300 nm in thick β-Ga₂O₃ films were deposited by HF magnetron sputtering at temperature of 650 ºC and subsequently annealed at temperature of 900 ºC. Sensors with Pt contacts were produced based on these films. Structural and gas sensitive properties of β-Ga₂O₃ thin films were studied. As-deposited films demonstrated relatively low response and significant drift of gas sensitive characteristics. Subsequently annealing of films led to significant increase in responses to hydrogen-containing gases and long-term stability of gas sensitive characteristics. The responses to H₂ and CH₄ for annealed samples were ~140 arb. un. and ~38 arb. un. at operating temperatures of 600 ºC and 650 ºC, correspondingly. Changes in gas sensitive characteristics of β-Ga₂O₃ thin films are caused by refurbishment of microstructure of their surface. High temperature deposition and subsequently annealing are promising tools for optimizing the gas sensitive properties of β-Ga₂O₃ thin films for high-temperature applications.

Wireless sensor nodes (WSNs) have attracted much attention in recent years as a device that can be used for structural health monitoring. Most of the WSNs are used battery as a source of power supply and the energy will be depleted over the time. The self-power WSNs play an important role and electromagnetic vibration energy harvesters (EMVEHs) is one of the options for the continuous energy supply. As an important component of the electromagnetic vibration energy harvesting system, rectifiers play an important role in the entire system which convert the AC power generated by vibration into DC power. The basic half-wave rectifier consists of a rectifier diode, a filter capacitor, and a load, while the basic full-bridge rectifier is composed of four rectifier diodes, a filter capacitor, and a load. Compared to commonly adopted full-bridge rectifiers, half-wave rectifiers are often not favored by researchers due to their inability to obtain full cycle AC power. However, limited researchers focused on the cantilever beam type electromagnetic vibration energy harvester, half-wave rectifiers may exhibit more advantages than full-bridge rectifiers in some cases. This paper presented an equivalent circuit model of the harvester derived from mechanical parameters, and then verifies the feasibility of the rectifier circuit applied to the harvester using finite element simulation software. The experimental results present the performance of two circuits under different sweep mode condition and the comparative analysis on the maximum power and bandwidth is studied. The results indicate that although the ripple coefficient of the half-wave rectifier is greater than that of the full-bridge circuit, it exhibits certain advantages in bandwidth and maximum power. Finally, a preliminary analysis of the mechanism of the experimental phenomenon is provided.

Tin oxide semiconducting material has been emerged as attractive active thin-layer alternative for developing high-performance and low-cost microelectronic devices. The effects of oxygen concentration and growth technique during the deposition process on the electrical properties of SnO_x alloy should be investigated for developing new eco-friendly photosensors and photovoltaic devices. The present work aims to predict the electrical key governing parameters throughout the device developing processes such as the Energy level values and band-gap energy as function of the injected oxygen concentrations. For realization, over 1000 data points were collected by modeling the effect of oxygen contents on the SnO_x electrical properties using Density Function Theory (DFT). Through extensive Machine Learning (ML) analysis, the impact of the oxygen concentration on the electrical properties and the material type is well predicted, where the applied ML prediction model for band-gap energy showed a good correlation between predicted values and the calculated ones using DFT computations. It is revealed that the combined DFT-ML-based approach can be robust, accurate and easy-to-implement tool to study and accelerate the developing of new highly efficient materials for microelectronic applications

Antibiotics are commonly used in veterinary medicine purposes to prevent and treat bacterial infections in food-producing animals [1]. As a result, antibiotic residues may be present in animal-derived products (meat, eggs, milk and dairy products) and eventually consumed by humans, potentially causing resistance to antibiotics and other health issues [2]. Oxytetracycline (OTC) is a commonly used antibiotic in farming animals; in order to protect consumers, a maximum residue limit (MRL) of 100 µg L^-1 has been established by the European Union for OTC residues in milk [3].

Although liquid chromatography is considered the default and most powerful analytical methodology for antibiotics residues [4], biosensors constitute an alternative technology that offers rapidity, low-cost and scope for on-site or field assays. Biosensors are based on a signal arising from the interaction event between the target and a selective bioreceptor [5]. In particular, aptasensors are biosensing devices utilizing aptamers (short single-stranded oligonucleotides with high affinity towards specific targets) as bioreceptors; aptamers exhibit some unique and advantageous properties compared to other bioreceptors and, therefore, different aptasensing strategies have been applied to the assay of antibiotics including OTC [6].

In this work, we describe microfabricated gold-based electrochemical aptasensors for label-free detection of OTC. Thin-film gold electrodes were fabricated by sputtering of gold on a Kapton film. A thiol-modified OTC-specific aptamer was immobilized on the electrode surface exploiting thiol-gold interactions. Then, the sample was incubated with the aptamer-modified electrode to achieve selective capture of OTC. Finally, the binding between the immobilized aptamer and OTC was monitored electrochemically using cyclic voltammetry (CV), differential-pulse voltammetry (DPV) and electrochemical impedance spectroscopy (EIS) using the Fe(CN)₆⁴⁺/Fe(CN)₆³⁺ redox probe. Different experimental variables were studied and the metrological features for OTC were derived. Finally, proof-of-principle applicability of the sensors was demonstrated for the determination of OTC in milk.

This work was supported by European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No. 101007299.

[1] Virto M., Santamarina-García G, Amores G, Hernández I (2022) Antibiotics in Dairy Production: Where Is the Problem? Dairy 3:541-564, doi:https://doi.org/10.3390/dairy3030039
[2] Landers, T.F.; Cohen, B.; Wittum, T.E., Larson, E.L. A Review of Antibiotic Use in Food Animals: Perspective, Policy, and Potential. Public Health Rep. 2012, 127(1), 4–22. doi: http://www.jstor.org/stable/41639470.
[3] Commission Regulation (EU) No. 37/2010, https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32010R0037
[4] Peris-Vicente, J.; Peris-García, E.; Albiol-Chiva, J.; Durgbanshi, A.; Ochoa-Aranda, E.; Carda-Broch, S.; Bose, D.; Esteve-Romero, J. Liquid chromatography, a valuable tool in the determination of antibiotics in biological, food and environmental samples. Microchem. J. 2022, 177, 107309. doi: https://doi.org/10.1016/j.microc.2022.107309.
[5] Zhou, C.; Zou, H.; Sun, C.; Li. Y. Recent advances in biosensors for antibiotic detection: Selectivity and signal amplification with nanomaterials. Food Chem. 2021, 361, 130109. doi: https://doi.org/10.1016/j.foodchem.2021.130109.
[6] Liang, G.; Song, L.; Gao, Y.; Wu, K.; Guo, R.; Chen, R.; Zhen, J.; Pan, L. Aptamer Sensors for the Detection of Antibiotic Residues— A Mini-Review. Toxics 2023, 11, 513. doi: https://doi.org/10.3390/toxics11060513.

In this paper, we present a comparative study of several machine learning (ML) approaches for accurate office room occupancy detection through the analysis of multi-sensor data. Our study utilizes the Occupancy Detection dataset, which incorporates data from Temperature, Humidity, Light, and CO2 sensors, with ground-truth labels obtained from time-stamped images captured at minute intervals. Traditional ML techniques including Decision Trees, Gaussian Naïve Bayes, K-Nearest Neighbors, Logistic Regression (LR), Support Vector Machines (SVM), Multilayer Perceptron (MLP), and Quadratic Discriminant Analysis are compared alongside advanced ensemble methods like Random Forest, Bagging, AdaBoost, GradientBoosting, ExtraTrees as well as our custom voting and multiple stacking classifiers. Hyperparameter optimization is performed for selected models before being integrated into ensemble methods. The performances of the models were evaluated through rigorous cross-validation experiments. The results obtained highlight the efficacy and suitability of varying candidate and ensemble methods, demonstrating the potential of machine learning techniques for enhancing the detection accuracy. Notably, LR and SVM exhibited superior performance, achieving average accuracies of 98.88 ± 0.70% and 98.65 ± 0.96%, respectively. Additionally, our custom voting and stacking ensembles demonstrated improvements in classification outcomes compared to base ensemble schemes, as indicated by various the various evaluation metrics.

The use of near-infrared (NIR) spectroscopic sensing system for the prediction of milk quality indicators in cow milk was described in this study. The objective of this study was to investigate the measurement accuracy of the sensing system developed in our study. Three major milk constituents (fat, protein, and lactose), milk urea nitrogen (MUN) and somatic cell count (SCC) of two Holstein cows belonging to Hokkaido University dairy farm were measured. Milk spectra with a wavelength range of 700 to 1050 nm and milk samples were collected in every 20 seconds during milking using the NIR spectroscopic sensing system. Calibration models were developed using partial least squares (PLS) regression analysis and the precision and accuracy of the models were validated. The statistical results obtained for milk fat and protein contents were very high while the results obtained for milk lactose, MUN and SCC were sufficiently high. This suggest that the NIR spectroscopic sensing system developed in this study could be used for online real-time milk quality determination of each cows’ milk constituents, MUN and SCC during milking. This sensing system could assist dairy farmers in solving the challenge of effective individual cow management, resulting in smart dairy farming.

Individuals who use the majority of their period of time indoors are especially sensitive to indoor air quality (IAQ), which has a significant impact on their general well-being and health. Traditional IAQ measurement techniques, however, are frequently pricy, complicated, and labor-intensive. In this study, we used a low-cost, simple-to-use, and handy sensor system to track the levels of carbon dioxide (CO₂), nitrogen dioxide (NO₂), ozone (O₃), particulate matter (PM_1.0, PM_2.5, and PM₁₀), temperature and relative humidity (RH) in a laboratory at the Bamidele Olomilua University of Education, Science, and Technology in Ikere-Ekiti for a month. We contrasted the outcomes with other benchmarks and WHO recommendations. However, the NO₂ levels (144.00-303.00 ppb) exceeded the suggested levels (National Institute for Occupational Safety and Health (NIOSH) - 70 ppb; National Ambient Air Quality Standards (NAAQS) - 100 ppb; National Environmental Standards and Regulations Enforcement Agency (NESREA) - 120 ppb); and World Health Organization (WHO) - 25 ppb), suggesting a possible cause of indoor contaminants. We also noticed that the temperature and humidity varied considerably all through the day, which had an impact on the inhabitants' thermal comfort and ventilation. The PCA findings indicate that particulate matter, the weather, photochemical reactions, and combustion processes are the key contributors to fluctuation in the air quality measurements. Based on their quantities and relationships, these elements can have a variety of effects on both the natural environment as well as well-being. Our monitoring device can give immediate information and warnings, assisting in locating and reducing indoor airborne pollutants sources and enhancing their indoor air quality (IAQ). This work shows that adopting a low-cost sensor system for IAQ measurement in underdeveloped nations, where such data are sparse and frequently erroneous, is both feasible and beneficial.

While direct analysis of real biosamples is a complex process compared to other systems, it also yields a set of novel and valuable results. Ascorbic acid or Vitamin C plays a pivotal role in our immune functions; however, in contrast to other vitamins, the human body cannot generate it. Therefore, there is a high demand for the development of new tools capable of onsite monitoring and fast analysis of natural sources without any treatment, even in the field and during the transfer and storage of food products. By leveraging the unique features of microelectrodes, we have developed novel combined microelectrodes by modifying carbon fiber (33 μm) coated with an Au nano-film to serve as a working electrode and tiny silver wire as a reference electrode. This microscale tool allows for the direct microelectroanalysis of ascorbic acid in lemon and cactus bodies, serving as biological matrices. Beyond the potential for direct electroanalysis in these bio-matrices, our primary objectives include examining the distribution of ascorbic acid content across different sections of lemon fruit and parts of the cactus plant. Both lemon and cactus are recognized sources of vitamin C. Notably, the micro-size of the combined sensor provides sufficient resolution for microscale analysis in fruit and plant samples without sample treatment. Our electrochemical measurements revealed that the center of the lemon contains notably higher levels of Vitamin C compared to its sides. Also, the levels of Vitamin C are higher in the fresh arms of the cactus compared to the older arms and the cactus trunk. Furthermore, our observations indicated that improper storage of lemon products in the presence of daylight in one week significantly reduces the vitamin C level. Finally, we believe these findings hold significance and practical applicability in the agricultural, medicinal, and food industries.

Climate change is an undoubtable phenomenon. Extreme weather conditions occur at several locations throughout the globe. Prolonged and severe rainfalls, especially when combined to deforestation, often lead to massive river floods. Such natural disasters pose a great danger to humanity.

The present study aims to design a low-cost smart AI powered node, to serve as a flood Early Warning System complete solution. The node is designed to predict forthcoming flood events by capturing and combining critical data related to such phenomena. Such data are the water level at rivers or other water discharge basins, rainfall, soil moisture, and river bank slides. The node will autonomously monitor the above quantities at a high frequency rate, and selectively upload them to a server only when verified conditions for a forthcoming flood will exist. These conditions will be evaluated by the local ML model. Network access of the node is aided by the utilization of an LTE modem provided that cellular network is present.

Further on, datasets referring to actual flood phenomena will be used to train the tinyML AI flood prediction models. After the models are validated, the AI neural networks will be integrated to the node’s Firmware. This will allow each node to reliable predict flood events and issue local and remote alarms. Combination of several nodes at an area of interest will form a robust and reliable Early Warning System.

Additive manufacturing (AM) is a process of creating three-dimensional objects by adding material layer by layer based on a digital model. However, several defects may be encountered during the process that can adversely influence the quality and mechanical properties of the parts. Micrograph data refers to high-resolution images capturing the internal microstructure of the printed material, providing crucial insights for analysis and evaluation of the part's mechanical properties and overall quality. However, obtaining such images can be time consuming, which can slow down the process of quality control. Micrograph data can be used to study different types of pores in the manufactured parts. Pores can vary in size, shape, and distribution, making their accurate classification a complex task. The four main types of pores commonly encountered in AM are process pores, gas pores, lack of fusion pores and cracks. Identifying and differentiating these pore types is crucial to understanding the reasons for porosity and applying effective solutions to reduce their occurrence. This paper presents a hybrid machine learning (ML) approach, which combines image processing and supervised ML algorithm for detecting and classifying the pores in AM from the micrograph data. We compute several pixel based features for e.g. by using Sobel, Gaussian filters on the input micrograph image data. To generate output labels, we use standard image processing algorithm for contour detection of the pore defects, calculate their features for e.g. area, convexity, aspect ratio, circularity etc. and use these features for annotating four different types of pores. Next, we use these input and output data to train a Random Forest classifier, which achieves high accuracy. We will compare our hybrid model-data-driven classifier with a traditional pure data-driven CNN pixel classifier. Our future work involves studying the relationships between process variables, material properties, and pores.