Search

Kumite is a karate sparring competition in which two players face off and perform offensive and defensive techniques. Depending on the players, there may be preliminary actions (hereinafter referred to as "pre-actions"), such as pulling the arms or legs, lowering the shoulders, etc., just before a technique is performed. Since the presence of a pre-action allows the opponent to know the timing of the technique, it is important to reduce pre-actions in order to improve the kumite. However, it is difficult for beginners and intermediate players to accurately identify their pre-actions and to improve them through practice. Therefore, this study aims to construct a practice support system that enables beginners and intermediate players to understand their pre-actions. In this paper, we focus on the forefist punch, one of kumite's punching techniques. We propose a method to estimate the presence or absence of a pre-action based on the similarity between the acceleration data of an arbitrary forefist punch and a previously prepared dataset consisting of acceleration data of the forefist punch without a pre-action. We found that the proposed method can estimate the presence or absence of a pre-action in an arbitrary forefist punch with an accuracy of 86%. We also developed KARATECH as a system to support the practice of reducing pre-actions using the proposed method. KARATECH shows the presence or absence of pre-actions through videos and graphs. The evaluation results confirmed that the group using KARATECH had a lower pre-action rate. [ABSTRACT FROM AUTHOR]

Estimation of temporospatial clinical features of gait (CFs), such as step count and length, step duration, step frequency, gait speed, and distance traveled, is an important component of community-based mobility evaluation using wearable accelerometers. However, accurate unsupervised computerized measurement of CFs of individuals with Duchenne muscular dystrophy (DMD) who have progressive loss of ambulatory mobility is difficult due to differences in patterns and magnitudes of acceleration across their range of attainable gait velocities. This paper proposes a novel calibration method. It aims to detect steps, estimate stride lengths, and determine travel distance. The approach involves a combination of clinical observation, machine-learning-based step detection, and regression-based stride length prediction. The method demonstrates high accuracy in children with DMD and typically developing controls (TDs) regardless of the participant's level of ability. Fifteen children with DMD and fifteen TDs underwent supervised clinical testing across a range of gait speeds using 10 m or 25 m run/walk (10 MRW, 25 MRW), 100 m run/walk (100 MRW), 6-min walk (6 MWT), and free-walk (FW) evaluations while wearing a mobile-phone-based accelerometer at the waist near the body's center of mass. Following calibration by a trained clinical evaluator, CFs were extracted from the accelerometer data using a multi-step machine-learning-based process and the results were compared to ground-truth observation data. Model predictions vs. observed values for step counts, distance traveled, and step length showed a strong correlation (Pearson's r = −0.9929 to 0.9986, p < 0.0001). The estimates demonstrated a mean (SD) percentage error of 1.49% (7.04%) for step counts, 1.18% (9.91%) for distance traveled, and 0.37% (7.52%) for step length compared to ground-truth observations for the combined 6 MWT, 100 MRW, and FW tasks. Our study findings indicate that a single waist-worn accelerometer calibrated to an individual's stride characteristics using our methods accurately measures CFs and estimates travel distances across a common range of gait speeds in both DMD and TD peers. [ABSTRACT FROM AUTHOR]

The escalating global water usage and the increasing strain on major cities due to water shortages highlights the critical need for efficient water management practices. In water-stressed regions worldwide, significant water wastage is primarily attributed to leakages, inefficient use, and aging infrastructure. Undetected water leakages in buildings' pipelines contribute to the water waste problem. To address this issue, an effective water leak detection method is required. In this paper, we explore the application of edge computing in smart buildings to enhance water management. By integrating sensors and embedded Machine Learning models, known as TinyML, smart water management systems can collect real-time data, analyze it, and make accurate decisions for efficient water utilization. The transition to TinyML enables faster and more cost-effective local decision-making, reducing the dependence on centralized entities. In this work, we propose a solution that can be adapted for effective leakage detection in real-world scenarios with minimum human intervention using TinyML. We follow an approach that is similar to a typical machine learning lifecycle in production, spanning stages including data collection, training, hyperparameter tuning, offline evaluation and model optimization for on-device resource efficiency before deployment. In this work, we considered an existing water leakage acoustic dataset for polyvinyl chloride pipelines. To prepare the acoustic data for analysis, we performed preprocessing to transform it into scalograms. We devised a water leak detection method by applying transfer learning to five distinct Convolutional Neural Network (CNN) variants, which are namely EfficientNet, ResNet, AlexNet, MobileNet V1, and MobileNet V2. The CNN models were found to be able to detect leakages where a maximum testing accuracy, recall, precision, and F1 score of 97.45%, 98.57%, 96.70%, and 97.63%, respectively, were observed using the EfficientNet model. To enable seamless deployment on the Arduino Nano 33 BLE edge device, the EfficientNet model is compressed using quantization resulting in a low inference time of 1932 ms, a peak RAM usage of 255.3 kilobytes, and a flash usage requirement of merely 48.7 kilobytes. [ABSTRACT FROM AUTHOR]

The study of Domain Generalization (DG) has gained considerable momentum in the Machine Learning (ML) field. Human Activity Recognition (HAR) inherently encompasses diverse domains (e.g., users, devices, or datasets), rendering it an ideal testbed for exploring Domain Generalization. Building upon recent work, this paper investigates the application of regularization methods to bridge the generalization gap between traditional models based on handcrafted features and deep neural networks. We apply various regularizers, including sparse training, Mixup, Distributionally Robust Optimization (DRO), and Sharpness-Aware Minimization (SAM), to deep learning models and assess their performance in Out-of-Distribution (OOD) settings across multiple domains using homogenized public datasets. Our results show that Mixup and SAM are the best-performing regularizers. However, they are unable to match the performance of models based on handcrafted features. This suggests that while regularization techniques can improve OOD robustness to some extent, handcrafted features remain superior for domain generalization in HAR tasks. [ABSTRACT FROM AUTHOR]

Injection moulding (IM) is an important industrial process, known to be the most used plastic formation technique. Demand for faster cycle times and higher product customisation is driving interest in additive manufacturing (AM) as a new method for mould tool manufacturing. The use of AM offers advantages such as greater design flexibility and conformal cooling of components to reduce cycle times and increase product precision. However, shortcomings of metal additive manufacturing, such as porosity and residual stresses, introduce uncertainties about the reliability and longevity of AM tooling. The injection moulding process relies on high volumes of produced parts and a minimal amount of tool failures. This paper reviews the demands for tool condition monitoring systems for AM-manufactured mould tools; although tool failures in conventionally manufactured tooling are rare, they do occur, usually due to cracking, deflection, and channel blockages. However, due to the limitations of the AM process, metal 3D-printed mould tools are susceptible to failures due to cracking, delamination and deformation. Due to their success in other fields, acoustic emission, accelerometers and ultrasound sensors offer the greatest potential in mould tool condition monitoring. Due to the noisy machine environment, sophisticated signal processing and decision-making algorithms are required to prevent false alarms or the missing of warning signals. This review outlines the state of the art in signal decomposition and both data- and model-based approaches to determination of the current state of the tool, and how these can be employed for IM tool condition monitoring. The development of such a system would help to ensure greater industrial uptake of additive manufacturing of injection mould tooling, by increasing confidence in the technology, further improving the efficiency and productivity of the sector. [ABSTRACT FROM AUTHOR]

Activity monitoring has become a necessary demand for weak people to guarantee their safety. The paper proposed a Parallel Training Logical Execution (PTLE) system using machine learning (ML) models on a microelectromechanical system (MEMS) accelerometer to detect coughs, falls, and other normal activities. When there are many categories, the ML prediction can be confused between these activities with each other. The PTLE system trains several models in parallel with more specific activity classes in each dataset. The shared tasks between parallel models relieve the complexity for a single one. There are six additional parameters for accelerometer characteristics, which were calculated from three axes accelerations as input features to improve the ML's consciousness. Once all models were trained, the system was ready to receive the input accelerations and activated the logical flow to manage link operation between these ML models for output predictions. Random Forest (RF) had the highest potential among the ML classification algorithms after the validation. In the experiment, the comparison between the PTLE model and the regular ML model were carried out with real-time data from an M5stickC wearable device on the user's chest to the trained models on PC. The result showed the advancement of the proposed method in term of precision, recall, F1-score with an overall accuracy of 98% in the real-time test. The accelerations from the wearable device were sent to ML models via Wi-Fi with Message Queue Telemetry Transport (MQTT) broker, and the activity predictions were transferred to the cloud for the family members or doctor care based on Internet of Things (IoT) communication. [ABSTRACT FROM AUTHOR]

In this paper, a dual-axis Fabry–Pérot (FP) accelerometer assembled on single optical fiber is proposed. The sensor is equipped with a special beam-splitting prism to split the light into two perpendicular directions (the X- and Y-axes); the prism surface coated with semi-permeable film and the reflective sheet on the corresponding Be-Cu vibration-sensitive spring form two sets of FP cavities of different sizes. When the Be-Cu spring with a proof mass (PM) is subjected to the vibration signal, the cavity length of the corresponding FP cavity is changed and the interference signal returns to the collimator through the original path of the prism. After bandpass filtering and demodulation, the two cavity lengths are obtained, and the acceleration measurement in dual-axis directions is completed. The resonant frequency of the proposed dual-axis fiber optic accelerometer is around 280 Hz. The results of the spectral measurements show 3.93 μ m / g (g = 9.8 m/s2: gravity constant) and 4.19 μ m / g for the applied acceleration along the X- and Y-axes, respectively, and the cross-axis sensitivity is below 5.1%. Within the angle range of 180°, the maximum error of measured acceleration is less than 3.77%. The proposed fiber optic dual-axis FP accelerometer has high sensitivity and strong immunity to electromagnetic interference. The size of the sensor mainly depends on the size of the prism, which is easy to reduce and mass produce. Moreover, this FP construction method has high flexibility and development potential. [ABSTRACT FROM AUTHOR]

Improper cycling posture is linked to a variety of spinal musculoskeletal diseases, including structural malformation of the spine and back discomfort. This paper presents a novel smart shirt integrated tri-axial gyroscope and accelerometer that can detect postural variation in terms of spinal curvature changes. To provide accurate feedback to the wearer and improve the wearer's correct movement, the garment is able to recognize trunk body posture. The gyroscope/accelerometer was placed around the upper and mid trunk of the user to record tri-axial angular velocity data. The device can also be used to help determine the trunk bending angle and monitor body postures in order to improve optimal orientation and position. The garment enables continuous measurement in the field at high sample rates (50 Hz), and the sensor has a large measurement range (16 g, 2000°/s). As electronic components are non-washable, instead of encapsulating them, a detachable module was created. In this, magnets are embedded in the jersey, and allow the positioning and removal of the sensor. The test results show that the average trunk-bending angle was 21.5°, and 99 percent of the observed angle fell within the standard (ranging from 8° to 35°). The findings demonstrate the feasibility of employing the smart shirt sensor to estimate trunk motions in the field on a regular basis. [ABSTRACT FROM AUTHOR]

Human Activity Recognition (HAR) has been studied extensively, yet current approaches are not capable of generalizing across different domains (i.e., subjects, devices, or datasets) with acceptable performance. This lack of generalization hinders the applicability of these models in real-world environments. As deep neural networks are becoming increasingly popular in recent work, there is a need for an explicit comparison between handcrafted and deep representations in Out-of-Distribution (OOD) settings. This paper compares both approaches in multiple domains using homogenized public datasets. First, we compare several metrics to validate three different OOD settings. In our main experiments, we then verify that even though deep learning initially outperforms models with handcrafted features, the situation is reversed as the distance from the training distribution increases. These findings support the hypothesis that handcrafted features may generalize better across specific domains. [ABSTRACT FROM AUTHOR]

In safety-critical systems such as industrial plants or aircraft, failure occurs inevitably during operation, and it is important to prevent it in order to maintain high availability. To reduce this risk, a lot of efforts are directed from developing sensing technologies to failure prognosis algorithms to enable predictive maintenance. The success of effective and reliable predictive maintenance not only relies on robust prognosis algorithms but also on the selection of sensors or data acquisition strategy. However, there are not many in-depth studies on a trade-off between sensor quality and data storage in the view of prognosis performance. The information about (1) how often data should be measured and (2) how good sensor quality should be for reliable failure prediction can be highly impactful for practitioners. In this paper, the authors evaluate the efficacy of the two factors in terms of remaining useful life (RUL) prediction accuracy and its uncertainty. In addition, since knowing true degradation information is almost impossible in practice, the authors validated the use of the prognosis metric without requiring the true degradation information. A numerical case study is conducted to identify the relationship between sensor quality and data storage. Then, real bearing run-to-failure (RTF) datasets acquired from accelerometer (contact type) and microphone (non-contact type) sensors are evaluated based on the prognosis performance metric and compared in terms of the sensors' cost-effectiveness for predictive maintenance. [ABSTRACT FROM AUTHOR]

Assistive Technology helps to assess the daily living and safety of frail people, with particular regards to the detection and prevention of falls. In this paper, a comparison is provided among different strategies to analyze postural sway, with the aim of detecting unstable postural status in standing condition as precursors of potential falls. Three approaches are considered: (i) a time-based features threshold algorithm, (ii) a time-based features Neuro-Fuzzy inference system, and (iii) a Neuro-Fuzzy inference fed by Discrete-Wavelet-Transform-based features. The analysis was performed across a wide dataset and exploited performance indexes aimed at assessing the accuracy and the reliability of predictions provided by the above-mentioned strategies. The results obtained demonstrate valuable performances of the three considered strategies in correctly distinguishing among stable and unstable postural status. However, the analysis of robustness against noisy data highlights better performance of Neuro-Fuzzy inference systems with respect to the threshold-based algorithm. [ABSTRACT FROM AUTHOR]

Both as an aid for less experienced clinicians and to enhance objectivity and sharp clinical skills in professionals, quantitative technologies currently bring the equine lameness diagnostic closer to evidence-based veterinary medicine. The present paper describes an original, inertial sensor-based wireless device system, the Lameness Detector 0.1, used in ten horses with different lameness degrees in one fore- or hind-leg. By recording the impulses on three axes of the incorporated accelerometer in each leg of the assessed horse, and then processing the data using custom-designed software, the device proved its usefulness in lameness identification and severity scoring. Mean impulse values on the horizontal axis calculated for five consecutive steps above 85, regardless of the leg, indicated the slightest subjectively recognizable lameness, increasing to 130 in severe gait impairment. The range recorded on the same axis (between 61.2 and 67.4) in the sound legs allowed a safe cut-off value of 80 impulses for diagnosing a painful limb. The significance of various comparisons and several correlations highlighted the potential of this simple, affordable, and easy-to-use lameness detector device for further standardization as an aid for veterinarians in diagnosing lameness in horses. [ABSTRACT FROM AUTHOR]

This paper presents a tidal current meter that is based on the inertial acceleration principle for offshore infrastructures in deep water. Focusing on the marine installations of the aquaculture industry, we studied the forces of tides at a depth of 15 m by measuring the acceleration. In addition, we used a commercial MEMS triaxial accelerometer to record the acceleration values. A prototype of the tidal measurement unit was developed and tested at a real offshore aquaculture infrastructure in Gran Canaria, which is one of the Canary Islands in the Atlantic Ocean. The proposed tidal measurement unit was used as a recorder to assess the complexity of measuring the frequency of tidal currents in the short (10 min), medium (one day) and long term (one week). The acquired data were studied in detail, in both the time and frequency domains, to determine the frequency of the forces that were involved. Finally, the complexity of the frequency measurements from the captured data was analyzed in terms of sampling ratio and recording duration, from the point of view of using our proposed measurement unit as an ultra-low-power embedded system. The proposed device was tested for more than 180 days using a lithium-ion battery. This working period was three times greater than the best alternative in the literature because of the ultra-low-power design of the on-board embedded system. The measurement accuracy error was lower than 1% and the resolution was 0.01 cm/s for the 0.8 m/s velocity scale. This performance was similar to the best Doppler solution that was found in the literature. [ABSTRACT FROM AUTHOR]

As mobile devices become more and more popular, users gain many conveniences. It has also made smartphone makers install new software and prebuilt hardware on their products, including many kinds of sensors. With improved storage and computing power, users also become accustomed to storing and interacting with personally sensitive information. Due to convenience and efficiency, mobile devices use gait authentication widely. In recent years, protecting the information security of mobile devices has become increasingly important. It has become a hot research area because smartphones are vulnerable to theft or unauthorized access. This paper proposes a novel attack model called a collusion attack. Firstly, we study the imitation attack in the general state and its results and propose and verify the feasibility of our attack. We propose a collusion attack model and train participants with quantified action specifications. The results demonstrate that our attack increases the attacker's false match rate only using an acceleration sensor in some systems sensor. Furthermore, we propose a multi-cycle defense model based on acceleration direction changes to improve the robustness of smartphone-based gait authentication methods against such attacks. Experimental results show that our defense model can significantly reduce the attacker's success rate. [ABSTRACT FROM AUTHOR]

Physical activity (PA) is globally recognized as a pillar of general health. Step count, as one measure of PA, is a well known predictor of long-term morbidity and mortality. Despite its popularity in consumer devices, a lack of methodological standards and clinical validation remains a major impediment to step count being accepted as a valid clinical endpoint. Previous works have mainly focused on device-specific step-count algorithms and often employ sensor modalities that may not be widely available. This may limit step-count suitability in clinical scenarios. In this paper, we trained neural network models on publicly available data and tested on an independent cohort using two approaches: generalization and personalization. Specifically, we trained neural networks on accelerometer signals from one device and either directly applied them or adapted them individually to accelerometer data obtained from a separate subject cohort wearing multiple distinct devices. The best models exhibited highly accurate step-count estimates for both the generalization (96–99%) and personalization (98–99%) approaches. The results demonstrate that it is possible to develop device-agnostic, accelerometer-only algorithms that provide highly accurate step counts, positioning step count as a reliable mobility endpoint and a strong candidate for clinical validation. [ABSTRACT FROM AUTHOR]

Coronavirus 2019 (COVID-19) has posed a serious threat to the lives and health of the majority of people worldwide. Since the early days of the outbreak, South Korea's government and citizens have made persistent efforts to provide effective prevention against further spread of the disease. In particular, the participation of individual citizens in complying with the necessary code of conduct to prevent spread of the infection, through measures such as social distancing and mask wearing, is as instrumental as the geographical tracking of the trajectory of the infected. In this paper, we propose an activity recognition method based on a wristband equipped with an IR array and inertial measurement unit (IMU) to detect individual compliance with codes of personal hygiene management, such as mask wearing, which are recommended to prevent the spread of infectious diseases. The results of activity recognition were comparatively analyzed by applying conventional machine learning algorithms and convolutional neural networks (CNNs) to the IMU time series and IR array thermal images collected from 25 subjects. When CNN and 24 × 32 thermal images were used, 97.8% accuracy was achieved (best performance), and when 6 × 8 low-resolution thermal images were used, similar performance with 97.1% accuracy was obtained. In the case of using IMU, the performance of activity recognition was lower than that obtained with the IR array, but an accuracy of 93% was achieved even in the case of applying machine learning algorithms, indicating that it is more suitable for wearable devices with low computational capability. [ABSTRACT FROM AUTHOR]

Drowning is considered amongst the top 10 causes of unintentional death, according to the World Health Organization (WHO). Therefore, anti-drowning systems that can save lives by preventing and detecting drowning are much needed. This paper proposes a robust and waterproof sensor-based device to detect distress in swimmers at varying depths and different types of water environments. The proposed device comprises four main components, including heart rate, blood oxygen level, movement, and depth sensors. Although these sensors were designed to work together to boost the system's capability as an anti-drowning device, each could operate independently. The sensors were able to determine the heart rate to an accuracy of 1 beat per minute (BPM), 1% SpO2, the acceleration with adjustable sensitivities of ±2 g, ±4 g, ±8 g, and ±16 g, and the depth up to 12.8 m. The data obtained from the sensors were sent to a microcontroller that compared the input data to adjustable threshold values to detect dangerous situations. Being in hazardous situations for more than a specific time activated the alarming system. Based on the comparison made in the program and measuring the time of submersion, a message indicating drowning or safe was sent to a lifeguard to continuously monitor the swimmer' condition via Wi-Fi to an IP address reachable by a mobile phone or laptop. It is also possible to continuously monitor the sensor outputs on the device's display or the connected mobile phone or laptop. The threshold values could be adjusted based on biometric parameters such as swimming conditions (swimming pool, beach, depth, etc.) and swimmers health and conditions. The functionality of the proposed device was thoroughly tested over a wide range of parameters and under different conditions, both in air and underwater. It was demonstrated that the device could detect a range of potentially hazardous aquatic situations. This work will pave the way for developing an effective drowning sensing system that could save tens of thousands of lives across the globe every year. [ABSTRACT FROM AUTHOR]

Discrete particle dynamics is one of the least understood aspects of river bedload transport, but in situ measurement of stone movement during floods poses a significant technical challenge. A promising approach to address this knowledge gap is to use sensors embedded within stones. Sensors must be waterproof and recoverable after being transported downstream and potentially buried by other sediment. To address this challenge rugged sensors (Kinematic Loggers) were developed for deployment inside stones (ranging in size from cobbles to boulders) during floods. The sensors feature a 9-axis inertial measurement unit, 3-axis high-g accelerometer, 128 MB flash memory, and a 433 MHz LoRa radio transmission module for sensor recovery. The sensors are enclosed in rugged waterproof housings for deployment in extreme conditions (i.e., bedload transport during floods). Novel relay units and drone-based recovery systems were also developed for finding the sensors after field deployments. Firmware to control the sensors and relay units was developed, as well as software for configuring the sensors and an android application for communicating with the sensors via the LoRa radio transmission module. This paper covers the technical development of the sensors, mounting them inside stones, and field recovery tests. Although designed for measurement of coarse bedload transport and particle dynamics during floods, the sensors are equally applicable for deployment in other harsh environments, such as to study landslide and rockfall dynamics. [ABSTRACT FROM AUTHOR]

This paper presents methods for continuous condition monitoring of railway switches and crossings (S&C, turnout) via sleeper-mounted accelerometers at the crossing transition. The methods are developed from concurrently measured sleeper accelerations and scanned crossing geometries from six in situ crossing panels. These measurements combined with a multi-body simulation (MBS) model with a structural track model and implemented scanned crossing geometries are used to derive the link between the crossing geometry condition and the resulting track excitation. From this analysis, a crossing condition indicator C λ 1 − λ 2 ,   γ is proposed. The indicator is defined as the root mean square (RMS) of a track response signal γ that has been band-passed between frequencies corresponding to track deformation wavelength bounds of λ 1 and λ 2 for the vehicle passing speed (f = v/   λ ). In this way, the indicator ignores the quasi-static track response with wavelengths predominantly above λ 1 and targets the dynamic track response caused by the kinematic wheel-crossing interaction governed by the crossing geometry. For the studied crossing panels, the indicator C 1 − 0.2   m ,   γ ( λ 1 = 1 and λ 2 = 0.2 ) was evaluated for γ = u, v, or a as in displacements, velocities, and accelerations, respectively. It is shown that this condition indicator has a strong correlation with vertical wheel–rail contact forces that is sustained for various track conditions. Further, model calibrations were performed to measured sleeper displacements for the six investigated crossing panels. The calibrated models show (1) a good agreement between measured and simulated sleeper displacements for the lower frequency quasi-static track response and (2) improved agreement for the dynamic track response at higher frequencies. The calibration also improved the agreement between measurements and simulation for the crossing condition indicator demonstrating the value of model calibration for condition monitoring purposes. [ABSTRACT FROM AUTHOR]

The role and importance of geodesists in the planning and building of civil engineering constructions are well known. However, the importance and benefits of collected data during maintenance in exploitation have arisen in the last thirty years due primarily to the development of Global Positioning Systems (GPS) and Global Navigation Satellite System (GNSS) instruments, sensors and systems, which can receive signals from multiple GPS systems. In the last fifteen years, the development of Terrestrial Laser Scanners (TLS) and Image-Assisted Total Stations (IATS) has enabled much wider integration of these types of geodetic instruments with their sensors into monitoring systems for the displacement and deformation monitoring of structures, as well as for regular structure inspections. While GNSS sensors have certain limitations regarding their accuracy, their suitability in monitoring systems, and the need for a clean horizon, IATS do not have these limitations. The latest development of Total Stations (TS) called IATS is a theodolite that consists of a Robotic Total Station (RTS) with integrated image sensors. Today, IATS can be used for structural and geo-monitoring, i.e., for the determination of static and dynamic displacements and deformations, as well as for the determination of civil engineering structures' natural frequencies. In this way, IATS can provide essential information about the current condition of structures. However, like all instruments and sensors, they have their advantages and disadvantages. IATS's biggest advantage is their high level of accuracy and precision and the fact that they do not need to be set up on the structure, while their biggest disadvantage is that they are expensive. In this paper, the developed low-cost IATS prototype, which consists of an RTS Leica TPS1201 instrument and GoPro Hero5 camera, is presented. At first, the IATS prototype was tested in the laboratory where simulated dynamic displacements were determined. After the experiment, the IATS prototype was used in the field for the purpose of static and dynamic load testing of the railway bridge Kloštar, after its reconstruction according to HRN ISO NORM U.M1.046—Testing of bridges by load test. In this article, the determination of bridge dynamic displacements and results of the computation of natural frequencies using FFT from the measurement data obtained by means of IATS are presented. During the load testing of the bridge, the frequencies were also determined by accelerometers, and these data were used as a reference for the assessment of IATS accuracy and suitability for dynamic testing. From the conducted measurements, we successfully determined natural bridge frequencies as they match the results gained by accelerometers. [ABSTRACT FROM AUTHOR]

Existing accelerometer-based human activity recognition (HAR) benchmark datasets that were recorded during free living suffer from non-fixed sensor placement, the usage of only one sensor, and unreliable annotations. We make two contributions in this work. First, we present the publicly available Human Activity Recognition Trondheim dataset (HARTH). Twenty-two participants were recorded for 90 to 120 min during their regular working hours using two three-axial accelerometers, attached to the thigh and lower back, and a chest-mounted camera. Experts annotated the data independently using the camera's video signal and achieved high inter-rater agreement (Fleiss' Kappa = 0.96 ). They labeled twelve activities. The second contribution of this paper is the training of seven different baseline machine learning models for HAR on our dataset. We used a support vector machine, k-nearest neighbor, random forest, extreme gradient boost, convolutional neural network, bidirectional long short-term memory, and convolutional neural network with multi-resolution blocks. The support vector machine achieved the best results with an F1-score of 0.81 (standard deviation: ± 0.18 ), recall of 0.85 ± 0.13 , and precision of 0.79 ± 0.22 in a leave-one-subject-out cross-validation. Our highly professional recordings and annotations provide a promising benchmark dataset for researchers to develop innovative machine learning approaches for precise HAR in free living. [ABSTRACT FROM AUTHOR]

Mediumfrequency fiber Bragg grating (FBG) acceleration sensors are used in important applications in mechanical, aerospace and weapon equipment, and have strict requirements in terms of resonance frequency and sensitivity. A novel medium-frequency accelerometer, based on fiber Bragg grating and flexible hinges, is proposed in this paper. The differential structure doubles the sensitivity of the sensor while avoiding temperature effects. The structure model and principle for the sensor are introduced, the sensor's sensing characteristics are theoretically analyzed, and the structure parameters for the sensor are determined through numerical analysis. The sensing experiments show that the resonance frequency of the sensor is approximately 2800 Hz, the sensitivity is 21.8 pm/g in the flat frequency range of 50–1000 Hz, and the proposed sensor has a good temperature self-compensation function and lateral anti-interference capability. [ABSTRACT FROM AUTHOR]

At present, people spend most of their time in passive rather than active mode. Sitting with computers for a long time may lead to unhealthy conditions like shoulder pain, numbness, headache, etc. To overcome this problem, human posture should be changed for particular intervals of time. This paper deals with using an inertial sensor built in the smartphone and can be used to overcome the unhealthy human sitting behaviors (HSBs) of the office worker. To monitor, six volunteers are considered within the age band of 26 ± 3 years, out of which four were male and two were female. Here, the inertial sensor is attached to the rear upper trunk of the body, and a dataset is generated for five different activities performed by the subjects while sitting in the chair in the office. Correlation-based feature selection (CFS) technique and particle swarm optimization (PSO) methods are jointly used to select feature vectors. The optimized features are fed to machine learning supervised classifiers such as naive Bayes, SVM, and KNN for recognition. Finally, the SVM classifier achieved 99.90% overall accuracy for different human sitting behaviors using an accelerometer, gyroscope, and magnetometer sensors. [ABSTRACT FROM AUTHOR]

Nowadays, engineers are widely using accelerometers to record the vibration of structures for structural verification purposes. The main obstacle for using these data acquisition systems is their high cost, which limits its use to unique structures with a relatively high structural health monitoring budget. In this paper, a Cost Hyper-Efficient Arduino Product (CHEAP) has been developed to accurately measure structural accelerations. CHEAP is a system that is composed of five low-cost accelerometers that are connected to an Arduino microcontroller as their data acquisition system. Test results show that CHEAP not only has a significantly lower price (14 times cheaper in the worst-case scenario) compared with other systems used for comparison but also shows better accuracy on low frequencies for low acceleration amplitudes. Moreover, the final output results of Fast Fourier Transformation (FFT) assessments showed a better observable resolution for CHEAP than the studied control systems. [ABSTRACT FROM AUTHOR]

The paper deals with performance enhancement of low-cost, consumer-grade inertial sensors realized by means of Micro Electro-Mechanical Systems (MEMS) technology. Focusing their attention on the reduction of bias instability and random walk-driven drift of cost-effective MEMS accelerometers and gyroscopes, the authors hereinafter propose a suitable method, based on a redundant configuration and complemented with a proper measurement procedure, to improve the performance of low-cost, consumer-grade MEMS sensors. The performance of the method is assessed by means of an adequate prototype and compared with that assured by a commercial, expensive, tactical-grade MEMS inertial measurement unit, taken as reference. Obtained results highlight the promising reliability and efficacy of the method in estimating position, velocity, and attitude of vehicles; in particular, bias instability and random walk reduction greater than 25% is, in fact, experienced. Moreover, differences as low as 0.025 rad and 0.89 m are obtained when comparing position and attitude estimates provided by the prototype and those granted by the tactical-grade MEMS IMU. [ABSTRACT FROM AUTHOR]

This paper presents an output offset minimized capacitance-to-digital interface for a MEMS accelerometer. With a gain-enhanced voltage-controlled oscillator (VCO)-based quantization loop, the interface is able to output a digital signal with improved dynamic range. For optimizing the output offset caused by nonideal factors (e.g., the bond-wire drift), a nested digital chopping feedback loop is embedded in the VCO-based quantization loop. It enables the interface to minimize the output offset without digital filtering and digital-to-analog conversion. The proposed architecture is well suited for dynamic range and offset improvements with low cost. Fabricated with a 0.18 μm Global Foundry (GF) CMOS process, the interface offers a 78 dB dynamic range with 0.4% nonlinearity from a single 2 V supply. With the input referred offset up to 1.3 pF, the offset cancellation loop keeps the DC output offset within 40 mV. The power dissipation is 6.5 mW with a bandwidth of 4 kHz. [ABSTRACT FROM AUTHOR]