Sensors

27 pages, 731 KiB

Open AccessArticle

Detection and Localization of Rotor Winding Inter-Turn Short Circuit Fault in DFIG Using Zero-Sequence Current Component Under Variable Operating Conditions

by Muhammad Shahzad Aziz, Jianzhong Zhang, Sarvarbek Ruzimov and Xu Huang

Sensors 2025, 25(9), 2815; https://doi.org/10.3390/s25092815 (registering DOI) - 29 Apr 2025

Abstract

DFIG rotor windings face high stress and transients from back-to-back converters, causing inter-turn short circuit (ITSC) faults. Rapid rotor-side dynamics, combined with the unique capability of DFIG to operate in multiple modes, make the fault detection in rotor windings more challenging. This paper [...] Read more.

DFIG rotor windings face high stress and transients from back-to-back converters, causing inter-turn short circuit (ITSC) faults. Rapid rotor-side dynamics, combined with the unique capability of DFIG to operate in multiple modes, make the fault detection in rotor windings more challenging. This paper presents a comprehensive methodology for online ITSC fault diagnosis in DFIG rotor windings based on zero-sequence current (ZSC) component analysis under variable operating conditions. Fault features are identified and defined through the analytical evaluation of the DFIG mathematical model. Further, a simple yet effective algorithm is presented for online implementation of the proposed methodology. Finally, the simulation of the DFIG model is carried out in MATLAB/Simulink under both sub-synchronous and super-synchronous modes, covering a range of variable loads and low-frequency conditions, along with different fault severity levels of ITSC in rotor windings. Simulation results confirm the effectiveness of the proposed methodology for online ITSC fault detection at a low-severity stage and precise location identification of the faulty phase within the DFIG rotor windings under both sub-synchronous and super-synchronous modes. Full article

(This article belongs to the Section Intelligent Sensors)

14 pages, 3143 KiB

Open AccessArticle

A Capacitive Pressure Sensor with a Hierarchical Microporous Scaffold Prepared by Melt Near-Field Electro-Writing

by Zhong Zheng, Yifan Pan and Hao Huang

Sensors 2025, 25(9), 2814; https://doi.org/10.3390/s25092814 (registering DOI) - 29 Apr 2025

Abstract

Flexible capacitive pressure sensors (CPSs) have been widely studied and applied due to their various advantages. Numerous studies have been carried out on improving their electromechanical sensing properties through microporous structures. However, it is challenging to effectively control these structures. In this work, [...] Read more.

Flexible capacitive pressure sensors (CPSs) have been widely studied and applied due to their various advantages. Numerous studies have been carried out on improving their electromechanical sensing properties through microporous structures. However, it is challenging to effectively control these structures. In this work, we controllably fabricate a hierarchical microporous capacitive pressure sensor (HMCPS) using melt near-field electro-writing technology. Thanks to the hierarchical microporous sensor, which provides a multi-level elastic modulus and relative dielectric constants, the HMCPS shows outstanding sensing properties. Its multi-range pressure response is sensitive: 3.127 kPa⁻¹ at low pressure, 0.124 kPa⁻¹ at medium pressure, and 0.025 kPa⁻¹ at high pressure. Also, it has a stability of over 5000 cycles and a response time of less than 100 ms. The HMCPS can monitor dynamic and static pressures across a broad pressure range. It has been successfully applied to monitor human motions, showing great potential in human–computer interaction and smart wearable devices. Full article

(This article belongs to the Section Physical Sensors)

► Show Figures

Graphical abstract

19 pages, 22744 KiB

Open AccessArticle

Semantic-Aware Remote Sensing Change Detection with Multi-Scale Cross-Attention

by Xingjian Zheng, Xin Lin, Linbo Qing and Xianfeng Ou

Sensors 2025, 25(9), 2813; https://doi.org/10.3390/s25092813 (registering DOI) - 29 Apr 2025

Abstract

Remote sensing image change detection plays a vital role in diverse real-world applications such as urban development monitoring, disaster assessment, and land use analysis. As deep learning strives, Convolutional Neural Networks (CNNs) have shown their effects in image processing applications. There are two [...] Read more.

Remote sensing image change detection plays a vital role in diverse real-world applications such as urban development monitoring, disaster assessment, and land use analysis. As deep learning strives, Convolutional Neural Networks (CNNs) have shown their effects in image processing applications. There are two problems in old-school change detection techniques: First, the techniques do not fully use the effective information of the global and local features, which causes their semantic comprehension to be less accurate. Second, old-school methods usually simply rely on differences and computation at the pixel level without giving enough attention to the information at the semantic level. To address these problems, we propose a multi-scale cross-attention network (MSCANet) based on a CNN in this paper. First, a multi-scale feature extraction strategy is employed to capture and fuse image information across different spatial resolutions. Second, a cross-attention module is introduced to enhance the model’s ability to comprehend semantic-level changes between bitemporal images. Compared to the existing methods, our approach better integrates spatial and semantic features across scales, leading to more accurate and coherent change detection. Experiments on three public datasets (LEVIR-CD, CDD, and SYSU-CD) demonstrate competitive performance. For example, the model achieves an F1-score of 96.19% and an IoU of 92.67% on the CDD dataset. Additionally, robustness tests with Gaussian noise show that the model maintains high accuracy under input degradation, highlighting its potential for real-world applications. These findings suggest that our MSCANet effectively improves semantic awareness and robustness, offering a promising solution for change detection in complex and noisy remote sensing environments. Full article

(This article belongs to the Section Environmental Sensing)

20 pages, 8702 KiB

Open AccessArticle

Quantitative Prediction of Residual Stress, Surface Hardness, and Case Depth in Medium Carbon Steel Plate Based on Multifunctional Magnetic Testing Techniques

by Changjie Xu, Xianxian Wang, Haijiang Dong, Juanjuan Li, Liting Wang, Xiucheng Liu and Cunfu He

Sensors 2025, 25(9), 2812; https://doi.org/10.3390/s25092812 - 29 Apr 2025

Abstract

In this study, the methods of tangential magnetic field (TMF), magnetic Barkhausen noise (MBN), and incremental permeability (IP) were employed for in the simultaneous, quantitative prediction of target properties (bidirectional residual stress, surface hardness, and case depth) in the 45 steel plate. The [...] Read more.

In this study, the methods of tangential magnetic field (TMF), magnetic Barkhausen noise (MBN), and incremental permeability (IP) were employed for in the simultaneous, quantitative prediction of target properties (bidirectional residual stress, surface hardness, and case depth) in the 45 steel plate. The bidirectional magnetic signals and target properties were measured experimentally. The results of Pearson correlation analyses revealed that most parameters of the MBN and IP signals are strongly correlated with both residual stress and surface hardness under the influence of multiple target properties. The multiple linear regression (MLR) model demonstrated highly accurate quantitative prediction of residual stress and hardness in the y-direction. However, the simultaneous prediction of residual stress and case depth in the x-direction proved less effective than expected. To address this limitation, an inversion method was developed based on the regression model with the single parameter as the dependent variable and the target properties as the independent variable. By incorporating known magnetic parameters and target properties, the model effectively determined the unknown target properties. After applying the method, the coefficient of determination (R²) for x-direction residual stress increased from 0.89 to 0.96 and the absolute error (AE) of case depth decreased from 0.10 mm to 0.04 mm for case depths below 0.15. Full article

(This article belongs to the Section Physical Sensors)

► Show Figures

Figure 1

17 pages, 5448 KiB

Open AccessArticle

Deep Learning-Based Multimode Fiber Distributed Temperature Sensing

by Luxuan Yang, Xiaoyan Wang, Tong Wu, Huichuan Lin, Songjie Luo, Ziyang Chen, Yongxin Liu and Jixiong Pu

Sensors 2025, 25(9), 2811; https://doi.org/10.3390/s25092811 (registering DOI) - 29 Apr 2025

Abstract

As a laser beam passes through a multimode fiber (MMF), a speckle pattern is generated, which is sensitive to temperature, thereby making the MMF a temperature-sensing element. A deep learning technique is employed to the MMF-based temperature sensor, to obtain high-precision temperature sensing. [...] Read more.

As a laser beam passes through a multimode fiber (MMF), a speckle pattern is generated, which is sensitive to temperature, thereby making the MMF a temperature-sensing element. A deep learning technique is employed to the MMF-based temperature sensor, to obtain high-precision temperature sensing. We designed an MMF-based temperature-sensing configuration and developed a dual-output Convolutional Neural Network (CNN) for predicting both the temperature and the position of the heating point, and we constructed a dataset. It was shown that the location prediction accuracy reached 100%, while the temperature prediction accuracy (within a ±1 °C error margin) was 100% and 95.12% in the two experiments, respectively. The precision of the predicting heating point was less than 1 cm. Different types of MMFs were used in temperature measurements, showing that the accuracy remained quite high. This non-contact, high-precision MMF-based temperature measurement method, driven by deep learning, is suitable for applications in hazardous environments. Full article

(This article belongs to the Section Optical Sensors)

► Show Figures

Figure 1

20 pages, 3618 KiB

Open AccessArticle

Crowd Evacuation in Stadiums Using Fire Alarm Prediction

by Afnan A. Alazbah, Osama B. Rabie and Abdullah Al-Barakati

Sensors 2025, 25(9), 2810; https://doi.org/10.3390/s25092810 - 29 Apr 2025

Abstract

Ensuring rapid and efficient evacuation in high-density environments, such as stadiums, is critical for public safety during fire emergencies. Traditional fire alarm systems rely on reactive detection mechanisms, often resulting in delayed response times, increased panic, and overcrowding. This study introduces an AI-driven [...] Read more.

Ensuring rapid and efficient evacuation in high-density environments, such as stadiums, is critical for public safety during fire emergencies. Traditional fire alarm systems rely on reactive detection mechanisms, often resulting in delayed response times, increased panic, and overcrowding. This study introduces an AI-driven predictive fire alarm and evacuation model that leverages machine learning algorithms and real-time environmental sensor data to anticipate fire hazards before ignition, improving emergency response efficiency. To detect early fire risk indicators, the system processes data from 62,630 sensor measurements across 15 ecological parameters, including temperature, humidity, total volatile organic compounds (TVOC), CO₂ levels, and particulate matter. A comparative analysis of six machine learning models—Logistic Regression, Support Vector Machines (SVM), Random Forest, and proposed EvacuNet—demonstrates that EvacuNet outperforms all other models, achieving an accuracy of 99.99%, precision of 1.00, recall of 1.00, and an AUC-ROC score close to 1.00. The predictive alarm system significantly reduces false alarm rates and enhances fire detection speed, allowing emergency responders to take preemptive action. Moreover, integrating AI-driven evacuation optimization minimizes bottlenecks and congestion, reduces evacuation times, and improves structured crowd movement. These findings underscore the necessity of intelligent fire detection systems in high-occupancy venues, demonstrating that AI-based predictive modeling can drastically improve fire response and evacuation efficiency. Future research should focus on integrating IoT-enabled emergency navigation, reinforcement learning algorithms, and real-time crowd management systems to further enhance predictive accuracy and minimize casualties. By adopting such advanced technologies, large-scale venues can significantly improve emergency preparedness, reduce evacuation delays, and enhance public safety. Full article

(This article belongs to the Section Internet of Things)

► Show Figures

Figure 1

14 pages, 3160 KiB

Open AccessArticle

CVNet: Lightweight Cross-View Vehicle ReID with Multi-Scale Localization

by Wenji Yin, Baixuan Han, Yueping Peng, Hexiang Hao, Zecong Ye, Yu Shen, Yanjun Cai and Wenchao Kang

Sensors 2025, 25(9), 2809; https://doi.org/10.3390/s25092809 - 29 Apr 2025

Abstract

Cross-view vehicle re-identification (ReID) between aerial and ground perspectives is challenging due to limited computational resources on edge devices and significant scale variations. We propose CVNet, a lightweight network with two key modules: the multi-scale localization (MSL) module and the deep–shallow filtrate collaboration [...] Read more.

Cross-view vehicle re-identification (ReID) between aerial and ground perspectives is challenging due to limited computational resources on edge devices and significant scale variations. We propose CVNet, a lightweight network with two key modules: the multi-scale localization (MSL) module and the deep–shallow filtrate collaboration (DFC) module. The MSL module employs multi-scale depthwise separable convolutions and a localization attention mechanism to extract multi-scale features and localize salient regions, addressing viewpoint variations. DFC employs a dual-branch design comprising deep and shallow branches, integrating a filtration module optimized via neural architecture search, a collaboration module, and lightweight convolutions. This design effectively captures both unique and shared cross-view features, ensuring efficient and robust feature representation. We also release a new CVPair v1.0 dataset, the first benchmark for cross-view ReID, containing 14,969 images of 894 vehicle identities, offering results of traditional and lightweight methods. CVNet achieves state-of-the-art performance on CVPair v1.0, VehicleID, and VeRi776, advancing cross-view vehicle ReID. The dataset will be released publicly. Full article

(This article belongs to the Section Sensor Networks)

19 pages, 2499 KiB

Open AccessData Descriptor

SILF Dataset: Fault Dataset for Solar Insecticidal Lamp Internet of Things Node

by Xing Yang, Liyong Zhang, Lei Shu, Xiaoyuan Jing and Zhijun Zhang

Sensors 2025, 25(9), 2808; https://doi.org/10.3390/s25092808 - 29 Apr 2025

Abstract

Solar insecticidal lamps (SILs) are commonly used agricultural pest control devices that attract pests through a lure lamp and eliminate them using a high-voltage metal mesh. When integrated with Internet of Things (IoT) technology, SIL systems can collect various types of data, e.g., [...] Read more.

Solar insecticidal lamps (SILs) are commonly used agricultural pest control devices that attract pests through a lure lamp and eliminate them using a high-voltage metal mesh. When integrated with Internet of Things (IoT) technology, SIL systems can collect various types of data, e.g., pest kill counts, meteorological conditions, soil moisture levels, and equipment status. However, the proper functioning of SIL-IoT is a prerequisite for enabling these capabilities. Therefore, this paper introduces the component composition and fault analysis of SIL-IoT. By examining long-term operational data from seven nodes deployed in real-world scenarios, different fault modes are identified. Six typical machine methods are adopted to verify the validity of the proposed dataset. The results indicate that machine learning algorithms can achieve high accuracy on the proposed dataset. Notably, voltage, current, and meteorological data play a crucial role in the fault diagnosis process for both SIL-IoT and other related agricultural IoT devices. Full article

(This article belongs to the Section Cross Data)

► Show Figures

Figure 1

14 pages, 2978 KiB

Open AccessArticle

A Miniaturized FBG Tactile Sensor for the Tip of a Flexible Ureteroscope

by Shiyuan Dong, Sen Ma, Tenglong Zhou, Yuyang Lou, Xuanwei Xiong, Keyu Wei, Dong Luo, Jianwei Wu, Huanhuan Liu, Ran Tao, Tianyu Yang and Yuming Dong

Sensors 2025, 25(9), 2807; https://doi.org/10.3390/s25092807 - 29 Apr 2025

Abstract

This work introduces a novel fiber Bragg grating (FBG)-based tactile sensor specifically developed for real-time force monitoring at the tips of flexible ureteroscopes. With a diameter of only 1.5 mm, the sensor features a dual-FBG configuration that effectively separates temperature effects from force [...] Read more.

This work introduces a novel fiber Bragg grating (FBG)-based tactile sensor specifically developed for real-time force monitoring at the tips of flexible ureteroscopes. With a diameter of only 1.5 mm, the sensor features a dual-FBG configuration that effectively separates temperature effects from force signals, integrated with an innovative elastomer structure based on staggered parallelogram elements. Finite element analyses comparing traditional spiral and parallel groove designs indicate that the new configuration not only enhances axial sensitivity through optimized deformation characteristics but also significantly improves resistance to transverse forces via superior stress distribution and structural stability. In the sensor, a suspended lateral FBG is employed for thermal compensation, while an axially constrained FBG is dedicated to force detection. Calibration using a segmented approach yielded dual-range sensitivities of approximately 283.85 pm/N for the 0–0.5 N range and 258.57 pm/N for the 0.5–1 N range, with a maximum error of 0.07 N. Ex vivo ureteroscopy simulations further demonstrated the sensor’s capability to detect tissue–instrument interactions and to discriminate contact events effectively. This miniaturized solution offers a promising approach to achieving precise force feedback in endoscopic procedures while conforming to the dimensional constraints of standard ureteroscopes. Full article

(This article belongs to the Special Issue Recent Advances in Optoelectronic Materials and Device Engineering)

20 pages, 2340 KiB

Open AccessArticle

Modeling and Analysis of Mixed Traffic Flow Considering Driver Stochasticity and CAV Connectivity Uncertainty

by Qi Zeng, Siyuan Hao, Nale Zhao and Ruiche Liu

Sensors 2025, 25(9), 2806; https://doi.org/10.3390/s25092806 - 29 Apr 2025

Abstract

As connected and autonomous vehicle (CAV) technologies are rapidly integrated into modern transportation systems, understanding the dynamics of mixed traffic flow involving both human-driven vehicles (HVs) and CAVs is becoming increasingly important, particularly under uncertain conditions. In this paper, we propose a car-following [...] Read more.

As connected and autonomous vehicle (CAV) technologies are rapidly integrated into modern transportation systems, understanding the dynamics of mixed traffic flow involving both human-driven vehicles (HVs) and CAVs is becoming increasingly important, particularly under uncertain conditions. In this paper, we propose a car-following model framework to investigate the combined effects of driver stochasticity and connectivity uncertainties of CAVs on mixed traffic flow. The proposed framework can capture the inherent stochastic variations in human driving behavior by extending the classic intelligent driver model (IDM) with a Langevin-type stochastic differential equation. A car-following model with multi-anticipation control is developed for CAVs, explicitly incorporating sensor noise, communication delays, and dynamic connectivity. Extensive numerical simulations demonstrate that higher CAV penetration leads to more stable traffic flows. Even with certain levels of connectivity uncertainty, CAVs can still effectively stabilize the traffic. However, driver stochasticity has a pronounced impact on traffic stability—greater variability in driver behavior tends to reduce overall stability. Furthermore, sensitivity analyses reveal that in pure CAV environments, sensor noise, communication delays and communication ranges can affect traffic stability and energy consumption. In contrast, in mixed traffic conditions, the inherent instability of HV behavior tends to dominate and diminish the relative influence of CAV connectivity-related uncertainties. These findings underscore the necessity of robust sensor fusion and error compensation strategies to fully realize the potential of CAV technology. In mixed traffic environments, measures should be taken to minimize the adverse effects of HVs on CAV performance. Full article

(This article belongs to the Special Issue Intelligent Transportation Systems with Connected Vehicle, Cloud Computing, and Internet of Things Technologies)

► Show Figures

Figure 1

17 pages, 3128 KiB

Open AccessArticle

Transmission Raman Spectroscopy for Inner Layers Chemical Analysis of Fresh Produce

by Rani Arielly

Sensors 2025, 25(9), 2805; https://doi.org/10.3390/s25092805 - 29 Apr 2025

Abstract

Identification of chemical properties in the inner tissues of fresh produce would enable us to identify major issues plaguing the agriculture supply chain, like off-flavors and core rot since these are caused by or accompanied by known chemical elements. We show the development [...] Read more.

Identification of chemical properties in the inner tissues of fresh produce would enable us to identify major issues plaguing the agriculture supply chain, like off-flavors and core rot since these are caused by or accompanied by known chemical elements. We show the development of transmission Raman spectroscopy system for identifying these elements by addressing several issues: we located an optimal spectral window by conducting optical attenuation measurements and calculated the required LASER power in that range. For apple tissues, this optical window was found in the 700–950 nm range, and the required LASER power range was calculated to be in the 40–700 mW range. We also calculated that the optimal shifted-excitation Raman difference spectroscopy wavelengths should be separated by 0.7 nm in order to optimally produce narrow and high-intensity Raman peak features and eliminate the competing fluorescence signal. Finally, we provide a complete optical system design with exact optimal parameters. In contrast to other fields like pharmaceuticals and medicine, transmission Raman spectroscopy has not been applied extensively in agriculture. Therefore, this study fills a gap in that field’s applicability. Full article

(This article belongs to the Section Chemical Sensors)

► Show Figures

Figure 1

31 pages, 7561 KiB

Open AccessArticle

Centralized Measurement Level Fusion of GNSS and Inertial Sensors for Robust Positioning and Navigation

by Mohamed F. Elkhalea, Hossam Hendy, Ahmed Kamel, Ashraf Abosekeen and Aboelmagd Noureldin

Sensors 2025, 25(9), 2804; https://doi.org/10.3390/s25092804 - 29 Apr 2025

Abstract

In the current era, which is characterized by increasing demand for high-precision location and navigation capabilities, various industries, including those involved in intelligent vehicle systems, logistics, augmented reality, and more, heavily rely on accurate location information to optimize processes and deliver personalized experiences. [...] Read more.

In the current era, which is characterized by increasing demand for high-precision location and navigation capabilities, various industries, including those involved in intelligent vehicle systems, logistics, augmented reality, and more, heavily rely on accurate location information to optimize processes and deliver personalized experiences. In this context, the integration of Global Navigation Satellite System (GNSS) and inertial sensor technologies in smartphones has emerged as a critical solution to meet these demands. This research paper presents an algorithm that combines a GNSS with a modified downdate algorithm (MDDA) for satellite selection and integrates inertial navigation systems (INS) in both loosely and tightly coupled configurations. The primary objective was to harness the inherent strengths of these onboard sensors for navigation in challenging environments. These algorithms were meticulously designed to enhance performance and address the limitations encountered in harsh terrain. To evaluate the effectiveness of these proposed systems, vehicular experiments were conducted under diverse GNSS observation conditions. The experimental results clearly illustrate the considerable improvements achieved by the recommended tightly coupled (TC) algorithm when integrated with MDDA, in contrast to the loosely coupled (LC) algorithm. Specifically, the TC algorithm demonstrated a remarkable reduction of over 90% in 2D position root mean square error (RMSE) and a 75% reduction in 3D position RMSE when compared to solutions utilizing the weighting matrix provided by Google with all visible satellites. These findings underscore the substantial advancements in precision resulting from the integration of GNSS and INS technologies, thereby unlocking the full potential of transformative applications in the realm of intelligent vehicle navigation. Full article

(This article belongs to the Section Navigation and Positioning)

► Show Figures

Figure 1

16 pages, 1107 KiB

Open AccessReview

Potential for Wearable Sensor-Based Field-Deployable Diagnosis and Monitoring of Mild Traumatic Brain Injury: A Scoping Review

by Hope C. Davis-Wilson, Erika Maldonado-Rosado, Meghan Hegarty-Craver and Dorota S. Temple

Sensors 2025, 25(9), 2803; https://doi.org/10.3390/s25092803 - 29 Apr 2025

Abstract

Studies have shown that wearable commercial off-the-shelf sensors, such as accelerometers, inertial measurement units, and heart monitors, can distinguish between individuals with a mild traumatic brain injury (mTBI) and uninjured controls. However, there is no consensus on which metrics derived from wearable sensors [...] Read more.

Studies have shown that wearable commercial off-the-shelf sensors, such as accelerometers, inertial measurement units, and heart monitors, can distinguish between individuals with a mild traumatic brain injury (mTBI) and uninjured controls. However, there is no consensus on which metrics derived from wearable sensors are best to use for objective identification of mTBI symptoms. The primary aim of this scoping review was to map the state of knowledge of wearable sensor-based assessments for mTBI, based on previously published research. Data sources included Web of Science and PubMed. Original peer-reviewed articles were selected if mTBI was clinically diagnosed, an uninjured control cohort was included, and data collection used at least one digital sensor worn on the body. After screening 507 articles, 21 studies were included in the analysis. Overall, the studies identified multiple wearables-derived physiological metrics that differ between individuals with mTBI and uninjured controls. Some metrics associated with static balance, walking tasks, and postural changes to initiate an autonomic nervous system response were shown to support diagnosis of mTBI in retrospective studies with acceptable to outstanding accuracy. Further studies are needed to formulate standard protocols, reproduce results in large heterogeneous cohorts in prospective studies, and develop improved models that can diagnose mTBI with sufficient sensitivity and specificity in targeted populations. Full article

(This article belongs to the Special Issue Wearable Sensors for Movement, Postural Control and Locomotion Analysis)

► Show Figures

Figure 1

21 pages, 9171 KiB

Open AccessReview

Progress in Avalanche Photodiodes for Laser Ranging

by Zhenxing Liu, Ning An, Xingwei Han, Natalia Edith Nuñez, Liang Jin and Chengzhi Liu

Sensors 2025, 25(9), 2802; https://doi.org/10.3390/s25092802 - 29 Apr 2025

Abstract

Laser ranging is a high-precision geodetic technique that plays an indispensable role in the field of geodynamics. Avalanche photodiodes (APDs) offer a series of advantages over other photodetector technologies, including photomultiplier tubes (PMTs) and superconducting single-photon detectors (SNSPDs). These advantages include high sensitivity, [...] Read more.

Laser ranging is a high-precision geodetic technique that plays an indispensable role in the field of geodynamics. Avalanche photodiodes (APDs) offer a series of advantages over other photodetector technologies, including photomultiplier tubes (PMTs) and superconducting single-photon detectors (SNSPDs). These advantages include high sensitivity, small size, high integration, and low power consumption, which have contributed to the widespread use of APDs in laser ranging applications. This paper analyses the key role of APDs in enhancing the accuracy and stability of laser ranging through the examination of application examples, including Si-APD and InGaAs/InP APD. Finally, based on the technological needs of laser ranging, the future development directions of APDs are envisioned, aiming to provide a reference for the research of photodetectors in high-precision and high-frequency laser ranging applications. Full article

(This article belongs to the Section Electronic Sensors)

► Show Figures

Figure 1

30 pages, 7314 KiB

Open AccessArticle

Performance Evaluation of Fixed-Point Continuous Monitoring Systems: Influence of Averaging Time in Complex Emission Environments

by David Ball, Nathan Eichenlaub and Ali Lashgari

Sensors 2025, 25(9), 2801; https://doi.org/10.3390/s25092801 - 29 Apr 2025

Abstract

Quantifying methane emissions from facilities with complex emissions profiles can present a substantial challenge. Real-world emission scenarios can involve dynamic operational background emissions and temporally overlapping asynchronous emission events with varying rates from multiple sources. Previous studies have involved simpler testing setups, often [...] Read more.

Quantifying methane emissions from facilities with complex emissions profiles can present a substantial challenge. Real-world emission scenarios can involve dynamic operational background emissions and temporally overlapping asynchronous emission events with varying rates from multiple sources. Previous studies have involved simpler testing setups, often with synchronous emission sources and constant rates. This work is among the first to assess the performance of continuous monitoring systems (CMSs) under dynamic, overlapping emission scenarios with time-varying baselines. The data were collected as part of a novel single-blind controlled release study, where release sources and emission rates are not disclosed during the testing period. Several error metrics are defined and evaluated across a range of relevant averaging times, demonstrating that despite significant error variance in short-duration estimates, the low bias of the system results in substantially improved emission estimates when aggregated to longer timescales. Over the 4-week duration of this study, 700 kg of methane was released by the testing center, while the estimated quantity shows a final mass of 673 kg, an underestimation by 27 kg (4%). These results demonstrate that advanced CMSs can accurately quantify cumulative site-level emissions in complex scenarios, highlighting their potential for enhanced future emissions monitoring and regulatory applications in the oil and gas sector. Full article

(This article belongs to the Special Issue Gas Sensing for Air Quality Monitoring)

► Show Figures

Figure 1

19 pages, 2258 KiB

Open AccessArticle

A Multidimensional Particle Swarm Optimization-Based Algorithm for Brain MRI Tumor Segmentation

by Zsombor Boga, Csanád Sándor and Péter Kovács

Sensors 2025, 25(9), 2800; https://doi.org/10.3390/s25092800 - 29 Apr 2025

Abstract

Particle Swarm Optimization (PSO) has been extensively applied to optimization tasks in various domains, including image segmentation. In this work, we present a clustering-based segmentation algorithm that employs a multidimensional variant of PSO. Unlike conventional methods that require a predefined number of segments, [...] Read more.

Particle Swarm Optimization (PSO) has been extensively applied to optimization tasks in various domains, including image segmentation. In this work, we present a clustering-based segmentation algorithm that employs a multidimensional variant of PSO. Unlike conventional methods that require a predefined number of segments, our approach automatically selects an optimal segmentation granularity based on specified similarity criteria. This strategy effectively isolates brain tumors by incorporating both grayscale intensity and spatial information across multiple MRI modalities, allowing the method to be reliably tuned using a limited amount of training data. We further demonstrate how integrating these initial segmentations with a random forest classifier (RFC) enhances segmentation precision. Using MRI data from the RSNA-ASNR-MICCAI brain tumor segmentation (BraTS) challenge, our method achieves robust results with reduced reliance on extensive labeled datasets, offering a more efficient path toward accurate, clinically relevant tumor segmentation. Full article

(This article belongs to the Special Issue Sensors and Machine-Learning Based Signal Processing)

► Show Figures

Figure 1

19 pages, 6884 KiB

Open AccessArticle

Design of Computer Numerical Control System for Fiber Placement Machine Based on Siemens 840D sl

by Kun Xia, Di Zhao, Qingqing Yuan, Jingxia Wang and Aodong Shen

Sensors 2025, 25(9), 2799; https://doi.org/10.3390/s25092799 - 29 Apr 2025

Abstract

To address the manufacturing demands of large-scale aerospace composite components, this study systematically investigates the coordinated motion characteristics of multi-axis systems in fiber placement equipment. This investigation is based on the structural features and process specifications of the equipment. A comprehensive motion control [...] Read more.

To address the manufacturing demands of large-scale aerospace composite components, this study systematically investigates the coordinated motion characteristics of multi-axis systems in fiber placement equipment. This investigation is based on the structural features and process specifications of the equipment. A comprehensive motion control scheme for grid-based fiber placement machines was developed using the Siemens 840D CNC system, integrating filament-winding and tape-laying functionalities on a unified control platform while enabling 10-axis synchronous motion. To mitigate thermal-induced errors, a compensation method incorporating a BP neural network optimized by a genetic algorithm with an enhanced fitness function (GA-BP) was proposed. Experimental results demonstrate significant improvements: the maximum thermal errors of the Z-axis and X3-axis were reduced by 36.7% and 53.3%, respectively, while the core mold placement time was reduced to 61% of the specified duration, with notable enhancements in trajectory accuracy and processing efficiency. This research provides a technical framework for the design of multi-axis cooperative control systems and thermal error compensation in automated fiber placement equipment, offering critical insights for advancing manufacturing technologies in aerospace composite applications. The proposed methodology highlights practical value in balancing precision, efficiency, and system integration for complex composite component production. Full article

(This article belongs to the Section Sensor Materials)

► Show Figures

Figure 1

19 pages, 7282 KiB

Open AccessArticle

Mature Rice Biomass Estimation Using UAV-Derived RGB Vegetation Indices and Growth Parameters

by Mengguang Liao, Yun Wang, Nan Chu, Shaoning Li, Yifan Zhang and Dongfang Lin

Sensors 2025, 25(9), 2798; https://doi.org/10.3390/s25092798 - 29 Apr 2025

Abstract

The biomass of rice at maturity serves as a vital indicator for assessing overall productivity, and its accurate estimation holds significant importance for ensuring food security and promoting sustainable agriculture. To improve the precision of current biomass estimation methods for mature rice, this [...] Read more.

The biomass of rice at maturity serves as a vital indicator for assessing overall productivity, and its accurate estimation holds significant importance for ensuring food security and promoting sustainable agriculture. To improve the precision of current biomass estimation methods for mature rice, this study employed support vector regression to integrate RGB vegetation indices from rice canopy images with growth parameters, thereby developing a biomass estimation model. The model was validated by applying it to the experimental area. The results indicated that screening RGB vegetation indices and combining them with growth parameters enhanced estimation accuracy. Specifically, the model integrating RGB vegetation indices (g, RGBVI) with rice plant height and moisture content demonstrated high estimation accuracy (R² = 0.78, RMSE = 0.32 kg/m²). The absolute difference between the estimated and measured biomass values ranged from 0.15 to 0.39 kg/m². Additionally, the estimated biomass showed a strong correlation with yield (R² = 0.86), with a fitted equation of y = 0.04x + 0.59. These results suggest that the model is reliable for large-area estimation of mature rice biomass. However, the degree of rice maturity and the lodging phenomenon were identified as the primary factors influencing the precision of model application. Overall, integrating RGB vegetation indices of the rice canopy, obtained via UAV-based remote sensing technology, with growth parameters provides an effective method for estimating mature rice biomass and offers a valuable reference for efficient yield estimation. Full article

(This article belongs to the Section Smart Agriculture)

► Show Figures

Figure 1

21 pages, 10414 KiB

Open AccessArticle

SPL-PlaneTR: Lightweight and Generalizable Indoor Plane Segmentation Based on Prompt Learning

by Zhongchen Deng, Yuanlong Ge, Xiatian Qi, Kai Sun, Ruixi Wan, Bingxu Zhang, Shenman Zhang, Xun Zhang and Yan Meng

Sensors 2025, 25(9), 2797; https://doi.org/10.3390/s25092797 - 29 Apr 2025

Abstract

Single-image plane segmentation plays an important role in understanding 3D indoor scenes, including applications such as 3D indoor reconstruction. In recent years, PlaneTR, a transformer-based architecture, has achieved remarkable performance in single-image plane instance segmentation. It has garnered significant attention from researchers and [...] Read more.

Single-image plane segmentation plays an important role in understanding 3D indoor scenes, including applications such as 3D indoor reconstruction. In recent years, PlaneTR, a transformer-based architecture, has achieved remarkable performance in single-image plane instance segmentation. It has garnered significant attention from researchers and remains one of the most advanced algorithms in this field. However, PlaneTR has the following two major limitations: its ineffective utilization of line segment information within images and the high number of parameters. In this study, we propose an improved version of PlaneTR, named Spatial Prompt Learning PlaneTR (SPL-PlaneTR), to address these issues. Our approach effectively balances model complexity and performance. Specifically, to more effectively leverage structural information provided by line segments, we replace the original line segment’s transformer branch with a lightweight line segment prompt module and line segment prompt adapter. Additionally, we introduce spatial queries to replace conventional position queries, enabling the network to accurately localize planes across diverse indoor scenes. The experimental results demonstrate that our model, with fewer parameters, outperforms PlaneTR on both the original and noise-corrupted ScanNet datasets. Furthermore, SPL-PlaneTR achieves superior zero-shot transfer performance on the Matterport3D, ICL-NUIM RGB-D, and 2D-3D-S datasets compared to PlaneTR. Notably, our lightweight SPL-PlaneTR also surpasses several state-of-the-art algorithms in this domain. Our code and model have been publicly available. Full article

(This article belongs to the Section Sensing and Imaging)

► Show Figures

Figure 1

18 pages, 5643 KiB

Open AccessArticle

A New Hybrid Reinforcement Learning with Artificial Potential Field Method for UAV Target Search

by Fang Jin, Zhihao Ye, Mengxue Li, Han Xiao, Weiliang Zeng and Long Wen

Sensors 2025, 25(9), 2796; https://doi.org/10.3390/s25092796 - 29 Apr 2025

Abstract

Autonomous navigation and target search for unmanned aerial vehicles (UAVs) have extensive application potential in search and rescue, surveillance, and environmental monitoring. Reinforcement learning (RL) has demonstrated excellent performance in real-time UAV navigation through dynamic optimization of decision-making strategies, but its application in [...] Read more.

Autonomous navigation and target search for unmanned aerial vehicles (UAVs) have extensive application potential in search and rescue, surveillance, and environmental monitoring. Reinforcement learning (RL) has demonstrated excellent performance in real-time UAV navigation through dynamic optimization of decision-making strategies, but its application in large-scale environments for target search and obstacle avoidance is still limited by slow convergence and low computational efficiency. To address this issue, a hybrid framework combining RL and artificial potential field (APF) is proposed to improve the target search algorithm. Firstly, a task scenario and training environment for UAV target search are constructed. Secondly, RL is integrated with APF to form a framework that combines global and local strategies. Thirdly, the hybrid framework is compared with standalone RL algorithms through training and analysis of their performance differences. The experimental results demonstrate that the proposed method significantly outperforms standalone RL algorithms in terms of target search efficiency and obstacle avoidance performance. Specifically, the SAC-APF hybrid framework achieves a 161% improvement in success rate compared to the baseline SAC model, increasing from 0.282 to 0.736 in obstacle scenarios. Full article

(This article belongs to the Special Issue Artificial-Intelligence-Driven Intelligent Fault Prediction and Health Management Techniques in Manufacturing Systems: 2nd Edition)

► Show Figures

Figure 1

33 pages, 678 KiB

Open AccessReview

Internet of Medical Things Systems Review: Insights into Non-Functional Factors

by Giovanni Donato Gallo and Daniela Micucci

Sensors 2025, 25(9), 2795; https://doi.org/10.3390/s25092795 - 29 Apr 2025

Abstract

Internet of Medical Things (IoMT) is a rapidly evolving field with the potential to bring significant changes to healthcare. While several surveys have examined the structure and operation of these systems, critical aspects such as interoperability, sustainability, security, runtime self-adaptation [...] Read more.

Internet of Medical Things (IoMT) is a rapidly evolving field with the potential to bring significant changes to healthcare. While several surveys have examined the structure and operation of these systems, critical aspects such as interoperability, sustainability, security, runtime self-adaptation, and configurability are sometimes overlooked. Interoperability is essential for integrating data from various devices and platforms to provide a comprehensive view of a patient’s health. Sustainability addresses the environmental impact of IoMT technologies, crucial in the context of green computing. Security ensures the protection of sensitive patient data from breaches and manipulation. Runtime self-adaptation allows systems to adjust to changing patient conditions and environments. Configurability enables IoMT frameworks to monitor diverse patient conditions and manage different treatment paths. This article reviews current techniques addressing these aspects and highlights areas requiring further research. Full article

(This article belongs to the Section Biomedical Sensors)

► Show Figures

Figure 1

25 pages, 1405 KiB

Open AccessReview

A Survey of the Multi-Sensor Fusion Object Detection Task in Autonomous Driving

by Hai Wang, Junhao Liu, Haoran Dong and Zheng Shao

Sensors 2025, 25(9), 2794; https://doi.org/10.3390/s25092794 - 29 Apr 2025

Abstract

Multi-sensor fusion object detection is an advanced method that improves object recognition and tracking accuracy by integrating data from different types of sensors. As it can overcome the limitations of a single sensor in complex environments, the method has been widely applied in [...] Read more.

Multi-sensor fusion object detection is an advanced method that improves object recognition and tracking accuracy by integrating data from different types of sensors. As it can overcome the limitations of a single sensor in complex environments, the method has been widely applied in fields such as autonomous driving, intelligent monitoring, robot navigation, drone flight and so on. In the field of autonomous driving, multi-sensor fusion object detection has become a hot research topic. To further explore the future development trends of multi-sensor fusion object detection, we introduce the mainstream framework Transformer model of the multi-sensor fusion object detection algorithm, and we also provide a comprehensive summary of the feature fusion algorithms used in multi-sensor fusion object detection, specifically focusing on the fusion of camera and LiDAR data. This article provides an overview of feature fusion’s development into feature-level fusion and proposal-level fusion, and it specifically reviews multiple related algorithms. We discuss the application of current multi-sensor object detection algorithms. In the future, with the continuous advancement of sensor technology and the development of artificial intelligence algorithms, multi-sensor fusion object detection will show great potential in more fields. Full article

(This article belongs to the Special Issue Advanced Environment Perception, Decision Planning, and Intelligent Control of Transport Equipment)

► Show Figures

Figure 1

18 pages, 2786 KiB

Open AccessArticle

Clock Glitch Fault Attacks on Deep Neural Networks and Their Countermeasures

by Sangwon Lee, Suhyung Kim, Seongwoo Hong and Jaecheol Ha

Sensors 2025, 25(9), 2793; https://doi.org/10.3390/s25092793 - 29 Apr 2025

Abstract

Recently, deep neural networks (DNNs) have been widely used in various fields, such as autonomous vehicles and smart homes. Since these DNNs can be directly implemented on edge devices, they offer advantages such as real-time processing in low-power and low-bandwidth environments. However, the [...] Read more.

Recently, deep neural networks (DNNs) have been widely used in various fields, such as autonomous vehicles and smart homes. Since these DNNs can be directly implemented on edge devices, they offer advantages such as real-time processing in low-power and low-bandwidth environments. However, the deployment of DNNs in embedded systems, including edge devices, exposes them to threats such as fault injection attacks. This paper introduces a method of inducing misclassification using clock glitch fault attacks in devices where DNN models are executed. As a result of experiments on a microcontroller with a DNN implemented for two types of image classification (multi-class and binary classification using MNIST, CIFAR-10, and Kaggle datasets), we show that clock glitch fault attacks can lead—with a high probability—to the occurrence of serious misclassifications. Furthermore, we propose countermeasures to defeat the glitch attacks on each Softmax function and Sigmoid function at the algorithm level, and we confirm that these methods can effectively prevent misclassification incidents. Full article

(This article belongs to the Special Issue Internet of Things for Smart Homes (4th Edition))

► Show Figures

Figure 1

16 pages, 5226 KiB

Open AccessArticle

Enhanced Mask R-CNN Incorporating CBAM and Soft-NMS for Identification and Monitoring of Offshore Aquaculture Areas

by Jiajun Zhang, Yonggui Wang, Yaxin Zhang and Yanxin Zhao

Sensors 2025, 25(9), 2792; https://doi.org/10.3390/s25092792 - 29 Apr 2025

Abstract

The use of remote sensing images to analyze the change characteristics of large-scale aquaculture areas and monitor aquaculture violations is of great significance for exploring the law of marine aquaculture and assisting the monitoring and standardization of aquaculture areas. In this study, a [...] Read more.

The use of remote sensing images to analyze the change characteristics of large-scale aquaculture areas and monitor aquaculture violations is of great significance for exploring the law of marine aquaculture and assisting the monitoring and standardization of aquaculture areas. In this study, a violation monitoring framework for marine aquaculture areas based on image recognition using an enhanced Mask R-CNN architecture incorporating a convolutional block attention module (CBAM) and soft non-maximum suppression (Soft-NMS) is proposed and applied in Sandu’ao. The results show that the modified Mask R-CNN, when compared to the most basic Mask R-CNN model, exhibits higher accuracy in identifying marine aquaculture areas. The aquaculture patterns in the Xiapu region are characterized by two peak periods of aquaculture area fluctuations, occurring in March and October. Conversely, July marks the month with the smallest aquaculture area in the region and is influenced by factors such as water temperature and aquaculture cycle. Significant changes in the aquaculture area were observed in January, March, June, August, and October, necessitating rigorous monitoring. Furthermore, monitoring and analysis of aquaculture areas have revealed that despite the reduction in illegal aquaculture acreage since 2017 due to the implementation of functional zone planning for marine aquaculture areas, illegal aquaculture activities remain prevalent in prohibited and restricted zones in Xiapu, accounting for a considerable proportion. Full article

(This article belongs to the Section Smart Agriculture)

► Show Figures

Figure 1

8 pages, 3501 KiB

Open AccessCommunication

Angle-Engineered Bi_0.94La_0.06CuSeO Thin Films for High-Performance Transverse Thermoelectric Devices

by Mingjing Chen, Chenming Yue, Tianchang Qin, Haixu Liu, Guoying Yan and Shufang Wang

Sensors 2025, 25(9), 2791; https://doi.org/10.3390/s25092791 - 29 Apr 2025

Abstract

BiCuSeO has emerged as a highly promising material for transverse thermoelectric (TTE) applications, with its performance significantly enhanced through La doping. In this study, we investigate the effect of inclination angle on the TTE performance of inclined Bi_0.94La_0.06CuSeO thin [...] Read more.

BiCuSeO has emerged as a highly promising material for transverse thermoelectric (TTE) applications, with its performance significantly enhanced through La doping. In this study, we investigate the effect of inclination angle on the TTE performance of inclined Bi_0.94La_0.06CuSeO thin films fabricated using the pulsed laser deposition technique. A huge output voltage of 31.4 V was achieved in the 10° inclined Bi_0.94La_0.06CuSeO film under 308 nm ultraviolet pulsed laser irradiation. Furthermore, the films also exhibited significant response with excellent linearity when exposed to continuous-wave lasers across a broad spectral range (360 nm to 10,600 nm) and a point-like heat source. Notably, the voltage is directly proportional to sin2θ, where θ is the inclination angle. These findings not only provide a clear optimization strategy for TTE performance through inclination angle engineering but also highlight the material’s great potential for developing high-performance optical and thermal sensing TTE devices. Full article

(This article belongs to the Section Nanosensors)

► Show Figures

Figure 1

21 pages, 6300 KiB

Open AccessArticle

Electrospun (La,Ba)FeO₃ Nanofibers as Materials for Highly Sensitive VOC Gas Sensors

by Vadim Platonov, Nikolai Malinin, Darya Filatova, Ivan Sapkov and Marina Rumyantseva

Sensors 2025, 25(9), 2790; https://doi.org/10.3390/s25092790 - 28 Apr 2025

Abstract

In this work, we report the synthesis of perovskite-type Ba-doped LaFeO₃ (La_1−xBa_xFeO₃, x = 0.00, 0.02, 0.04, and 0.06) nanofibers (NFs) using the electrospinning method. The synthesized La_1−xBa_xFeO₃ materials have a [...] Read more.

In this work, we report the synthesis of perovskite-type Ba-doped LaFeO₃ (La_1−xBa_xFeO₃, x = 0.00, 0.02, 0.04, and 0.06) nanofibers (NFs) using the electrospinning method. The synthesized La_1−xBa_xFeO₃ materials have a fibrous structure with an average fiber diameter of 250 nm. The fibers, in turn, consist of smaller crystalline particles of 20–50 nm in size. The sensor properties of La_1−xBa_xFeO₃ nanofibers were studied when detecting 20 ppm CO, CH₄, methanol, and acetone in dry air in the temperature range of 50–350 °C. Doping with barium leads to a significant increase in sensor response and a decrease in operating temperature when detecting volatile organic compounds (VOCs). The process of acetone oxidation on the surface of the most sensitive La_0.98Ba_0.02FeO₃ material was studied using in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) and temperature-programmed desorption in combination with mass spectrometry (TPD-MS). A mechanism for the sensor signal formation is proposed. Full article

(This article belongs to the Special Issue Recent Advances in Sensors for Chemical Detection Applications)

► Show Figures

Figure 1

44 pages, 5982 KiB

Open AccessArticle

Adaptive Augmented Reality User Interfaces for Real-Time Defect Visualization and On-the-Fly Reconfiguration for Zero-Defect Manufacturing

by George Margetis, Katerina Valakou, Stavroula Ntoa, Despoina Gavgiotaki and Constantine Stephanidis

Sensors 2025, 25(9), 2789; https://doi.org/10.3390/s25092789 - 28 Apr 2025

Abstract

Zero-defect manufacturing is one of the most promising strategies to mitigate failures within manufacturing processes, allowing industries to increase product quality efficiently and effectively. One of the challenges faced in the practical adoption of zero-defect manufacturing is that the most important aspect of [...] Read more.

Zero-defect manufacturing is one of the most promising strategies to mitigate failures within manufacturing processes, allowing industries to increase product quality efficiently and effectively. One of the challenges faced in the practical adoption of zero-defect manufacturing is that the most important aspect of manufacturing, people, is often neglected. Aiming to support shop floor operators, this work introduces a human-centric approach assisting them to become aware of defects in the production line and imminently reconfigure it. Our system comprises an Augmented Reality application that encompasses interfaces that dynamically adapt to different contexts of use and enable operators to interact naturally and effectively and reconfigure the manufacturing process. The system leverages the efficiency of the shop floor operators in monitoring and controling the production line they are working on, according to the task they are performing, and their level of expertise, to produce appropriate visual components. To demonstrate the versatility and generality of the proposed system we evaluated it in three different production lines, conducting cognitive walkthroughs with experts and user-based evaluations with thirty shop floor operators. The results demonstrate that the system is intuitive and user-friendly, facilitating operator engagement and situational awareness, enhancing operator attentiveness, and achieving improved operational outcomes. Full article

(This article belongs to the Special Issue Intelligent Sensors and Signal Processing in Industry)

► Show Figures

Figure 1

21 pages, 2233 KiB

Open AccessArticle

Vehicle Recognition and Driving Information Detection with UAV Video Based on Improved YOLOv5-DeepSORT Algorithm

by Binshuang Zheng, Jing Zhou, Zhengqiang Hong, Junyao Tang and Xiaoming Huang

Sensors 2025, 25(9), 2788; https://doi.org/10.3390/s25092788 - 28 Apr 2025

Abstract

To investigate whether the skid resistance of the ramp meets the requirements of vehicle driving safety and stability, the simulation using the ideal driver model is inaccurate. Therefore, considering the driver’s driving habits, this paper proposes the use of Unmanned aerial vehicles (UAVs) [...] Read more.

To investigate whether the skid resistance of the ramp meets the requirements of vehicle driving safety and stability, the simulation using the ideal driver model is inaccurate. Therefore, considering the driver’s driving habits, this paper proposes the use of Unmanned aerial vehicles (UAVs) for the collection and extraction of vehicle driving information. To process the collected UAV video, the Google Collaboration platform is used to modify and compile the “You Only Look Once” version 5 (YOLOv5) algorithm with Python 3.7.12, and YOLOv5 is retrained with the captured video. The results show that the precision rate P and recall rate R have satisfactory results with an F1 value of 0.86, reflecting a good P-R relationship. The loss function also stabilized at a very low level after 70 training epochs. Then, the trained YOLOv5 is used to replace the Faster R-CNN detector in the DeepSORT algorithm to improve the detection accuracy and speed and extract the vehicle driving information from the perspective of UAV. By coding, the coordinate information of the vehicle trajectory is extracted, the trajectory is smoothed, and the frame difference method is used to calculate the real-time speed information, which is convenient for the establishment of a real driver model. Full article

(This article belongs to the Special Issue Intelligent Vehicle, Infrastructure Perception and Control Based on Imaging and Sensing)

17 pages, 453 KiB

Open AccessArticle

Online Meta-Recommendation of CUSUM Hyperparameters for Enhanced Drift Detection

by Jessica Fernandes Lopes, Sylvio Barbon Junior and Leonimer Flávio de Melo

Sensors 2025, 25(9), 2787; https://doi.org/10.3390/s25092787 - 28 Apr 2025

Viewed by 3

Abstract

With the increasing demand for time-series analysis, driven by the proliferation of IoT devices and real-time data-driven systems, detecting change points in time series has become critical for accurate short-term prediction. The variability in patterns necessitates frequent analysis to sustain high performance by [...] Read more.

With the increasing demand for time-series analysis, driven by the proliferation of IoT devices and real-time data-driven systems, detecting change points in time series has become critical for accurate short-term prediction. The variability in patterns necessitates frequent analysis to sustain high performance by acquiring the hyperparameter. The Cumulative Sum (CUSUM) method, based on calculating the cumulative values within a time series, is commonly used for change detection due to its early detection of small drifts, simplicity, low computational cost, and robustness to noise. However, its effectiveness heavily depends on the hyperparameter configuration, as a single setup may not be universally suitable across the entire time series. Consequently, fine-tuning is often required to achieve optimal results, yet this selection process is traditionally performed through trial and error or prior expert knowledge, which introduces subjectivity and inefficiency. To address this challenge, several strategies have been proposed to facilitate hyperparameter optimizations, as traditional methods are impractical. Meta-learning-based techniques present viable alternatives for periodic hyperparameter optimization, enabling the selection of configurations that adapt to dynamic scenarios. This work introduces a meta-modeling scheme designed to automate the recommendation of hyperparameters for the CUSUM algorithm. Benchmark datasets from the literature were used to evaluate the proposed framework. The results indicate that this framework preserves high accuracy while significantly reducing time requirements compared to Grid Search and Genetic Algorithm optimization. Full article

(This article belongs to the Section Internet of Things)

► Show Figures

Figure 1

29 pages, 3827 KiB

Open AccessArticle

Path Planning in Narrow Road Scenarios Based on Four-Layer Network Cost Structure Map

by Ping Wang, Hao Zhang and Youming Tang

Sensors 2025, 25(9), 2786; https://doi.org/10.3390/s25092786 - 28 Apr 2025

Viewed by 25

Abstract

To address the issues of insufficient safety distance and unsmooth paths in AGV path planning for narrow road scenarios, this paper proposes a method that integrates Voronoi-skeleton-based custom layers with traditional cost maps. First, key nodes of the Voronoi skeleton are extracted to [...] Read more.

To address the issues of insufficient safety distance and unsmooth paths in AGV path planning for narrow road scenarios, this paper proposes a method that integrates Voronoi-skeleton-based custom layers with traditional cost maps. First, key nodes of the Voronoi skeleton are extracted to generate a custom layer, which is then combined with static, obstacle, and expansion layers to form a new four-layer network cost map. This approach accurately distinguishes obstacle influences and enhances algorithm robustness. The A* algorithm based on this new map guides the automated guided vehicle (AGV) to travel safely along the road center. Second, an improved A* algorithm is employed for global planning to ensure safe navigation. Finally, B-spline smoothing is applied to the global path to enhance the AGV’s efficiency and stability in complex environments. The experimental results show that in narrow road scenarios, the proposed algorithm improves AGV path planning safety by 82%, reduces the number of spatial turning points by 55.85%, and shortens planning time by 48.98%. Overall, this algorithm significantly enhances the robustness and real-time performance of path planning in narrow roads, ensuring the AGV moves safely in an optimal manner. Full article

(This article belongs to the Section Sensing and Imaging)

Journal Menu

Journal Browser

Sensors, Volume 25, Issue 9 (May-1 2025) – 182 articles

Further Information

Guidelines

MDPI Initiatives

Follow MDPI