The study aimed at obtaining precise view of modification of heart rate variability (HRV) and respiratory rate with help of “smart clothes” (the Hexoskin Smart Shirt, Hexoskin Smart Sensors & AI, Montreal, QC, Canada) during 45 min session of ground-based microgravity modeled by “dry” immersion (DI). Eight healthy subjects aged from 19 to 21 years participated in the study. Hexoskin Smart Shirt provided a .wav sound file, and the ecg_peaks function of the neurokit2 library was used. The resulting data set of RR intervals on ECG was filtered with cutoff heart rate < 45 and >180 beat per min. HRV parameters were calculated within 5-min segments with help of the pyHRV toolbox. Time-domain (HR, SDNN) and frequency-domain (TP, HF, LF, VLF) HRV parameters, sample and approximate entropy were calculated. The dynamics of HRV parameters and respiratory rate during DI session acquired with “smart” clothes were consistent with those obtained with conventional ECG devices. Thus, the “smart clothes” technology appears as a reliable telemetric instrument to monitor cardiac and respiratory regulation with HRV parameters during the DI session. This technology enabled to continuously monitor HRV parameters under DI, which allowed to evidence that HRV parameters change quite early (within minutes) under DI conditions. This findigs can be appliyed to monitor body functions in extreme environment, inncluding space conditions.

Computer vision applications require substantial amounts of data for effective training and accurate inference in a variety of tasks. However, in many real-world scenarios, data insufficiency is a common issue, which can arise due to various factors such as the rarity of certain conditions, difficulties in data collection, or high costs associated with data acquisition. This insufficiency often leads to computational models with inadequate performance, particularly in terms of their generalization capabilities. Traditional data augmentation techniques are widely used to mitigate overfitting and improve model robustness by artificially increasing the diversity of the training data. However, the application of these techniques is not always feasible or desirable, especially in cases where the augmented data does not accurately represent the underlying distribution. In response to these challenges, this paper explores an alternative data augmentation approach specifically designed for classification tasks. The method leverages adversarial images generated using the Fast Gradient Sign Method (FGSM) with added noise to address sample imbalance and enhance classifier performance. The technique was validated on a dataset of images related to the classification of diseases in coffee plants caused by nutrient deficiencies in the soil. The experimental results demonstrated a significant improvement in model performance, highlighting the effectiveness of the proposed method as a viable alternative to traditional data augmentation techniques.

Gas-sensitive devices have great potential for use in a variety of applications, including environmental monitoring, biomedical devices, and the pharmaceutical industry. Zinc oxide is one of the most widely used gas-sensitive materials due to its wide-band n-type semiconductor properties, high biocompatibility, chemical stability, environmental safety, and low cost. However, the stability of gas sensors based on zinc oxide can be significantly affected by the presence of water vapor in the atmosphere. To address this issue, methods for synthesizing gas-sensitive layers of zinc oxide with improved resistance to water vapor have been explored. The effects of different seed layers and the use of additional precursors have been studied in order to optimize the sensor properties of zinc oxide nanostructures. Methods have been developed for the synthesis of sacrifically doped ZnO nanorods using ZnO-SiO2 net-like nanocomposites and ZnO nanoparticles as seed layers. Analysis of the composition of the resulting layers showed that the use of the sacrifice doping approach allows controlling the content of oxygen vacancies on the surface. Sensor layers consisting of zinc oxide nanorods synthesized using sacrifice doping on the seed layer of ZnO nanoparticles exhibited a response to vapor of volatile organic compounds, with almost no response to water vapor. Nanorods can significantly increase the active adsorption area for water vapor, as they can act as capture traps for water molecules and do not interfere with the measurement readings from sensors. Sensor layers on the seed layers of ZnO-SiO2 nanocomposites responded to both vapor of volatile organic compounds and water vapor. This phenomenon can be explained by the presence of silicon dioxide in the composite material, which attracts water molecules. This affects the final performance of the gas sensor. The results can be used to develop highly efficient sensors for industrial, medical, and food applications.

Driven by the growth of Industry 4.0, advanced studies on machine integration and manufacturing automation through IoT systems are progressing rapidly. Specifically, in milling applications of CFRP (Carbon Fiber Reinforced Polymer) composites, acoustic emission sensors employing piezoelectric transducers have been used to generate acoustic maps. These maps are crucial for monitoring the condition of both the tool and the workpiece, providing a visual analysis of the tool-workpiece interaction that facilitates decision-making by the operator in case of failures. Traditionally, creating acoustic maps that visualize the process and correlate with machining conditions requires an external synchronization signal, usually provided by an encoder attached to the spindle. This study introduces an innovative technique that uses the image generated by the acoustic map to perform automatic alignment during the map's production, eliminating the need for an external synchronization signal. Implemented in Matlab software, the algorithm uses digital filters to extract features, recognizing the pattern of the cutting edges in the map. Based on the misalignment of the cutting edges in the image, the algorithm automatically adjusts the rotation parameters in the map reconstruction, resulting in an accurate representation of the process. The results demonstrate that under specific machining conditions, the need for an external synchronization signal to construct an acoustic map is unnecessary, making the data acquisition system simpler, more economical, and computationally less demanding. This advancement significantly contributes to the development of embedded IoT sensor solutions tailored for Industry 4.0.

This paper introduces a model-based fault diagnosis approach that combines Bayesian Networks with an Extended Kalman Filter (EKF) to detect and diagnose faults in a vehicle's lateral dynamic system. Model-based techniques, particularly Bayesian Networks, have gained popularity in engineering applications due to their robustness and ability to reduce the computational cost associated with empirical models. The proposed method leverages the strengths of these techniques by calculating residuals for yaw rate, wheel slip rate, and steering angle, comparing sensor data with data obtained from analytical models. This comparison enables the identification of discrepancies that may indicate faults within the system. The EKF plays a crucial role in estimating the vehicle's speed by fusing data from GPS and accelerometer sensors. This estimation allows the system to detect potential errors in the wheel speed sensors, which are critical for maintaining accurate vehicle dynamics. In the event of an incorrect wheel speed measurement, the system detects the error, and the erroneous data is replaced with the speed value derived from sensor fusion. The proposed fault diagnosis method was implemented in C++ within the Robot Operating System (ROS) framework. To enhance usability and provide real-time error visualization during tests, a Human-Machine Interface (HMI) was developed. Real-time testing of the system has been performed on a test vehicle in a controlled traffic-free area. To highlight the impact of the EKF on system performance test scenarios for the left and right wheel speed sensors were chosen where faults have been injected to the measurements. The results clearly indicate that the designed algorithm can accurately detect and diagnose the faults correctly while ensuring the reliability of the dynamic model, demonstrating its effectiveness and potential for real-world automotive applications.

Essential oils are widely used in aromatherapy, food and pharmaceutical industries. They contain a range of electroactive components in particular natural phenolic antioxidants like eugenol, trans-anethole, thymol, carvacrol, and vanillin. Therefore, sensitive voltammetric determination of these compounds is of practical interest. To solve this problem, sensors based on the electropolymerized dyes have been developed. Pyrogallol red, equimolar mixture of phenolic red and p-coumaric acid, thymolphthalein, bromocresol purple have been used as monomers for the formation of non-conductive coverages as confirmed by cyclic voltammetry data. Polymer layer improves the selectivity and sensitivity of sensors to target analytes. Layer-by-layer combination with carbon nanomaterials (single- or multi-walled nanotubes) has been applied to provide sufficient electroconductivity of the electrodes. The effect of electropolymerization conditions has been studied on the basis of target phenolic antioxidants response. The surface and electron transfer parameters of the developed sensors have been characterized by scanning electron microscopy, voltammetry, and electrochemical impedance spectroscopy. The sensors have been used in quantification of eugenol, trans-anethole, thymol, carvacrol, and vanillin under conditions of differential pulse voltammetry in Britton-Robinson buffer. The detection limits in the range of 0.037–0.730 μM have been achieved. The analytical characteristics of the sensors are comparable or improved vs. existing ones. The major advantage of the sensors developed is high selectivity of response in the presence of other natural phenolic antioxidants.

Parkinson’s disease (PD) is a neurodegenerative disorder that affects the central nervous system and leads to progressive degeneration of neurons that results in movement slowness, mental health problems, speaking difficulties, etc. In the past 25 years, the prevalence of PD has doubled. Global estimates revealed that over 8.5 million cases have been identified so far. Thus, early and accurate detection of PD is crucial for treatment. Traditional detection methods are subjective and prone to delays as they are reliant on clinical evaluation and imaging. Alternatively, artificial intelligence (AI) has recently emerged as a transformative technology in the healthcare sector showing decent and promising results. However, an effective algorithm needs to be investigated for the most accurate prediction of a particular disease. Thus, this paper explores the ability of different machine learning algorithms for the effective detection of PD. A total of 26 algorithms were implemented using the Scikit-Learn library on the Oxford PD detection dataset. It is a collection of 195 voice measurements recorded from 31 individuals, of which 23 have PD. The implemented algorithms are logistic regression, decision tree, random forest, k-nearest neighbors, support vector machine, Gaussian naïve bayes, multi-layered perceptron (MLP), extreme gradient boosting, adaptive boosting, stochastic gradient descent, gradient boosting machine, extra tree classifier, light gradient boosting machine, categorical boosting, Bernoulli naïve bayes, complement naïve bayes, multinomial naïve bayes, histogram-based gradient boosting, nearest centroid, radius neighbors classifier, logistic regression with elastic net regularization, extreme learning machine, ridge classifier, huber classifier, perceptron classifier, and voting classifier. Among them, MLP outperformed the other algorithms by testing accuracy of 95%, precision of 94%, sensitivity of 100%, F1 score of 97%, and AUC of 98%. Thus, it successfully discriminates healthy individuals from those with PD, thereby helping for accurate early detection of PD for new patients using their voice measurement.

Gaussian derivatives offer valuable capabilities for analyzing image characteristics such as structure, edges, texture, and features, which are essential aspects in the assessment of image quality. Present days Convolutional Neural Networks (CNN) gained its importance in all computer vision applications and also in image quality assessment domain. Because of these characteristics of gaussian derivative that plays a major role in assessing image quality, this work is carried by combining these characteristics with the CNNs to better extract the features for assessing the quality of an image. While CNNs have demonstrated their ability to handle distortion effectively, they are limited in their capacity to capture features at different scales, making them inadequate in dealing with significant variations in object size. Consequently, the concept of spatial pyramid pooling (SPP) has been introduced to address this limitation in image quality assessment (IQA). SPP involves pooling the spatial feature maps from the highest convolutional layers into a feature representation of fixed length. Additionally, through the utilization of convolutional block attention module (CBAM) a module designed for the interpretation of images and local importance pooling (LIP) proposed method for No-reference image quality assessment has demonstrated improved accuracy, generalization, and efficiency compared to conventional (or) traditional IQA methods.

The regulation of the orientation of a flying aircraft under autopilot is a multifaceted and crucial task that requires accuracy and flexibility. To do this, it is essential to have a complex control system that is furnished with an advanced controller capable of actively monitoring and modifying the flying characteristics of the aircraft. This must possess the ability to react dynamically to a range of disturbances experienced throughout the flight, including turbulence, fluctuations in wind, and other pertinent environmental elements. Through real-time adjustment of the flying attitude, the control system guarantees that the aircraft maintains its planned trajectory, stability, and safety along the whole trajectory. Typically, PID controllers are used to regulate the longitudinal direction of flights. However, these offline tuned controllers lack automation and are unable to adjust parameters in response to inherent disturbances seen in practice. Thus, this paper proposes online tuning techniques that are created using artificial intelligence (AI) mechanisms such as fuzzy logic and neural networks. The philosophy involved in this work is the online tuning of PID gain parameters by applying both aforementioned intelligent methods. The study also implements many traditional PID tuning techniques and compares the most effective tuning method with online approaches. To evaluate the effectiveness of online controllers and the optimal traditional PID controller, these controllers are subjected to different disturbances, and their performance is evaluated based on time-domain transient characteristics. The analysis revealed that the intelligent fuzzy controller-based PID controller outperformed alternative tuning techniques in terms of time performance indices such as delay time, rise time, peak time, and settling time, which are improved by 5.88%, 3.26%, 8.05%, and 55.71% respectively when compared to traditional PID tuning methods. The overall comprehensive analysis is conducted using MATLAB/Simulink, and the most optimal online tuning approach is recommended for the controller design.

The manufacturing process can be classified as one of the most crucial for the functioning of the industry, as it forms the foundation of the entire industrial mechanism. Over the years, with the advent of Industry 4.0, this process has been refined, achieving a high degree of precision. The pillar of Industry 4.0 is the sensors that perform data acquisition and enable subsequent analyses. One of the sensors used in this process is the Acoustic Emission (AE) sensor; however, its current use in the industry is highly complex, and still in the embryonic stage, relying on multiple software and computational tools for data acquisition and processing. In this context, this work proposes a system comprising hardware and embedded software designed to facilitate the acquisition of acoustic emission signals through a developing wireless IoT sensor. This system is part of a robust ecosystem, designed to support the implementation of fault diagnosis models, feature extraction, pattern classification, and integration with cloud storage systems. The results demonstrate that the developed system has become a viable solution within an ecosystem for applications using wireless IoT acoustic emission sensors, reducing all complex apparatus to a single tool. Moreover, it enabled remote configuration of the acoustic emission sensor during tests, supporting the inclusion of mathematical and computational models for feature extraction and failure analysis, allowing the registration and organization of tests through forms for registration and document management, without demanding external computers. Such contributions have allowed for the expansion of the use of acoustic emission sensors (AE), aligned with the demands of Industry 4.0, and have promoted significant advancements in the application of IoT sensors while contributing to the efficiency of manufacturing processes.