1,721,034 research outputs found
Metrological evaluation of contactless sleep position recognition using an accelerometric smart bed and machine learning
No full textPrecise categorization of sleep postures is essential for evaluating overall physical and mental condition. A smart bed was constructed with the microelectromechanical systems (MEMS) accelerometer sensor and an STM 32-bit microcontroller board. This work applies machine learning (ML) methods to acceleration data to accurately categorize four main sleep positions: right side, left side, prone, and supine without any wearable devices. In this work, the efficiency of 9 ML methods is examined. These algorithms include Logistic Regression (LR) with one-vs- rest and multinomial logistic regression types, Linear Discriminant Analysis (LDA), K-Nearest Neighbors Classification (KNN), Classification and Regression Trees (CART), Naive Bayes (NB), Support Vector Machines (SVMs) with one-vs-one and one-vs-rest types, and Random Forest (RF). The best hyperparameters of each model was accomplished, based on GridSearchCV. The K-fold cross-validation with the assessing measurement stability results indicate that the LG-OvR, LDA, and RF models have the best performance, whereas LG-OvR model possesses accuracy rates of almost 99 %. Furthermore, precision, recall and F1-score are calculated with minimum value of 0.95 for all chosen models. The training and test time are also presented for the selected models. This research has important implications for healthcare, sports medicine, and ergonomics, demonstrating the potential of Artificial Intelligence (AI) approaches in improving sleep monitoring methods
Comparison of Machine Learning Algorithms for Heartbeat Detection Based on Accelerometric Signals Produced by a Smart Bed
Full text linkThis work aims to compare the performance of Machine Learning (ML) and Deep Learning (DL) algorithms in detecting users’ heartbeats on a smart bed. Targeting non-intrusive, continuous heart monitoring during sleep time, the smart bed is equipped with a 3D solid-state accelerometer. Acceleration signals are processed through an STM 32-bit microcontroller board and transmitted to a PC for recording. A photoplethysmographic sensor is simultaneously checked for ground truth reference. A dataset has been built, by acquiring measures in a real-world set-up: 10 participants were involved, resulting in 120 min of acceleration traces which were utilized to train and evaluate various Artificial Intelligence (AI) algorithms. The experimental analysis utilizes K-fold cross-validation to ensure robust model testing across different subsets of the dataset. Various ML and DL algorithms are compared, each being trained and tested using the collected data. The Random Forest algorithm exhibited the highest accuracy among all compared models. While it requires longer training time compared to some ML models such as Naïve Bayes, Linear Discrimination Analysis, and K-Nearest Neighbour Classification, it keeps substantially faster than Support Vector Machine and Deep Learning models. The Random Forest model demonstrated robust performance metrics, including recall, precision, F1-scores, macro average, weighted average, and overall accuracy well above 90%. The study highlights the better performance of the Random Forest algorithm for the specific use case, achieving superior accuracy and performance metrics in detecting user heartbeats in comparison to other ML and DL models tested. The drawback of longer training times is not too relevant in the long-term monitoring target scenario, so the Random Forest model stands out as a viable solution for real-time ballistocardiographic heartbeat detection, showcasing potential for healthcare and wellness monitoring applications
Artificial Intelligence Implementation in Internet of Things Embedded System for Real-Time Person Presence in Bed Detection and Sleep Behaviour Monitor
No full textThis paper works on detecting a person in bed for sleep routine and sleep pattern monitoring based on the Micro-Electro-Mechanical Systems (MEMS) accelerometer and Internet of Things (IoT) embedded system board. This work provides sleep information, patient assessment, and elderly care for patients who live alone via tele-distance to doctors or family members. About 216,000 pieces of acceleration data were collected, including three classes: no person in bed, a static laying position, and a moving state for Artificial Intelligence (AI) application. Six well-known Machine-Learning (ML) algorithms were evaluated with precision, recall, F1-score, and accuracy in the workstation before implementing in the STM32-microcontroller for real-time state classification. The four best algorithms were selected to be programmed into the IoT board and applied for real-time testing. The results demonstrate the high accuracy of the ML performance, more than 99%, and the Classification and Regression Tree algorithm is among the best models with a light code size of 1583 bytes. The smart bed information is sent to the IoT dashboard of Node-RED via a Message Queuing Telemetry broker (MQTT)
A Standard Cell Hardware Implementation for Finite-Difference Time Domain (FDTD) Calculation
No full textSeveral inherent characteristics make the Finite -Difference Time Domain (FDTD) algorithm almost ideal for the analysis of a wide class of microwave and highfrequency circuits as testified by the great number of papers appeared in the last two decades and by the presence of many software packages on the present market. The application of the FDTD method to practical, three-dimensional problems, however, is often limited by the demand of very large computational resources. In this paper, the architecture of a digital system, dedicated to the solution of the 3D FDTD algorithm and based on a custom VLSI chip, which implements the “field-update” engine, is described. The system is conceived as a PCB module communicating with a host personal computer via a PCI bus and accommodating dedicated synchronous DRAM hanks as well. Expectations are that significant speed-up, with respect to state-of-the-art software implementations of the FDTD algorithm, can be achieved
Detection and analysis of heartbeats in seismocardiogram signals
Full text linkThis paper presents an unsupervised methodology to analyze SeismoCardioGram (SCG) signals. Starting from raw accelerometric data, heartbeat complexes are extracted and annotated, using a two-step procedure. An unsupervised calibration procedure is added to better adapt to different user patterns. Results show that the performance scores achieved by the proposed methodology improve over related literature: on average, 98.5% sensitivity and 98.6% precision are achieved in beat detection, whereas RMS (Root Mean Square) error in heartbeat interval estimation is as low as 4.6 ms. This allows SCG heartbeat complexes to be reliably extracted. Then, the morphological information of such waveforms is further processed by means of a modular Convolutional Variational AutoEncoder network, aiming at extracting compressed, meaningful representation. After unsupervised training, the VAE network is able to recognize different signal morphologies, associating each user to its specific patterns with high accuracy, as indicated by specific performance metrics (including adjusted random and mutual information score, completeness, and homogeneity). Finally, a Linear Model is used to interpret the results of clustering in the learned latent space, highlighting the impact of different VAE architectural parameters (i.e., number of stacked convolutional units and dimension of latent space)
Physical modeling of silicon microstrip detectors: influence of the electrode geometry on critical electric fields
No full textIn this paper, a computer-based analysis of AC-coupled silicon microstrip detectors is presented. The study aims at investigating the main geometrical parameters responsible for potentially critical effects, such as early micro-discharges and breakdown phenomena. The adoption of CAD tools allows for evaluating the actual field distribution within the device, and makes it possible to identify critical regions. The adoption of overhanging metal strips is shown to have a positive impact on the electric field distribution, reducing corner effects and thus minimizing breakdown risk
Seismocardiography-based detection of heartbeats for continuous monitoring of vital signs
No full textThis paper presents an automated procedure for analyzing SeismoCardioGraphic (SCG) traces, i.e. chest vibrations induced by heart activity. Such signals are acquired by means of an inexpensive, compact accelerometer device, placed over the subject's sternum and held in place by a light strap. The methodology for beat detection and annotation features a preliminary coarse beat detection phase, followed by a self-managed calibration, that is subsequently leveraged to carry out actual SCG annotation and beat localization. The performance of such process was measured and compared against the ECG gold standard, used only to provide ground-truth beat-to-beat intervals (i.e. R-peak landmarks). Results show high average values of sensitivity and precision (99.1% and 97.9%, respectively). The coefficient of determination R2 reaches a 0.984 value, which shows that the algorithm is able to temporally localize SCG complexes in a very consistent way, compared to the ECG gold standard; indeed, the variability between the ECG and SCG beat to beat measures is found to be very limited (σdiff ≈ 8.5 ms, i.e. less than one sample interval)
Analysis and test of overhanging-metal microstrip detectors
No full textThe adoption of overhanging-metal contacts have been suggested as an effective mean to limit breakdown risks in heavy-damaged, high-voltage biased microstrip detectors. In this summary, the influence of such overhangs on device noise parameters is analyzed, with particular reference to the interstrip capacitance. Data have been collected on a set of detectors featuring variable overhang extensions and different width/pitch ratios, and numerical simulation has been exploited to provide physical interpretation of the experimental findings. In particular, the non-trivial dependence of interstrip capacitance over geometrical parameters is discussed. By looking at leakage currents and charge-collection as well, it is shown that limited-extension overhangs still have highly beneficial effects on the breakdown properties, while having no practical drawbacks on the detector performanc
- …
