Sensors

15 pages, 3771 KB

Open AccessArticle

Early Motor Cortex Connectivity and Neuronal Reactivity in Intracerebral Hemorrhage: A Continuous-Wave Functional Near-Infrared Spectroscopy Study

by Nitin Kumar, Geetha Charan Duba, Nabeela Khan, Chetan Kashinkunti, Ashfaq Shuaib, Brian Buck and Mahesh Pundlik Kate

Sensors 2025, 25(20), 6377; https://doi.org/10.3390/s25206377 (registering DOI) - 15 Oct 2025

Abstract

Insights into motor cortex remodeling may enable the development of more effective rehabilitation strategies during the acute phase. We aim to assess the affected and unaffected motor/premotor/somatosensory cortex resting state functional connectivity (RSFC) and reactivity with continuous wave functional near-infrared spectroscopy (cw-fNIRS) in [...] Read more.

Insights into motor cortex remodeling may enable the development of more effective rehabilitation strategies during the acute phase. We aim to assess the affected and unaffected motor/premotor/somatosensory cortex resting state functional connectivity (RSFC) and reactivity with continuous wave functional near-infrared spectroscopy (cw-fNIRS) in patients with ICH compared to age, sex, and comorbidity-matched subjects. We enrolled patients with acute–subacute hemispheric ICH (n = 37; two were excluded due to artifacts) and grouped them according to the side (right and left) of the stroke. Matched participants or patients with recent transient ischemic attack were enrolled as control subjects for the study (n = 44; five were excluded due to artifacts). RSFC was assessed in both affected and unaffected hemispheres by group-level seed-based (primary motor cortex, priMC) correlation analysis. FT-associated relative oxyhemoglobin (ΔHbO) changes were analyzed in affected and unaffected hemispheres with generalized linear model regression. In left hemispheric ICH, the resting state coherence between the affected priMC and the affected premotor cortex (preMC) increased (β = 0.83, 95% CI = 0.19, 1.47, p = 0.01). In contrast, in right hemispheric ICH, the coherence between the unaffected priMC and the affected preMC decreased (β = −0.6, 95% CI = −1.12, −0.09, p = 0.02). In the left hemispheric ICH, the left-hand FT was associated with increased ΔHbO over the affected preMC (β = 0.01, 95% CI = 0.003, 0.02, p = 0.01). In contrast, in right hemispheric ICH, the left-hand FT was associated with increased ΔHbO over the unaffected preMC (β = 0.02, 95% CI = 0.006, 0.04, p = 0.01). Left hemispheric preMC may be involved in motor cortex reorganization in acute ICH in either hemisphere. Further studies may be required to assess longitudinal changes in motor cortex reorganization to inform acute motor rehabilitation. Full article

(This article belongs to the Special Issue Advances and Innovations in Optical Fiber Sensors)

► Show Figures

Figure 1

23 pages, 3462 KB

Open AccessArticle

Development and Modeling of a Novel Magnetorheological Elastomer Isolator in Hybrid Mode with a Compression–Shear Hybrid Fractional-Derivative Parametric Model

by Yun Tian, Zhongwei Hu, Yingqing Guo, Lihua Zhu, Jun Dai, Yuxuan Tao and Xin Wang

Sensors 2025, 25(20), 6376; https://doi.org/10.3390/s25206376 (registering DOI) - 15 Oct 2025

Abstract

Magnetorheological elastomers (MREs) are composed of soft silicone rubber, carbonyl iron particles (CIPs), and various additives. This study designs and fabricates a novel hybrid-mode MRE isolator that can operate in both compression and shear modes simultaneously. Experimental and modeling investigations are conducted to [...] Read more.

Magnetorheological elastomers (MREs) are composed of soft silicone rubber, carbonyl iron particles (CIPs), and various additives. This study designs and fabricates a novel hybrid-mode MRE isolator that can operate in both compression and shear modes simultaneously. Experimental and modeling investigations are conducted to examine the dynamic mechanical properties of the hybrid-mode MRE isolator under varying excitation frequencies, displacement amplitudes, and magnetic field strengths. The equivalent stiffness, energy dissipation, and equivalent damping of the MRE isolator are determined. Experimental results reveal that the hybrid-mode MRE isolator exhibits a pronounced MR effect by utilizing a hybrid magnetic field generation system, with all three parameters significantly increasing as the magnetic field strength increases. However, as the excitation frequency increases, the equivalent stiffness and energy dissipation increase, while the equivalent damping decreases. Based on the experimental findings, a compression–shear hybrid fractional-derivative parametric (CSHF) model is proposed to describe the impact of different operating conditions on the dynamic viscoelastic properties of the MRE isolator. A comparative analysis of the experimental results and model predictions indicates that the proposed mechanical model can accurately describe the dynamic mechanical characteristics of the hybrid-mode MRE isolator. Full article

(This article belongs to the Special Issue Structural Health Monitoring and Smart Disaster Prevention)

24 pages, 3843 KB

Open AccessArticle

Contributions to the Development of Fire Detection and Intervention Capabilities Using an Indoor Air Quality IoT Monitoring System

by Radu Nicolae Pietraru, Adriana Olteanu, Maximilian Nicolae and Robert-Alexandru Crăciun

Sensors 2025, 25(20), 6375; https://doi.org/10.3390/s25206375 (registering DOI) - 15 Oct 2025

Abstract

This paper presents a method for functionally extending an IoT indoor air quality monitoring network by adding a cloud-level fire detection logic component. The proposed method does not aim to replace traditional fire detection systems at this stage of research, but to propose [...] Read more.

This paper presents a method for functionally extending an IoT indoor air quality monitoring network by adding a cloud-level fire detection logic component. The proposed method does not aim to replace traditional fire detection systems at this stage of research, but to propose a solution for the development of fire detection capabilities and to improve the support provided to firefighting teams by providing a geospatial representation of the building in which a fire occurs. The proposed solution is based on a series of laboratory tests that demonstrated that air quality sensors can successfully detect the effects caused by an ignition event of common materials and can differentiate fire events from other events that can generate false-positive alarms by classic detection systems. The research involved five laboratory combustion tests based on the measurement of temperature, humidity, PM2.5 particle concentration, volatile organic compound index, and nitrogen oxide index. Following the tests, a warning mechanism and geospatial representation were designed using a system with ten IoT sensors to monitor the indoor air quality in a building on our university’s campus. Full article

(This article belongs to the Special Issue Sensors Solutions for the Development of Climate Neutral and Smart Cities)

► Show Figures

Figure 1

19 pages, 24197 KB

Open AccessArticle

Amplitude Normalization for Speed-Induced Modulation in Rotating Machinery Measurements

by Zhiwen Fang, Qing Zhang and Xinfa Shi

Sensors 2025, 25(20), 6374; https://doi.org/10.3390/s25206374 (registering DOI) - 15 Oct 2025

Abstract

Rotating machinery under variable-speed conditions suffers from amplitude modulation (AM) effects induced by speed fluctuations, complicating accurate fault detection. To address this issue, an amplitude normalization method based on support vector regression (SVR) is proposed to estimate and remove the AM effects. The [...] Read more.

Rotating machinery under variable-speed conditions suffers from amplitude modulation (AM) effects induced by speed fluctuations, complicating accurate fault detection. To address this issue, an amplitude normalization method based on support vector regression (SVR) is proposed to estimate and remove the AM effects. The method employs a correlation-based feature selection strategy to construct feature vectors strongly associated with rotational speed, thereby enabling the accurate quantification of speed-induced AM effects. The robust nonlinear fitting capability of SVR is then utilized to model and remove these effects, enhancing fault signal clarity. The proposed method is validated through two case studies and compared with advanced amplitude normalization techniques, demonstrating its superior accuracy, robustness, and reliability. Experimental results demonstrate that the proposed method accurately estimates and eliminates speed-induced AM, significantly improving fault diagnosis accuracy by up to 34.7%. Full article

(This article belongs to the Special Issue Smart Sensors for Machine Condition Monitoring and Fault Diagnosis)

► Show Figures

Figure 1

47 pages, 3335 KB

Open AccessSystematic Review

AI Video Analysis in Parkinson’s Disease: A Systematic Review of the Most Accurate Computer Vision Tools for Diagnosis, Symptom Monitoring, and Therapy Management

by Lazzaro di Biase, Pasquale Maria Pecoraro and Francesco Bugamelli

Sensors 2025, 25(20), 6373; https://doi.org/10.3390/s25206373 - 15 Oct 2025

Abstract

Background. Clinical assessment of Parkinson’s disease (PD) is limited by high subjectivity and inter-rater variability. Markerless video analysis, namely Computer Vision (CV), offers objective and scalable characterization of motor signs. We systematically reviewed CV technologies suited for PD diagnosis, symptom monitoring, and treatment [...] Read more.

Background. Clinical assessment of Parkinson’s disease (PD) is limited by high subjectivity and inter-rater variability. Markerless video analysis, namely Computer Vision (CV), offers objective and scalable characterization of motor signs. We systematically reviewed CV technologies suited for PD diagnosis, symptom monitoring, and treatment management. Methods. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, we searched PubMed for articles published between 1 January 1984 and 9 May 2025. We used the following search strategy: (“Parkinson Disease”[MeSH Terms] OR “parkinson’s disease” OR “parkinson disease”) AND (“computer vision” OR “video analysis” OR “pose estimation” OR “OpenPose” OR “DeepLabCut” OR “OpenFace” OR “YOLO” OR “MediaPipe” OR “markerless motion capture” OR “skeleton tracking”). Results. Out of 154 identified studies, 45 met eligibility criteria and were synthesized. Gait was assessed in 42% of studies, followed by bradykinesia items (17.7%). OpenPose and custom CV solutions were each used in 36% of studies, followed by MediaPipe (16%), DeepLabCut (9%), YOLO (4%). Across aims, CV pipelines consistently showed diagnostic discrimination and severity tracking aligned with expert ratings. Conclusions. CV non-invasively quantifies PD motor impairment, holding potential for objective diagnosis, longitudinal monitoring, and therapy response. Guidelines for standardized video-recording protocols and software usage are needed for real-world applications. Full article

(This article belongs to the Collection Sensors for Gait, Human Movement Analysis, and Health Monitoring)

24 pages, 2132 KB

Open AccessArticle

DL-AoD Estimation-Based 5G Positioning Using Directionally Transmitted Synchronization Signals

by Ivo Müürsepp and Muhammad Mahtab Alam

Sensors 2025, 25(20), 6372; https://doi.org/10.3390/s25206372 - 15 Oct 2025

Abstract

This paper introduces a method for estimating the Downlink Angle of Departure (DL-AoD) of 5G User Equipment (UE) from measured signal strengths of directionally transmitted synchronization signals. Based on estimated DL-AoD values, from two or more anchor nodes, the position of the UE [...] Read more.

This paper introduces a method for estimating the Downlink Angle of Departure (DL-AoD) of 5G User Equipment (UE) from measured signal strengths of directionally transmitted synchronization signals. Based on estimated DL-AoD values, from two or more anchor nodes, the position of the UE was estimated. Unlike most prior work, which is simulation-based or relies on custom testbeds, this study uses real measurements from an operational 5G network in an industrial factory environment. A deterministic estimator was derived, but multipath and unknown beam characteristics limit its accuracy. To address this, machine learning was applied to automatically adapt to the environment. Previous simulation studies reported 90th-percentile DL-AoD estimation errors below 2°, while experimental works achieved best-case accuracies of 5–6°. In this study, the experimental DL-AoD estimation error remained below 4° for 90% of the measurements, indicating improved real-world performance. Reported positioning errors in the literature range from 3.8 m to 140 m, whereas the 13.2 m error obtained here lies near the midpoint of this range, confirming the practicality of the proposed method in industrial environments. Compared to existing approaches, this work demonstrates high angular accuracy using only sub-6 GHz beams in a realistic industrial scenario without detailed knowledge of antenna beam patterns and channel state. The findings demonstrate that standard 5G signals can provide accurate indoor localization without additional infrastructure, offering a practical path toward cost-effective positioning in industrial IoT and automation. Full article

(This article belongs to the Special Issue Integrated Sensing and Communication in IoT Applications)

► Show Figures

Figure 1

23 pages, 4965 KB

Open AccessArticle

Direct Estimation of Electric Field Distribution in Circular ECT Sensors Using Graph Convolutional Networks

by Robert Banasiak, Zofia Stawska and Anna Fabijańska

Sensors 2025, 25(20), 6371; https://doi.org/10.3390/s25206371 - 15 Oct 2025

Abstract

The Electrical Capacitance Tomography (ECT) imaging pipeline relies on accurate estimation of electric field distributions to compute electrode capacitances and reconstruct permittivity maps. Traditional ECT forward model methods based on the Finite Element Method (FEM) offer high accuracy but are computationally intensive, limiting [...] Read more.

The Electrical Capacitance Tomography (ECT) imaging pipeline relies on accurate estimation of electric field distributions to compute electrode capacitances and reconstruct permittivity maps. Traditional ECT forward model methods based on the Finite Element Method (FEM) offer high accuracy but are computationally intensive, limiting their use in real-time applications. In this proof-of-concept study, we investigate the use of Graph Convolutional Networks (GCNs) for direct, one-step prediction of electric field distributions associated with a circular ECT sensor numerical model. The network is trained on FEM-simulated data and outputs of full 2D electric field maps for all excitation patterns. To evaluate physical fidelity, we compute capacitance matrices using both GCN-predicted and FEM-based fields. Our results show strong agreement in both direct field prediction and derived quantities, demonstrating the feasibility of replacing traditional solvers with fast, learned approximators. This approach has significant implications for further real-time ECT imaging and control applications. Full article

(This article belongs to the Section Sensing and Imaging)

► Show Figures

Figure 1

22 pages, 4901 KB

Open AccessArticle

Wind Speed Interval Prediction Based on Bayesian Optimized Spatio-Temporal Integration and Compression Deep Residual Network

by Yun Wu, Yongzhen Gong, Xiaoguo Chen, Xingang Wang and Xiaoyong Li

Sensors 2025, 25(20), 6370; https://doi.org/10.3390/s25206370 - 15 Oct 2025

Abstract

To address the challenge of high wind speed variability in wind farm planning, a small-sample-based spatio-temporal fusion and compression deep residual point prediction model, STiCDRS (Spatio-Temporal integration and Compression Deep Residual), is proposed. This model is designed to deeply explore the spatial and [...] Read more.

To address the challenge of high wind speed variability in wind farm planning, a small-sample-based spatio-temporal fusion and compression deep residual point prediction model, STiCDRS (Spatio-Temporal integration and Compression Deep Residual), is proposed. This model is designed to deeply explore the spatial and temporal characteristics within wind speed sequences to enhance the accuracy of point predictions. Initially, the spatio-temporal integration and compression deep residual network is employed to obtain point prediction results. Subsequently, an innovative hybrid model, STiCDRS-NKDE (STiCDRS-Nonparametric Kernel Density Estimation), is introduced to achieve interval predictions, thereby providing more reliable probabilistic forecasts of wind speed. The hyper-parameters of the model are optimized using Bayesian optimization, ensuring efficient and automated tuning. Finally, a case study involving wind speed forecasting at a wind farm in Inner Mongolia, China, is conducted, comparing the performance of the STiCDRS model with traditional models. Experimental results demonstrate that in comparison to other models, the proposed STiCDRS-NKDE model delivers superior point prediction accuracy, appropriate interval predictions, and reliable probabilistic forecasting outcomes, fully showcasing its significant potential in the domain of wind speed forecasting. Full article

(This article belongs to the Section Fault Diagnosis & Sensors)

► Show Figures

Figure 1

24 pages, 1084 KB

Open AccessArticle

Data-Driven Optimization of Healthcare Recommender System Retraining Pipelines in MLOps with Wearable IoT Data

by Yohan Park, Jonghyeok Mun, Yejung Lee, Jihwan Um, Jongsun Choi and Jaeyoung Choi

Sensors 2025, 25(20), 6369; https://doi.org/10.3390/s25206369 - 15 Oct 2025

Abstract

Personalized healthcare recommender systems are increasingly being deployed in edge AI environments through wearable devices. In such environments, cloud servers leverage high-performance GPUs to train base models, which are then optimized for data reduction deployment on edge devices, enabling the delivery of personalized [...] Read more.

Personalized healthcare recommender systems are increasingly being deployed in edge AI environments through wearable devices. In such environments, cloud servers leverage high-performance GPUs to train base models, which are then optimized for data reduction deployment on edge devices, enabling the delivery of personalized services. However, the base model may experience a gradual decline in accuracy over time, a phenomenon known as model drift. Recommender systems that do not keep up with changes in user preferences risk generating predictions based on outdated behavior, which can negatively impact the user experience. Therefore, it is essential to adopt retraining approaches that incorporate both past training data and new data from wearable devices. To address the drift problem, we propose a dynamic data management strategy, integrated into an automated training pipeline based on machine learning operations (MLOps). This approach enables adaptive model updates in response to continuously evolving IoT data. To preserve base model performance, our strategy leverages data reduction and feature selection algorithms. By dynamically managing data with these techniques, we effectively mitigate data drift and enhance resource efficiency during model retraining. We validated our approach through experiments on personalized fitness recommendations using FitRec wearable data from 1104 users, achieving improved computational efficiency during retraining while preserving model accuracy. Consequently, our dynamic data management method ensures faster training and the sustained performance of data reduction base models essential for edge AI applications. Moreover, this approach presents a compelling solution for continuously refining personalized recommendation services in alignment with evolving user preferences. Full article

(This article belongs to the Special Issue Operationalize Edge AI for Next-Generation IoT Applications)

26 pages, 965 KB

Open AccessReview

Sensing While Drilling and Intelligent Monitoring Technology: Research Progress and Application Prospects

by Xiaoyu Li, Zongwei Yao, Tao Zhang and Zhiyong Chang

Sensors 2025, 25(20), 6368; https://doi.org/10.3390/s25206368 - 15 Oct 2025

Abstract

Obtaining accurate information on stratigraphic conditions and drilling status is necessary to ensure the safety of the drilling process and to guarantee the production of oil and gas. Sensing while drilling and intelligent monitoring technology, which employ multiple sensors and involve the use [...] Read more.

Obtaining accurate information on stratigraphic conditions and drilling status is necessary to ensure the safety of the drilling process and to guarantee the production of oil and gas. Sensing while drilling and intelligent monitoring technology, which employ multiple sensors and involve the use of intelligent algorithms, can be used to collect downhole information in situ to ensure safe, reliable, and efficient drilling and mining operations. These approaches are characterized by effective sensing and comprehensive utilization of drilling information through the integration of multi-sensor signals and intelligent algorithms, a core component of machine learning. The article summarizes the current research status of domestic and international sensing while drilling and intelligent monitoring technology using systematically collected relevant information. Specifically, first, the drilling-sensing methods used for in situ acquisition of downhole information, including fiber-optic sensing, electronic-nose sensing, drilling engineering-parameter sensing, drilling mud-parameter sensing, drilling acoustic logging, drilling electromagnetic wave logging, and drilling seismic logging, are described. Next, the basic composition and development direction of each sensing technology are analyzed. Subsequently, the application of intelligent monitoring technology based on machine learning in various aspects of drilling- and mining-status identification, including bit wear monitoring, stuck drill real-time monitoring, well surge real-time monitoring, and real-time monitoring of oil and gas output, is introduced. Finally, the potential applications of sensing while drilling and intelligent monitoring technology in deep-earth, deep-sea, and deep-space contexts are discussed, and the challenges, constraints, and development trends are summarized. Full article

(This article belongs to the Topic Advances in Oil and Gas Wellbore Integrity, 2nd Edition)

22 pages, 4701 KB

Open AccessReview

The Status of Intelligent Control Technology for the Working Height of a Crop Harvesting Header

by Chenxu Zhao, Feng Wu, Fengwei Gu, Xinsheng Zhou, Yanqin Zhang, Peng Chen, Jiayong Pei and Hongguang Yang

Sensors 2025, 25(20), 6367; https://doi.org/10.3390/s25206367 - 15 Oct 2025

Abstract

As is well known, intelligence and efficiency are important development directions for modern agriculture. The harvesting header, as key components of crop harvesters, have significant implications for achieving intelligent control of their working height, which has a notable impact on reducing harvest loss. [...] Read more.

As is well known, intelligence and efficiency are important development directions for modern agriculture. The harvesting header, as key components of crop harvesters, have significant implications for achieving intelligent control of their working height, which has a notable impact on reducing harvest loss. To understand the current state of intelligent control technology for the working height of a crop harvesting header, and to explore their application potential, this article provides a relatively systematic literature review. Firstly, we analyzed the structure and principle of the harvesting header of typical grain and oil crops such as rice and peanuts. Secondly, we briefly described the current methods for controlling the working height of the harvesting header. They mainly use two methods: mechanical profiling and electro-hydraulic profiling. Thirdly, we focused on researching and analyzing the measurement methods and control algorithms for the working height of the harvesting header. Finally, we pointed out the problems in the current height control of the harvesting header. These problems mainly include insufficient measurement accuracy of working height in complex terrain, slow response and large delay of working height hydraulic control system, incompatibility between working height control models and strategies, and relatively single working height measurement methods. Full article

(This article belongs to the Section Smart Agriculture)

► Show Figures

Figure 1

21 pages, 10888 KB

Open AccessArticle

Analysis Method for the Spatial Layout Equilibrium of Highway Transportation Network Based on Community Detection

by Yuanyuan Zhang, Weidong Song, Jinguang Sun and Peng Dai

Sensors 2025, 25(20), 6366; https://doi.org/10.3390/s25206366 - 15 Oct 2025

Abstract

Analyzing the spatial layout equilibrium of highway transportation networks is essential for optimizing transportation networks, enhancing system efficiency and sustainability. To promote the equitable distribution and management of highway traffic resources, this study introduces a framework for assessing the spatial layout equilibrium of [...] Read more.

Analyzing the spatial layout equilibrium of highway transportation networks is essential for optimizing transportation networks, enhancing system efficiency and sustainability. To promote the equitable distribution and management of highway traffic resources, this study introduces a framework for assessing the spatial layout equilibrium of highway networks based on community structure. A new algorithm, named the C-Louvain algorithm, is introduced in this paper to address improving the stability of detection results in unconnected networks. The method first constructs a spatial node-based network, then detects the community structure of the highway network using the C-Louvain algorithm, and identifies key communities of the community structure network through a depth-first search. Network spatial layout imbalance is quantitatively assessed through supply–demand equilibrium analysis based on the Gini coefficient. This methodology is applied to the regional highway network in Shenyang, China. Results indicate that the C-Louvain method is optimal, excelling in accuracy, volatility, and efficiency compared to the classic FN, Leiden, and Louvain algorithms, providing a valuable contribution to the literature on graph clustering and data mining. There are significant differences in the number of communities within different connected components, which reflects the heterogeneity of the network’s structure. By this method, the imbalanced area in the highway transportation network layout is quickly found, and the equitable distribution of traffic resources is quantitatively evaluated. The research results can provide a theoretical basis for managers to make scientific investment decisions for road network construction. Full article

(This article belongs to the Section Intelligent Sensors)

► Show Figures

Figure 1

19 pages, 3060 KB

Open AccessArticle

Design and Optimization of a Hybrid Design for Quantum Transduction

by Enrico Bargagna, Julian Delgado, Changqing Wang, Ivan Gonin, Vyacheslav P. Yakovlev, Paolo Neri, Donato Passarelli and Silvia Zorzetti

Sensors 2025, 25(20), 6365; https://doi.org/10.3390/s25206365 - 15 Oct 2025

Abstract

This study presents the mechanical design and analysis of a quantum electro-optical transducer engineered to operate at millikelvin temperatures within a dilution refrigerator. The transducer enables bidirectional microwave-optical frequency conversion through a hybrid architecture that integrates a superconducting radiofrequency (SRF) cavity with an [...] Read more.

This study presents the mechanical design and analysis of a quantum electro-optical transducer engineered to operate at millikelvin temperatures within a dilution refrigerator. The transducer enables bidirectional microwave-optical frequency conversion through a hybrid architecture that integrates a superconducting radiofrequency (SRF) cavity with an electro-optic optical cavity. Among several design options investigated, the configuration offering the best thermal and mechanical performance was selected, yielding a robust solution with reduced sensitivity to fabrication tolerances, improved heat dissipation, as well as alignment precision. The design ensures uniform temperature distribution, enabling higher laser pump powers and, thus, increased conversion efficiency, while maintaining mechanical stresses safely below the material yield strength. Electromagnetic simulations further validate the design, demonstrating enhanced coupling between the optical and microwave modes, as well as a broader tuning range achieved with smaller tuner displacements. Full article

(This article belongs to the Section Physical Sensors)

► Show Figures

Figure 1

22 pages, 1095 KB

Open AccessArticle

Two-Color Fluorescence Thermometry Using Lock-in Amplifiers for Background Suppression

by Benjamin R. Anderson and Hergen Eilers

Sensors 2025, 25(20), 6364; https://doi.org/10.3390/s25206364 - 15 Oct 2025

Abstract

We present a comprehensive theoretical and experimental study of the lock-in amplifier detection technique applied to two-color fluorescence thermometry. We modeled and measured the impact of different lock-in amplifier parameters (reference frequency, time constant, roll-off) and background pulse characteristics (width and steepness) on [...] Read more.

We present a comprehensive theoretical and experimental study of the lock-in amplifier detection technique applied to two-color fluorescence thermometry. We modeled and measured the impact of different lock-in amplifier parameters (reference frequency, time constant, roll-off) and background pulse characteristics (width and steepness) on the lock-in amplifier’s background suppression, as well as modeled the impact of the lock-in amplifier parameters on temperature measurements of short duration heating events. Based on our results, we provide a guide for designing experiments using lock-in amplifiers in two-color fluorescence thermometry to obtain the best possible results. Full article

(This article belongs to the Special Issue Advances in Optical Fiber-Based Sensors)

► Show Figures

Figure 1

18 pages, 6196 KB

Open AccessArticle

MSIMG: A Density-Aware Multi-Channel Image Representation Method for Mass Spectrometry

by Fengyi Zhang, Boyong Gao, Yinchu Wang, Lin Guo, Wei Zhang and Xingchuang Xiong

Sensors 2025, 25(20), 6363; https://doi.org/10.3390/s25206363 - 15 Oct 2025

Abstract

Extracting key features for phenotype classification from high-dimensional and complex mass spectrometry (MS) data presents a significant challenge. Conventional data representation methods, such as traditional peak lists or grid-based imaging strategies, are often hampered by information loss and compromised signal integrity, thereby limiting [...] Read more.

Extracting key features for phenotype classification from high-dimensional and complex mass spectrometry (MS) data presents a significant challenge. Conventional data representation methods, such as traditional peak lists or grid-based imaging strategies, are often hampered by information loss and compromised signal integrity, thereby limiting the performance of downstream deep learning models. To address this issue, we propose a novel data representation framework named MSIMG. Inspired by object detection in computer vision, MSIMG introduces a data-driven, “density-peak-centric” patch selection strategy. This strategy employs density map estimation and non-maximum suppression algorithms to locate the centers of signal-dense regions, which serve as anchors for dynamic, content-aware patch extraction. This process transforms raw mass spectrometry data into a multi-channel image representation with higher information fidelity. Extensive experiments conducted on two public clinical mass spectrometry datasets demonstrate that MSIMG significantly outperforms both the traditional peak list method and the grid-based MetImage approach. This study confirms that the MSIMG framework, through its content-aware patch selection, provides a more information-dense and discriminative data representation paradigm for deep learning models. Our findings highlight the decisive impact of data representation on model performance and successfully demonstrate the immense potential of applying computer vision strategies to analytical chemistry data, paving the way for the development of more robust and precise clinical diagnostic models. Full article

(This article belongs to the Section Chemical Sensors)

► Show Figures

Figure 1

16 pages, 5754 KB

Open AccessArticle

PPG-Net 4: Deep-Learning-Based Approach for Classification of Blood Flow Using Non-Invasive Dual Photoplethysmography (PPG) Signals

by Manisha Samant and Utkarsha Pacharaney

Sensors 2025, 25(20), 6362; https://doi.org/10.3390/s25206362 - 15 Oct 2025

Abstract

Cardiovascular disease diagnosis heavily relies on accurate blood flow assessments, traditionally performed using invasive and often uncomfortable methods like catheterization. This research introduces PPG-Net 4, an innovative deep learning approach for non-invasive blood flow pattern classification using dual photoplethysmography (PPG) signals. By leveraging [...] Read more.

Cardiovascular disease diagnosis heavily relies on accurate blood flow assessments, traditionally performed using invasive and often uncomfortable methods like catheterization. This research introduces PPG-Net 4, an innovative deep learning approach for non-invasive blood flow pattern classification using dual photoplethysmography (PPG) signals. By leveraging advanced machine learning techniques, the proposed method addresses critical limitations in current diagnostic technologies. The study employed a novel dual-sensor arrangement capturing PPG signals from two body locations, generating a comprehensive dataset from 75 participants. Advanced signal processing techniques, including mel spectrogram generation and mel-frequency cepstral coefficient extraction, enabled sophisticated feature representation. The deep learning model, PPG-Net 4, demonstrated good capability at classifying the following five distinct blood flow patterns: laminar, turbulent, stagnant, pulsatile, and oscillatory. The experimental results revealed strong classification performance, with F1-scores ranging from 0.86 to 0.92 across different flow patterns. The highest accuracy was observed for pulsatile flow (F1-score: 0.92), underscoring the model’s precision and reliability. This approach not only provides a non-invasive alternative to traditional diagnostic methods but also offers a potentially useful technique for early cardiovascular disease detection and continuous monitoring. Full article

(This article belongs to the Section Optical Sensors)

► Show Figures

Figure 1

21 pages, 3332 KB

Open AccessFeature PaperArticle

Intelligent Classification of Urban Noise Sources Using TinyML: Towards Efficient Noise Management in Smart Cities

by Maykol Sneyder Remolina Soto, Brian Amaya Guzmán, Pedro Antonio Aya-Parra, Oscar J. Perdomo, Mauricio Becerra-Fernandez and Jefferson Sarmiento-Rojas

Sensors 2025, 25(20), 6361; https://doi.org/10.3390/s25206361 - 14 Oct 2025

Abstract

Urban noise levels that exceed the World Health Organization (WHO) recommendations have become a growing concern due to their adverse effects on public health. In Bogotá, Colombia, studies by the District Department of Environment (SDA) indicate that 11.8% of the population is exposed [...] Read more.

Urban noise levels that exceed the World Health Organization (WHO) recommendations have become a growing concern due to their adverse effects on public health. In Bogotá, Colombia, studies by the District Department of Environment (SDA) indicate that 11.8% of the population is exposed to noise levels above the WHO limits. This research aims to identify and categorize environmental noise sources in real time using an embedded intelligent system. A total of 657 labeled audio clips were collected across eight classes and processed using a 60/20/20 train–validation–test split, ensuring that audio segments from the same continuous recording were not mixed across subsets. The system was implemented on a Raspberry Pi 2W equipped with a UMIK-1 microphone and powered by a 90 W solar panel with a 12 V battery, enabling autonomous operation. The TinyML-based model achieved precision and recall values between 0.92 and 1.00, demonstrating high performance under real urban conditions. Heavy vehicles and motorcycles accounted for the largest proportion of classified samples. Although airplane-related events were less frequent, they reached maximum sound levels of up to 88.4 dB(A), exceeding the applicable local limit of 70 dB(A) by approximately 18 dB(A) rather than by percentage. In conclusion, the results demonstrate that on-device TinyML classification is a feasible and effective strategy for urban noise monitoring. Local inference reduces latency, bandwidth usage, and privacy risks by eliminating the need to transmit raw audio to external servers. This approach provides a scalable and sustainable foundation for noise management in smart cities and supports evidence-based public policies aimed at improving urban well-being. This work presents an introductory and exploratory study on the application of TinyML for acoustic environmental monitoring, aiming to evaluate its feasibility and potential for large-scale implementation. Full article

(This article belongs to the Section Environmental Sensing)

► Show Figures

Figure 1

18 pages, 18957 KB

Open AccessArticle

Multi-Modal Data Fusion for 3D Object Detection Using Dual-Attention Mechanism

by Mengying Han, Benlan Shen and Jiuhong Ruan

Sensors 2025, 25(20), 6360; https://doi.org/10.3390/s25206360 - 14 Oct 2025

Abstract

To address the issue of missing feature information for small objects caused by the sparsity and irregularity of point clouds, as well as the poor detection performance on small objects due to their weak feature representation, this paper proposes a multi-modal 3D object [...] Read more.

To address the issue of missing feature information for small objects caused by the sparsity and irregularity of point clouds, as well as the poor detection performance on small objects due to their weak feature representation, this paper proposes a multi-modal 3D object detection method based on an improved PointPillars framework. First, LiDAR point clouds are fused with camera images at the data level, incorporating 2D semantic information to enhance small-object feature representation. Second, a Pillar-wise Channel Attention (PCA) module is introduced to emphasize critical features before converting pillar features into pseudo-image representations. Additionally, a Spatial Attention Module (SAM) is embedded into the backbone network to enhance spatial feature representation. Experiments on the KITTI dataset show that, compared with the baseline PointPillars, the proposed method significantly improves small-object detection performance. Specifically, under the bird’s-eye view (BEV) evaluation metrics, the Average Precision (AP) for pedestrians and cyclists increases by 7.06% and 3.08%, respectively; under the 3D evaluation metrics, these improvements are 4.36% and 2.58%. Compared with existing methods, the improved model also achieves relatively higher accuracy in detecting small objects. Visualization results further demonstrate the enhanced detection capability of the proposed method for small objects with different difficulty levels. Overall, the proposed approach effectively improves 3D object detection performance, particularly for small objects, in complex autonomous driving scenarios. Full article

(This article belongs to the Section Sensing and Imaging)

► Show Figures

Figure 1

23 pages, 2255 KB

Open AccessArticle

Design and Implementation of a YOLOv2 Accelerator on a Zynq-7000 FPGA

by Huimin Kim and Tae-Kyoung Kim

Sensors 2025, 25(20), 6359; https://doi.org/10.3390/s25206359 - 14 Oct 2025

Abstract

You Only Look Once (YOLO) is a convolutional neural network-based object detection algorithm widely used in real-time vision applications. However, its high computational demand leads to significant power consumption and cost when deployed in graphics processing units. Field-programmable gate arrays offer a low-power [...] Read more.

You Only Look Once (YOLO) is a convolutional neural network-based object detection algorithm widely used in real-time vision applications. However, its high computational demand leads to significant power consumption and cost when deployed in graphics processing units. Field-programmable gate arrays offer a low-power alternative. However, their efficient implementation requires architecture-level optimization tailored to limited device resources. This study presents an optimized YOLOv2 accelerator for the Zynq-7000 system-on-chip (SoC). The design employs 16-bit integer quantization, a filter reuse structure, an input feature map reuse scheme using a line buffer, and tiling parameter optimization for the convolution and max pooling layers to maximize resource efficiency. In addition, a stall-based control mechanism is introduced to prevent structural hazards in the pipeline. The proposed accelerator was implemented on the Zynq-7000 SoC board, and a system-level evaluation confirmed a negligible accuracy drop of only 0.2% compared with the 32-bit floating-point baseline. Compared with previous YOLO accelerators on the same SoC, the design achieved up to 26% and 15% reductions in flip-flop and digital signal processor usage, respectively. This result demonstrates feasible deployment on XC7Z020 with DSP 57.27% and FF 16.55% utilization. Full article

(This article belongs to the Special Issue Object Detection and Recognition Based on Deep Learning)

► Show Figures

Figure 1

Journal Description

Sensors

Latest Articles

Journal Menu

Journal Browser

Highly Accessed Articles

Latest Books

E-Mail Alert

News

Topics

Conferences

Special Issues

Topical Collections

Further Information

Guidelines

MDPI Initiatives

Follow MDPI