Search Results (978)

Cruciferous vegetables are popular in Asian dishes. However, striped flea beetles prefer to feed on leaves, which can damage the appearance of crops and reduce their economic value. Due to the lack of pest monitoring, the occurrence of pests is often irregular and unpredictable. Regular and quantitative spraying of pesticides for pest control is an alternative method. Nevertheless, this requires manual execution and is inefficient. This paper presents a system powered by solar energy, utilizing batteries and supercapacitors for energy storage to support the implementation of edge AI devices in outdoor environments. Raspberry Pi is utilized for artificial intelligence image recognition and the Internet of Things (IoT). YOLOv5 is implemented on the edge device, Raspberry Pi, for detecting striped flea beetles, and StyleGAN3 is also utilized for data augmentation in the proposed system. The recognition accuracy reaches 85.4%, and the results are transmitted to the server through a 4G network. The experimental results indicate that the system can operate effectively for an extended period. This system enhances sustainability and reliability and greatly improves the practicality of deploying smart pest detection technology in remote or resource-limited agricultural areas. In subsequent applications, drones can plan routes for pesticide spraying based on the distribution of pests. Full article

Salinity intrusion is a growing threat to freshwater resources, particularly in low-lying coastal and estuarine regions, necessitating the development of effective early warning systems (EWS) to support timely mitigation. Although various water quality monitoring technologies exist, many face challenges related to long-term sustainability, ongoing maintenance, and accessibility for local users. This study introduces a novel hybrid real-time salinity intrusion early warning system that uniquely integrates fixed and portable monitoring technologies with strong community participation—an approach not yet widely applied in comparable urban-adjacent delta regions. Unlike traditional systems, this model emphasizes local ownership, flexible data collection, and system scalability in resource-constrained environments. This study presents a real-time salinity intrusion early warning system for Bang Kachao, Thailand, combining eight fixed monitoring stations and 20 portable salinity measurement devices. The system was developed in response to community needs, with local input guiding both station placement and the design of mobile measurement tools. By integrating fixed stations for continuous, high-resolution data collection with portable devices for flexible, on-demand monitoring, the system achieves comprehensive spatial coverage and adaptability. A core innovation lies in its emphasis on community participation, enabling villagers to actively engage in monitoring and decision-making. The use of IoT-based sensors, Remote Telemetry Units (RTUs), and cloud-based data platforms further enhances system reliability, efficiency, and accessibility. Automated alerts are issued when salinity thresholds are exceeded, supporting timely interventions. Field deployment and testing over a seven-month period confirmed the system’s effectiveness, with fixed stations achieving 90.5% accuracy and portable devices 88.7% accuracy in detecting salinity intrusions. These results underscore the feasibility and value of a hybrid, community-driven monitoring approach for protecting freshwater resources and building local resilience in vulnerable regions. Full article

Monitoring Peripheral Oxygen Saturation (SpO₂) is an important vital sign both in Intensive Care Units (ICUs), during surgery and convalescence, and as part of remote medical consultations after of the COVID-19 pandemic. This has made the development of new SpO₂-measurement tools an area of active research and opportunity. In this paper, we present a new Deep Learning (DL) combined strategy to estimate SpO₂ without contact, using pre-magnified facial videos to reveal subtle color changes related to blood flow and with no calibration per subject required. We applied the Eulerian Video Magnification technique using the Hermite Transform (EVM-HT) as a feature detector to feed a Three-Dimensional Convolutional Neural Network (3D-CNN). Additionally, parameters and hyperparameter Bayesian optimization and an ensemble technique over the dataset magnified were applied. We tested the method on 18 healthy subjects, where facial videos of the subjects, including the automatic detection of the reference from a contact pulse oximeter device, were acquired. As performance metrics for the SpO₂-estimation proposal, we calculated the Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and other parameters from the Bland–Altman (BA) analysis with respect to the reference. Therefore, a significant improvement was observed by adding the ensemble technique with respect to the only optimization, obtaining 14.32% in RMSE (reduction from 0.6204 to 0.5315) and 13.23% in MAE (reduction from 0.4323 to 0.3751). On the other hand, regarding Bland–Altman analysis, the upper and lower limits of agreement for the Mean of Differences (MOD) between the estimation and the ground truth were 1.04 and −1.05, with an MOD (bias) of −0.00175; therefore, MOD $\pm 1.96\sigma$ = −0.00175 ± 1.04. Thus, by leveraging Bayesian optimization for hyperparameter tuning and integrating a Bagging Ensemble, we achieved a significant reduction in the training error (bias), achieving a better generalization over the test set, and reducing the variance in comparison with the baseline model for SpO₂ estimation. Full article

Diabetic foot ulcers (DFUs) are a serious complication of diabetes, associated with high recurrence and amputation rates. Adherence to offloading devices is critical for wound healing but remains inadequately monitored in real-world settings. This study evaluates the SmartBoot edge-computing system—a wearable, real-time remote monitoring solution integrating an inertial measurement unit (Sensoria Core) and smartwatch—for its validity in quantifying cadence and step count as digital biomarkers of frailty, and for detecting adherence. Twelve healthy adults wore two types of removable offloading boots (Össur and Foot Defender) during walking tasks at varied speeds; system outputs were validated against a gold-standard wearable and compared with staff-recorded adherence logs. Additionally, user experience was assessed using the Technology Acceptance Model (TAM) in healthy participants (n = 12) and patients with DFU (n = 81). The SmartBoot demonstrated high accuracy in cadence and step count across conditions (bias < 5.5%), with an adherence detection accuracy of 96% (Össur) and 97% (Foot Defender). TAM results indicated strong user acceptance and perceived ease of use across both cohorts. These findings support the SmartBoot system’s potential as a valid, scalable solution for real-time remote monitoring of adherence and mobility in DFU management. Further clinical validation in ongoing studies involving DFU patients is underway. Full article

This paper presents a novel framework for optimizing multimodal large language model (MLLM) inference through task offloading and resource allocation in UAV-assisted satellite edge computing (SEC) networks. MLLMs leverage transformer architectures to integrate heterogeneous data modalities for IoT applications, particularly real-time monitoring in remote areas. However, cloud computing dependency introduces latency, bandwidth, and privacy challenges, while IoT device limitations require efficient distributed computing solutions. SEC, utilizing low-earth orbit (LEO) satellites and unmanned aerial vehicles (UAVs), extends mobile edge computing to provide ubiquitous computational resources for remote IoTDs. We formulate the joint optimization of MLLM task offloading and resource allocation as a mixed-integer nonlinear programming (MINLP) problem, minimizing latency and energy consumption while optimizing offloading decisions, power allocation, and UAV trajectories. To address the dynamic SEC environment characterized by satellite mobility, we propose an action-decoupled soft actor–critic (AD-SAC) algorithm with discrete–continuous hybrid action spaces. The simulation results demonstrate that our approach significantly outperforms conventional deep reinforcement learning methods in convergence and system cost reduction compared to baseline algorithms. Full article

An electronic nose is a device capable of characterizing samples of substances and products by their aroma. The development of such devices relies on a series of non-specific sensors that react to gases and generate different signals, which can be used for compound identification and sample classification. The deployment of such devices often requires the possibility of having remote access over the Internet to manage their operation and to collect the sampled data. In this context, the application of web technologies to the monitoring and supervision of these systems connected to the Internet, which can be considered as an Internet of Things (IoT) device, offers the advantage of not requiring the development of client-side applications. Users can employ a browser to connect to the IoT device and monitor or control its operation. Moreover, web design enables the development of cross-platform web monitoring systems. In addition, the inclusion of the MQTT protocol and the utilization of a virtual private network (VPN) enable a secure transmission and collection of the sampled data. In this work, all these technologies have been applied in the development of a system to manage and collect data to monitor rot in lemons treated with sodium benzoate before harvest. Full article

Purpose: We aimed to evaluate the safety and feasibility of a remote electrocardiogram (ECG) monitoring-based cardiac rehabilitation (CR) program during an early postoperative period in patients who underwent cardiac surgery. Methods: Five days after cardiac surgery, patients were referred to a CR department and participated in a low-intensity inpatient CR program while wearing an ECG monitoring device. Prior to discharge, the patients underwent a cardiopulmonary exercise test (CPET) and squat endurance test to determine the suitable intensity and target heart rate (HR) for home-based CR (HBCR). During 2 weeks of the HBCR period after discharge, patients participated in aerobic and resistance exercises. Electrocardiogram data were transmitted to a cloud, where researchers closely monitored them through a website and provided feedback to the patients via telephone calls. Grip strength (GS), 6 min walk distance (6 MWD), EuroQol-5 dimension (EQ-5D), short-form 36-item health survey (SF-36), and Korean Activity Scale/Index (KASI) were measured at three different time points: 5 d post-surgery (T1), pre-discharge (T2), and 2 weeks after discharge (T3). Squat endurance tests and CPET were performed only at T2 and T3. Result: Sixteen patients completed the study, seven (44%) of whom underwent coronary artery bypass graft surgery (CABG). During the study period between T2 and T3, peak VO₂ improved from 12.39 ± 0.57 to 17.93 ± 1.25 mL/kg/min (p < 0.01). The squat endurance test improved from 16.69 ± 2.31 to 21.81 ± 2.31 (p < 0.01). In a comparison of values of time points between T1 and T3, the GS improved from 28.30 ± 1.66 to 30.40 ± 1.70 kg (p = 0.02) and 6 MWD increased from 249.33 ± 20.92 to 387.02 ± 22.77 m (p < 0.01). The EQ-5D and SF-36 improved from 0.59 ± 0.03 to 0.82 ± 0.03 (p < 0.01) and from 83.99 ± 3.40 to 122.82 ± 6.06 (p < 0.01), and KASI improved from 5.44 ± 0.58 to 26.11 ± 2.70 (p < 0.01). In a subgroup analysis, the CABG group demonstrated a greater increase in 6 MWD (102.29 m, p < 0.01) than the non-CABG group. At the end of the study, 75% of the patients expressed satisfaction with the early CR program guided by remote ECG monitoring. Conclusions: Our findings suggest that early remote ECG monitoring-based CR programs are safe and feasible for patients who have undergone cardiac surgery. Additionally, the program improved aerobic capacity, functional status, and quality of life. Full article

The growing mismanagement of energy resources is a pressing issue that poses significant risks to both individuals and the environment. As energy consumption continues to rise, the ramifications become increasingly severe, necessitating urgent action. In response, the rapid expansion of Internet of Things (IoT) devices offers a promising and innovative solution due to their adaptability, low power consumption, and transformative potential in energy management. This study describes a novel, integrative strategy that integrates IoT and Artificial Neural Networks (ANNs) in a smart monitoring mobile application intended to optimize energy usage and promote sustainability in residential settings. While both IoT and ANN technologies have been investigated separately in previous research, the uniqueness of this work is the actual integration of both technologies into a real-time, user-adaptive framework. The application allows for continuous energy monitoring via modern IoT devices and wireless sensor networks, while ANN-based prediction models evaluate consumption data to dynamically optimize energy use and reduce environmental effect. The system’s key features include simulated consumption scenarios and adaptive user profiles, which account for differences in household behaviors and occupancy patterns, allowing for tailored recommendations and energy control techniques. The architecture allows for remote device control, real-time feedback, and scenario-based simulations, making the system suitable for a wide range of home contexts. The suggested system’s feasibility and effectiveness are proved through detailed simulations, highlighting its potential to increase energy efficiency and encourage sustainable habits. This study contributes to the rapidly evolving field of intelligent energy management by providing a scalable, integrated, and user-centric solution that bridges the gap between theoretical models and actual implementation. Full article

The COVID-19 pandemic exposed the urgent need for scalable, reliable telemedicine tools to manage mild cases remotely and avoid overburdening healthcare systems. This study evaluates StepCare, a remote monitoring medical device, during the first pandemic wave at a single center in Spain. Among 35 patients monitored, StepCare showed high clinical reliability, aligning with physician assessments in 90.4% of cases. Patients and clinicians reported excellent usability and satisfaction. The system improved workflow efficiency, reducing triage time by 25% and associated costs by 84%. These results highlight StepCare’s value as a scalable, patient-centered solution for remote care during health crises. Full article

The energy consumption of current vertical-lifting underwater monitoring devices mainly falls into two categories: one fully supplied by battery packs; and the other partially by battery packs, with the rest from ocean thermal energy. Constrained by battery capacity, their operation time is limited, making long-term remote operations difficult. This study focuses on a device powered entirely by ocean thermal energy, which realizes the absorption and storage of energy through a phase change heat-exchange system, significantly extending its operation cycle and working area. A composite phase change material of n-hexadecane and graphite with a volume ratio of 9:1 is used. The Fluent software 2022 R1, based on the enthalpy-porosity method, simulates the phase change process of the device to analyze the effects of different structures and seawater temperatures. Results show that with the same phase change material volume and inner diameter of the cylindrical heat exchanger, a smaller outer diameter yields better phase change performance. Lower seawater temperature facilitates solidification. Due to natural convection in the liquid phase, the melting time is 520 s and solidification time is 4800 s, with the melting rate far exceeding the solidification rate. Full article

Continuous-time Recurrent Neural Networks (CTRNNs) are well-suited for modeling temporal dynamics in low-power neuromorphic and analog computing systems, making them promising candidates for edge-based human activity recognition (HAR) in healthcare. However, training CTRNNs remains challenging due to their continuous-time nature and the need to respect physical hardware constraints. In this work, we propose G-CTRNN, a novel gradient-based training framework for analog-friendly CTRNNs designed for embedded healthcare applications. Our method extends Backpropagation Through Time (BPTT) to continuous domains using TensorFlow’s automatic differentiation, while enforcing constraints on time constants and synaptic weights to ensure hardware compatibility. We validate G-CTRNN on the WISDM human activity dataset, which simulates realistic wearable sensor data for healthcare monitoring. Compared to conventional RNNs, G-CTRNN achieves superior classification accuracy with fewer parameters and greater stability—enabling continuous, real-time HAR on low-power platforms such as MEMS computing networks. The proposed framework provides a pathway toward on-device AI for remote patient monitoring, elderly care, and personalized healthcare in resource-constrained environments. Full article

In recent years, remote sensing imagery has become indispensable for applications such as environmental monitoring, land use classification, and urban planning. However, the physical constraints of satellite imaging systems frequently limit the spatial resolution of these images, impeding the extraction of fine-grained information critical to downstream tasks. Super-resolution (SR) techniques thus emerge as a pivotal solution to enhance the spatial fidelity of remote sensing images via computational approaches. While deep learning-based SR methods have advanced reconstruction accuracy, their high computational complexity and large parameter counts restrict practical deployment in real-world remote sensing scenarios—particularly on edge or low-power devices. To address this gap, we propose LSANet, a lightweight SR network customized for remote sensing imagery. The core of LSANet is the large separable kernel attention mechanism, which efficiently expands the receptive field while retaining low computational overhead. By integrating this mechanism into an enhanced residual feature distillation module, the network captures long-range dependencies more effectively than traditional shallow residual blocks. Additionally, a residual feature enhancement module, leveraging contrast-aware channel attention and hierarchical skip connections, strengthens the extraction and integration of multi-level discriminative features. This design preserves fine textures and ensures smooth information propagation across the network. Extensive experiments on public datasets such as UC Merced Land Use and NWPU-RESISC45 demonstrate LSANet’s competitive or superior performance compared to state-of-the-art methods. On the UC Merced Land Use dataset, LSANet achieves a PSNR of 34.33, outperforming the best-baseline HSENet with its PSNR of 34.23 by 0.1. For SSIM, LSANet reaches 0.9328, closely matching HSENet’s 0.9332 while demonstrating excellent metric-balancing performance. On the NWPU-RESISC45 dataset, LSANet attains a PSNR of 35.02, marking a significant improvement over prior methods, and an SSIM of 0.9305, maintaining strong competitiveness. These results, combined with the notable reduction in parameters and floating-point operations, highlight the superiority of LSANet in remote sensing image super-resolution tasks. Full article

This study evaluates the accuracy of a low-cost wearable system for the continuous monitoring of vital signs, including heart rate, blood oxygen saturation (Sp${\mathrm{O}}_2$), blood pressure trend (BPT), and body temperature. The prototype was built using the nRF52840 microcontroller, which integrates photoplethysmography and infrared sensors. The heart rate and Sp${\mathrm{O}}_2$ data were collected under three body positions (Rest, Sitting, and Standing), while all measurements were performed using both anatomical configurations: BPT-Finger and BPT-Earlobe. Results were compared against validated commercial devices: UT-100 for heart rate and Sp${\mathrm{O}}_2$, G-TECH LA800 for blood pressure, and G-TECH THGTSC3 for body temperature. Ten participants were monitored over a ten-day period. Bland–Altman analysis revealed clinically acceptable agreement thresholds of ±5 mmHg for blood pressure, ±5–10 bpm for heart rate, ±4% for Sp${\mathrm{O}}_2$, and ±0.5 °C for temperature. Both wearable configurations demonstrated clinically acceptable agreement across all vital signs. The BPT-Earlobe configuration exhibited superior stability and lower variability in the Rest and Sitting positions, likely due to reduced motion artifacts. Conversely, the BPT-Finger configuration showed higher Sp${\mathrm{O}}_2$ accuracy in the Standing position, with narrower limits of agreement. These findings highlight the importance of sensor placement in maintaining measurement consistency across physiological conditions. With an estimated cost of only ~USD 130—compared to ~USD 590 for the commercial alternatives—the proposed system presents a cost-effective, scalable, and accessible solution for decentralized health monitoring, particularly in underserved or remote environments. Full article

Field-based landslide mapping is a crucial task for geo-hydrological risk assessment but is often limited by the lack of integrated tools to capture accurate spatial and temporal data. This research investigates a Direct Numerical Cartography (DNC) system’s ability to capture both spatial and temporal landslide features during fieldwork. DNC enables fully digital surveys, minimizing errors and delivering real-time, spatially accurate data to experts on site. We tested an integrated approach combining the Ground Operative System for GIS Input Remote-data Acquisition (GOGIRA) with the Expert-based Multitemporal AI Detector (ExMAD). GOGIRA is a low-cost system for efficient georeferenced data collection, while ExMAD uses AI and multitemporal Sentinel-2 imagery to detect landslide triggering times. Upgrades to GOGIRA’s hardware and algorithms were carried out to improve its mapping accuracy. Field tests in Western Greece compared data to 64 expert-confirmed landslides, with the Range-R device showing a mean spatial error of 50 m, outperforming the tripod-based UGO device at 82 m. Operational factors like line-of-sight obstructions and terrain complexity affected accuracy. ExMAD applied a pre-trained U-Net convolutional neural network for automated temporal trend detection of landslide events. The combined DNC and AI-assisted remote sensing approach enhances landslide inventory precision and consistency while maintaining expert oversight, offering a scalable solution for landslide monitoring. Full article

Effective monitoring of fish passage through river barriers is essential for evaluating fishway performance and supporting adaptive river management. Traditional methods are often invasive, labor-intensive, or too costly to enable widespread implementation across most fishways. Infrared (IR) beam-break counters offer a promising alternative, but their adoption has been limited by high costs and a lack of flexibility. We developed and tested a novel, low-cost infrared beam-break counter—FishTracker—based on open-source Raspberry Pi and Arduino platforms. The system detects fish passages by analyzing interruptions in an IR curtain and reconstructing fish silhouettes to estimate movement, direction, speed, and morphometrics under a wide range of turbidity conditions. It also offers remote access capabilities for easy management. Field validation involved controlled tests with dummy fish, experiments with small-bodied live specimens (bleak) under varying turbidity conditions, and verification against synchronized video of free-swimming fish (koi carp). This first version of FishTracker achieved detection rates of 95–100% under controlled conditions and approximately 70% in semi-natural conditions, comparable to commercial counters. Most errors were due to surface distortion caused by partial submersion during the experimental setup, which could be avoided by fully submerging the device. Body length estimation based on passage speed and beam-interruption duration proved consistent, aligning with published allometric models for carps. FishTracker offers a promising and affordable solution for non-invasive fish monitoring in multispecies contexts. Its design, based primarily on open technology, allows for flexible adaptation and broad deployment, particularly in locations where commercial technologies are economically unfeasible. Full article