The use of wearable sensors for health monitoring is rapidly growing. Over the past decade, wearable technology has gained much attention from the tech industry for commercial reasons and the interest of researchers and clinicians for reasons related to its potential benefit on patients’ health. Wearable devices use advanced and specialized sensors able to monitor not only activity parameters, such as heart rate or step count, but also physiological parameters, such as heart electrical activity or blood pressure. Electrocardiogram (ECG) monitoring is becoming one of the most attractive health-related features of modern smartwatches, and, because cardiovascular disease (CVD) is one of the leading causes of death globally, the use of a smartwatch to monitor patients could greatly impact the disease outcomes on health care systems. Commercial wearable devices are able to record just single-lead ECG using a couple of metallic contact dry electrodes. This kind of measurement can be used only for arrhythmia diagnosis. For the diagnosis of other cardiac disorders, additional ECG leads are required. In this study, we characterized an electronic interface to be used with multiple contactless capacitive electrodes in order to develop a wearable ECG device able to perform several lead measurements. We verified the ability of the electronic interface to amplify differential biopotentials and to reject common-mode signals produced by electromagnetic interference (EMI). We developed a portable device based on the studied electronic interface that represents a prototype system for further developments. We evaluated the performances of the developed device. The signal-to-noise ratio of the output signal is favorable, and all the features needed for a clinical evaluation (P waves, QRS complexes and T waves) are clearly readable.

In this work an array of chemical sensors for gas detection has been developed, starting with a commercial sensor platform developed by Microchip (GestIC), which is normally used to detect, trace, and classify hand movements in space. The system is based on electric field changes, and in this work, it has been used as mechanism revealing the adsorption of chemical species CO₂ and O₂. The system is composed of five electrodes, and their responses were obtained by interfacing the sensors with an acquisition board based on an ATMEGA 328 microprocessor (Atmel MEGA AVR microcontroller). A dedicated measurement chamber was designed and prototyped in acrylonitrile butadiene styrene (ABS) using an Ultimaker3 3D printer. The measurement cell size is 120 × 85 mm. Anthocyanins (red rose) were used as a sensing material in order to functionalize the sensor surface. The sensor was calibrated using different concentrations of oxygen and carbon dioxide, ranging from 5% to 25%, mixed with water vapor in the range from 50% to 90%. The sensor exhibits good repeatability for CO₂ concentrations. To better understand the sensor response characteristics, sensitivity and resolution were calculated from the response curves at different working points. The sensitivity is in the order of magnitude of tens to hundreds of µV/% for CO₂, and of µV/% in the case of O₂. The resolution is in the range of 10⁻¹%–10⁻³% for CO₂, and it is around 10⁻¹% for O₂. The system could be specialized for different fields, for environmental, medical, and food applications.

Background: Continuous glucose monitoring (CGM) systems offer the advantage of noninvasive monitoring and continuous data on glucose fluctuations. This study introduces a new model that enables the generation of synthetic but realistic databases that integrate physiological variables and sensor attributes into a dataset generation model and this, in turn, enables the design of improved CGM systems. Methods: The presented approach uses a combination of physiological data and sensor characteristics to construct a model that considers the impact of these variables on the accuracy of CGM measures. A dataset of 500 sensor responses over a 15-day period is generated and analyzed using machine learning algorithms (random forest regressor and support vector regressor). Results: The random forest and support vector regression models achieved Mean Absolute Errors (MAEs) of $16.13$ mg/dL and $16.22$ mg/dL, respectively. In contrast, models trained solely on single sensor outputs recorded an average MAE of $11.01\pm 5.12$ mg/dL. These findings demonstrate the variable impact of integrating multiple data sources on the predictive accuracy of CGM systems, as well as the complexity of the dataset. Conclusions: This approach provides a foundation for developing more precise algorithms and introduces its initial application of Tiny Machine Control Units (MCUs). More research is recommended to refine these models and validate their effectiveness in clinical settings.

The current study aims to train and benchmark AI models tailored for the detection of microplastic in water from scattered signals. We trained two different models, the first based on a Multi-Layer Perceptron (MLP) and the second on a Gated Recurrent Unit (GRU). A Neural Architecture Search algorithm was used to determine the optimal configuration for each of the two models. Moreover, for deployment on edge devices, a specific custom-made compiler was designed and used. The compiler is specifically designed for TinyML applications and, therefore, for resource-constrained devices. It bypasses traditional inference engines, compiling the NNs to native C code using only standard C libraries. Our approach demonstrated better performance compared to state-of-the-art frameworks such as ONNX Runtime, achieving better latency, memory usage, energy consumption, and a higher portability. This highlights the potential of our method for efficient and effective microplastic detection in environmental monitoring.

Microplastics, small particles of plastic found in the environment, have become an increasingly worrying topic in recent years. This paper compares a statistical detection model to classifiers from various supervised learning paradigms in order to detect microplastics. The objective of this paper is to present a benchmark for detecting microplastics using statistical and machine learning models. The main goal is to assess and compare their performance when the defined parameters deviate from the optimal solution of the respective model. Results are presented in terms of probability error, comparing the performance of the machine learning techniques to the statistical model. The study considers a range of signal-to-noise ratios and a priori event probabilities, focusing on the classifiers’ ability to handle amplitude variability and threshold variation. Results show that as the number of analyzed particles in the flow increases, the detection performance improves, with Support Vector Machine, Linear Discriminant Analysis and Naive Bayes standing out from the other methods.