB2 and DB3 databases were used, which contain information on 50 different movements of 40 healthy subjects and 11 amputated subjects, respectively. The sEMG of each subject was acquired through 12 channels in a bipolar configuration. To carry out the classification, a convolutional neural network (CNN) was used and a comparison of four sets of features extracted in the time domain was made, three of which have shown good performance in previous works and one more that was used for the first time to train this type of network. Set one is composed of six features in the time domain (TD1), Set two has 10 features also in the time domain (TD2) including the autoregression model (AR), the third set has two features in the time domain derived from spectral moments (TD-PSD1), and finally, a set of five features also has information on the power spectrum of the signal obtained in the time domain (TD-PSD2). The selected features in each set were organized in four different ways for the formation of the training images. The results obtained show that the set of features TD-PSD2 obtained the best performance for all cases. With the set of features and the formation of images proposed, an increase in the accuracies of the models of 8.16% and 8.56% was obtained for the DB2 and DB3 databases, respectively, compared to the current state of the art that has used these databases. Full article

In anchor-free object detection, the center regions of bounding boxes are often highly weighted to enhance detection quality. However, the central area may become less significant in some situations. In this paper, we propose a novel dual attention-based approach for the adaptive weight assignment within bounding boxes. The proposed improved dual attention mechanism allows us to thoroughly untie spatial and channel attention and resolve the confusion issue, thus it becomes easier to obtain the proper attention weights. Specifically, we build an end-to-end network consisting of backbone, feature pyramid, adaptive weight assignment based on dual attention, regression, and classification. In the adaptive weight assignment module based on dual attention, a parallel framework with the depthwise convolution for spatial attention and the 1D convolution for channel attention is applied. The depthwise convolution, instead of standard convolution, helps prevent the interference between spatial and channel attention. The 1D convolution, instead of fully connected layer, is experimentally proved to be both efficient and effective. With the adaptive and proper attention, the correctness of object detection can be further improved. On public MS-COCO dataset, our approach obtains an average precision of 52.7%, achieving a great increment compared with other anchor-free object detectors. Full article

In this manuscript, an underwater target tracking problem with passive sensors is considered. The measurements used to track the target trajectories are (i) only bearing angles, and (ii) Doppler-shifted frequencies and bearing angles. Measurement noise is assumed to follow a zero mean Gaussian probability density function with unknown noise covariance. A method is developed which can estimate the position and velocity of the target along with the unknown measurement noise covariance at each time step. The proposed estimator linearises the nonlinear measurement using an orthogonal polynomial of first order, and the coefficients of the polynomial are evaluated using numerical integration. The unknown sensor noise covariance is estimated online from residual measurements. Compared to available adaptive sigma point filters, it is free from the Cholesky decomposition error. The developed method is applied to two underwater tracking scenarios which consider a nearly constant velocity target. The filter’s efficacy is evaluated using (i) root mean square error (RMSE), (ii) percentage of track loss, (iii) normalised (state) estimation error squared (NEES), (iv) bias norm, and (v) floating point operations (flops) count. From the simulation results, it is observed that the proposed method tracks the target in both scenarios, even for the unknown and time-varying measurement noise covariance case. Furthermore, the tracking accuracy increases with the incorporation of Doppler frequency measurements. The performance of the proposed method is comparable to the adaptive deterministic support point filters, with the advantage of a considerably reduced flops requirement. Full article

Cloud computing coupled with Internet of Things technology provides a wide range of cloud services such as memory, storage, computational processing, network bandwidth, and database application to the end users on demand over the Internet. More specifically, cloud computing provides efficient services such as “pay as per usage”. However, Utility providers in Smart Grid are facing challenges in the design and implementation of such architecture in order to minimize the cost of underlying hardware, software, and network services. In Smart Grid, smart meters generate a large volume of different traffics, due to which efficient utilization of available resources such as buffer, storage, limited processing, and bandwidth is required in a cost-effective manner in the underlying network infrastructure. In such context, this article introduces a QoS-aware Hybrid Queue Scheduling (HQS) model that can be seen over the IoT-based network integrated with cloud environment for different advanced metering infrastructure (AMI) application traffic, which have different QoS levels in the Smart Grid network. The proposed optimization model supports, classifies, and prioritizes the AMI application traffic. The main objective is to reduce the cost of buffer, processing power, and network bandwidth utilized by AMI applications in the cloud environment. For this, we developed a simulation model in the CloudSim simulator that uses a simple mathematical model in order to achieve the objective function. During the simulations, the effects of various numbers of cloudlets on the cost of virtual machine resources such as RAM, CPU processing, and available bandwidth have been investigated in cloud computing. The obtained simulation results exhibited that our proposed model successfully competes with the previous schemes in terms of minimizing the processing, memory, and bandwidth cost by a significant margin. Moreover, the simulation results confirmed that the proposed optimization model behaves as expected and is realistic for AMI application traffic in the Smart Grid network using cloud computing. Full article

Length of Stay (LOS) is an important performance metric in Australian Emergency Departments (EDs). Recent evidence suggests that an LOS in excess of 4 h may be associated with increased mortality, but despite this, the average LOS continues to remain greater than 4 h in many EDs. Previous studies have found that Data Mining (DM) can be used to help hospitals to manage this metric and there is continued research into identifying factors that cause delays in ED LOS. Despite this, there is still a lack of specific research into how DM could use these factors to manage ED LOS. This study adds to the emerging literature and offers evidence that it is possible to predict delays in ED LOS to offer Clinical Decision Support (CDS) by using DM. Sixteen potentially relevant factors that impact ED LOS were identified through a literature survey and subsequently used as predictors to create six Data Mining Models (DMMs). An extract based on the Victorian Emergency Minimum Dataset (VEMD) was used to obtain relevant patient details and the DMMs were implemented using the Weka Software. The DMMs implemented in this study were successful in identifying the factors that were most likely to cause ED LOS > 4 h and also identify their correlation. These DMMs can be used by hospitals, not only to identify risk factors in their EDs that could lead to ED LOS > 4 h, but also to monitor these factors over time. Full article

Unlimited access to information and data sharing wherever and at any time for anyone and anything is a fundamental component of fifth-generation (5G) wireless communication and beyond. Therefore, it has become inevitable to exploit the super-high frequency (SHF) and millimeter-wave (mmWave) frequency bands for future wireless networks due to their attractive ability to provide extremely high data rates because of the availability of vast amounts of bandwidth. However, due to the characteristics and sensitivity of wireless signals to the propagation effects in these frequency bands, more accurate path loss prediction models are vital for the planning, evaluating, and optimizing future wireless communication networks. This paper presents and evaluates the performance of several well-known machine learning methods, including multiple linear regression (MLR), polynomial regression (PR), support vector regression (SVR), as well as the methods using decision trees (DT), random forests (RF), K-nearest neighbors (KNN), artificial neural networks (ANN), and artificial recurrent neural networks (RNN). RNNs are mainly based on long short-term memory (LSTM). The models are compared based on measurement data to provide the best fitting machine-learning-based path loss prediction models. The main results obtained from this study show that the best root-mean-square error (RMSE) performance is given by the ANN and RNN-LSTM methods, while the worst is for the MLR method. All the RMSE values for the given learning techniques are in the range of $0.0216$ to $2.9008$ dB. Furthermore, this work shows that the models (except for the MLR model) perform excellently in fitting actual measurement data for wireless communications in enclosed indoor environments since they provide R-squared and correlation values higher than $0.91$ and $0.96$, respectively. The paper shows that these learning methods could be used as accurate and stable models for predicting path loss in the mmWave frequency regime. Full article

Power system configuration and performance are changing very quickly. Under the new paradigm of prosumers and energy communities, grids are increasingly influenced by microgeneration systems connected in both low and medium voltage. In addition, these facilities provide little or no information to distribution and/or transmission system operators, increasing power system management problems. Actually, information is a great asset to manage this new situation. The arrival of affordable and open Internet of Things (IoT) technologies is a remarkable opportunity to overcome these inconveniences allowing for the exchange of information about these plants. In this paper, we propose a monitoring solution applicable to photovoltaic self-consumption or any other microgeneration installation, covering the installations of the so-called ’prosumers’ and aiming to provide a tool for local self-consumption monitoring. A detailed description of the proposed system at the hardware level is provided, and extended information on the communication characteristics and data packets is also included. Results of different field test campaigns carried out in real PV self-consumption installations connected to the grid are described and analyzed. It can be affirmed that the proposed solution provides outstanding results in reliability and accuracy, being a popular solution for those who cannot afford professional monitoring platforms. Full article

Alteration in Land Use/Cover (LULC) considered a major challenge over the recent decades, as it plays an important role in diminishing biodiversity, altering the macro and microclimate. Therefore, the current study was designed to examine the past 30 years (1987–2017) changes in LULC and Land Surface Temperature (LST) and also simulated for next 30 years (2047). The LULC maps were developed based on maximum probability classification while the LST was retrieved from Landsat thermal bands and Radiative Transfer Equation (RTE) method for the respective years. Different approaches were used, such as Weighted Evidence (WE), Cellular Automata (CA) and regression prediction model for the year 2047. Resultantly, the LULC classification showed increasing trend in built-up and bare soil classes (13 km² and 89 km²), and the decreasing trend in vegetation class (−144 km²) in the study area. In the next 30 years, the built-up and bare soil classes would further rise with same speed (25 km² and 36.53 km²), and the vegetation class would further decline (−147 km²) until 2047. Similarly for LST, the temperature range for higher classes (27 -< 30 °C) increased by about 140 km² during 1987–2017, which would further enlarge (409 km²) until 2047. The lower LST range (15 °C to <21 °C) showed a decreasing trend (−54.94 km²) and would further decline to (−20 km²) until 2047 if it remained at the same speed. Prospective findings will be helpful for land use planners, climatologists and other scientists in reducing the increasing LST associated with LULC changes. Full article

The development of satellite sensors and interferometry synthetic aperture radar (InSAR) technology has enabled the exploitation of their benefits for long-term structural health monitoring (SHM). However, some restrictions cause this process to provide a small number of images leading to the problem of small data for SAR-based SHM. Conversely, the major challenge of the long-term monitoring of civil structures pertains to variations in their inherent properties by environmental and/or operational variability. This article aims to propose new hybrid unsupervised learning methods for addressing these challenges. The methods in this work contain three main parts: (i) data augmentation by the Markov Chain Monte Carlo algorithm, (ii) feature normalization, and (iii) decision making via Mahalanobis-squared distance. The first method presented in this work develops an artificial neural network-based feature normalization by proposing an iterative hyperparameter selection of hidden neurons of the network. The second method is a novel unsupervised teacher–student learning by combining an undercomplete deep neural network and an overcomplete single-layer neural network. A small set of long-term displacement samples extracted from a few SAR images of TerraSAR-X is applied to validate the proposed methods. The results show that the methods can effectively deal with the major challenges in the SAR-based SHM applications. Full article

Skin cancer (melanoma and non-melanoma) is one of the most common cancer types and leads to hundreds of thousands of yearly deaths worldwide. It manifests itself through abnormal growth of skin cells. Early diagnosis drastically increases the chances of recovery. Moreover, it may render surgical, radiographic, or chemical therapies unnecessary or lessen their overall usage. Thus, healthcare costs can be reduced. The process of diagnosing skin cancer starts with dermoscopy, which inspects the general shape, size, and color characteristics of skin lesions, and suspected lesions undergo further sampling and lab tests for confirmation. Image-based diagnosis has undergone great advances recently due to the rise of deep learning artificial intelligence. The work in this paper examines the applicability of raw deep transfer learning in classifying images of skin lesions into seven possible categories. Using the HAM1000 dataset of dermoscopy images, a system that accepts these images as input without explicit feature extraction or preprocessing was developed using 13 deep transfer learning models. Extensive evaluation revealed the advantages and shortcomings of such a method. Although some cancer types were correctly classified with high accuracy, the imbalance of the dataset, the small number of images in some categories, and the large number of classes reduced the best overall accuracy to 82.9%. Full article

The paper deals with a lead-through method of programming for industrial robots. The goal is to automatically reproduce 6DoF trajectories of a tool wielded by a human operator demonstrating a motion task. We present a novel motion-tracking system built around the HTC Vive pose estimation system. Our solution allows complete automation of the robot teaching process. Specific algorithmic issues of system calibration and motion data post-processing are also discussed, constituting the paper’s theoretical contribution. The motion tracking system is successfully deployed in a pilot application of robot-assisted spray painting. Full article

Aiming at non-stationary signals with complex components, the performance of a variational mode decomposition (VMD) algorithm is seriously affected by the key parameters such as the number of modes $K$, the quadratic penalty parameter $\alpha$ and the update step $\tau$. In order to solve this problem, an adaptive empirical variational mode decomposition (EVMD) method based on a binary tree model is proposed in this paper, which can not only effectively solve the problem of VMD parameter selection, but also effectively reduce the computational complexity of searching the optimal VMD parameters using intelligent optimization algorithm. Firstly, the signal noise ratio (SNR) and refined composite multi-scale dispersion entropy (RCMDE) of the decomposed signal are calculated. The RCMDE is used as the setting basis of the $\alpha$, and the SNR is used as the parameter value of the $\tau$. Then, the signal is decomposed into two components based on the binary tree mode. Before decomposing, the $\alpha$ and $\tau$ need to be reset according to the SNR and MDE of the new signal. Finally, the cycle iteration termination condition composed of the least squares mutual information and reconstruction error of the components determines whether to continue the decomposition. The components with large least squares mutual information (LSMI) are combined, and the LSMI threshold is set as 0.8. The simulation and experimental results indicate that the proposed empirical VMD algorithm can decompose the non-stationary signals adaptively, with lower complexity, which is O(n²), good decomposition effect and strong robustness. Full article

Health monitoring and related technologies are a rapidly growing area of research. To date, the electrocardiogram (ECG) remains a popular measurement tool in the evaluation and diagnosis of heart disease. The number of solutions involving ECG signal monitoring systems is growing exponentially in the literature. In this article, underestimated Orthogonal Matching Pursuit (OMP) algorithms are used, demonstrating the significant effect of concise representation parameters on improving the performance of the classification process. Cardiovascular disease classification models based on classical Machine Learning classifiers were defined and investigated. The study was undertaken on the recently published PTB-XL database, whose ECG signals were previously subjected to detailed analysis. The classification was realized for class 2, class 5, and class 15 cardiac diseases. A new method of detecting R-waves and, based on them, determining the location of QRS complexes was presented. Novel aggregation methods of ECG signal fragments containing QRS segments, necessary for tests for classical classifiers, were developed. As a result, it was proved that ECG signal subjected to algorithms of R wave detection, QRS complexes extraction, and resampling performs very well in classification using Decision Trees. The reason can be found in structuring the signal due to the actions mentioned above. The implementation of classification issues achieved the highest Accuracy of 90.4% in recognition of 2 classes, as compared to less than 78% for 5 classes and 71% for 15 classes. Full article

Due to the limited computing capacity of onboard devices, they can no longer meet a large number of computing requirements. Therefore, mobile edge computing (MEC) provides more computing and storage capabilities for vehicles. Inspired by a large number of roadside parking vehicles, this paper takes the roadside parking vehicles with idle computing resources as the task offloading platform and proposes a mobile edge computing task offloading strategy based on roadside parking cooperation. The resource sharing and mutual utilization among roadside vehicles, roadside units (RSU), and cloud servers (cloud servers) were established, and the collaborative offloading problem of computing tasks was transformed into a constraint problem. The hybrid genetic algorithm (HHGA) with a mountain-climbing operator was used to solve the multi-constraint problem, to reduce the delay and energy consumption of computing tasks. The simulation results show that when the number of tasks is 25, the delay and energy consumption of the HHGA algorithm is improved by 24.1% and 11.9%, respectively, compared with Tradition. When the task size is 1.0 MB, the HHGA algorithm reduces the system overhead by 7.9% compared with Tradition. Therefore, the proposed scheme can effectively reduce the total system cost during task offloading. Full article

In recent investigations of magnetoelectric sensors based on microelectromechanical cantilevers made of TiN/AlN/Ni, a complex eigenfrequency behavior arising from the anisotropic ΔE effect was demonstrated. Within this work, a FEM simulation model based on this material system is presented to allow an investigation of the vibrational properties of cantilever-based sensors derived from magnetocrystalline anisotropy while avoiding other anisotropic contributions. Using the magnetocrystalline ΔE effect, a magnetic hardening of Nickel is demonstrated for the (110) as well as the (111) orientation. The sensitivity is extracted from the field-dependent eigenfrequency curves. It is found, that the transitions of the individual magnetic domain states in the magnetization process are the dominant influencing factor on the sensitivity for all crystal orientations. It is shown, that Nickel layers in the sensor aligned along the medium or hard axis yield a higher sensitivity than layers along the easy axis. The peak sensitivity was determined to 41.3 T⁻¹ for (110) in-plane-oriented Nickel at a magnetic bias flux of 1.78 mT. The results achieved by FEM simulations are compared to the results calculated by the Euler–Bernoulli theory. Full article