Due to the diversified demands of quality of service（QoS） in volume multimedia application, QoS routings for multiservice are becoming a research hotspot in low earth orbit（LEO） satellite networks. A novel QoS sat...Due to the diversified demands of quality of service（QoS） in volume multimedia application, QoS routings for multiservice are becoming a research hotspot in low earth orbit（LEO） satellite networks. A novel QoS satellite routing algorithm for multi-class traffic is proposed. The goal of the routing algorithm is to provide the distinct QoS for different traffic classes and improve the utilization of network resources. Traffic is classified into three classes and allocated priorities based on their QoS requirements, respectively. A priority queuing mechanism guarantees the algorithm to work better for high-priority classes. In order to control the congestion, a blocking probability analysis model is built up based on the Markov process theory. Finally, according to the classification link-cost metrics, routings for different classes are calculated with the distinct QoS requirments and the status of network resource. Simulations verify the performance of the routing algorithm at different time and in different regions, and results demonstrate that the algorithm has great advantages in terms of the average delay and the blocking probability. Meanwhile, the robustness issue is also discussed.展开更多
A strict proof of the hyperbolicity of the multi-class LWR ( Lighthill-Whitham-Richards) traffic flow model, as well as the descriptions on those nonlinear waves characterized in the traffic flow problems were given. ...A strict proof of the hyperbolicity of the multi-class LWR ( Lighthill-Whitham-Richards) traffic flow model, as well as the descriptions on those nonlinear waves characterized in the traffic flow problems were given. They were mainly about the monotonicity of densities across shocks and in rarefactions. As the system had no characteristic decomposition explicitly, a high resolution and higher order accuracy WENO( weighted essentially non-oscillatory) scheme was introduced to the numerical simulation, which coincides with the analytical description.展开更多
Autonomous vehicles operate without direct human intervention,which introduces safety risks that differ from those of conventional vehicles.Although many studies have examined safety issues related to autonomous drivi...Autonomous vehicles operate without direct human intervention,which introduces safety risks that differ from those of conventional vehicles.Although many studies have examined safety issues related to autonomous driving,high-risk situations have often been defined using single indicators,making it difficult to capture the complex and evolving nature of accident risk.To address this limitation,this study proposes a structured framework for defining and analyzing high-risk situations throughout the traffic accident process.High-risk situations are described using three complementary indicators:accident likelihood,accident severity,and accident duration.These indicators explain how risk emerges,increases,and persists over time.Based on this concept,a framework for traffic accident visualization analysis is developed to support phase-specific risk assessment and visualization.The framework combines accident-phase information with factor-level risk contributions,allowing systematic identification of key factors and their interactions across different accident stages.Using combinations of the three indicators,high-risk situations are classified into twenty-seven distinct types,providing a clear typology for complex accident scenarios involving autonomous vehicles.The applicability of the proposed framework is demonstrated through two representative accident scenarioswith different risk characteristics.The results showthat the framework effectively captures interactions among multiple risk factors,explains how risk levels change from pre-crash to post-crash phases,and identifies contributing factors that are difficult to detect using conventional traffic accident investigation methods.Overall,the proposed framework offers a practical basis for autonomous vehicle accident analysis,safety evaluation,and policy-related decision-making.展开更多
As a fundamental component in computer vision,edges can be categorized into four types based on discontinuities in reflectance,illumination,surface normal,or depth.While deep CNNs have significantly advanced generic e...As a fundamental component in computer vision,edges can be categorized into four types based on discontinuities in reflectance,illumination,surface normal,or depth.While deep CNNs have significantly advanced generic edge detection,real-time multi-class semantic edge detection under resource constraints remains challenging.To address this,we propose a lightweight framework based on PiDiNet that enables fine-grained semantic edge detection.Our model simultaneously predicts background and four edge categories from full-resolution inputs,balancing accuracy and efficiency.Key contributions include:a multi-channel output structure expanding binary edge prediction to five classes,supported by a deep supervision mechanism;a dynamic class-balancing strategy combining adaptive weighting with physical priors to handle extreme class imbalance;and maintained architectural efficiency enabling real-time inference.Extensive evaluations on BSDS-RIND show our approach achieves accuracy competitive with state-of-the-art methods while operating in real time.展开更多
With the evolution of next-generation network technologies,the complexity of network management has significantly increased,and the means of network attacks are diversified,bringing new challenges to network traffic c...With the evolution of next-generation network technologies,the complexity of network management has significantly increased,and the means of network attacks are diversified,bringing new challenges to network traffic classification.This paper presents a general AIdriven network traffic classification workflow and elaborates on a traffic data and feature engineering framework.Most importantly,it analyzes the concept and causes of data distribution shifts in ne twork traffic,proposing detection methods and countermeasures.Experimental results on real traffic collected at different time intervals show that application evolution can induce data distribution shifts,which in turn lead to a noticeable degradation in traffic classification performance.Comparative drift detection experiments further confirm that such shifts are more evident over long-term intervals,while short-term traffic remains relatively stable.These findings demonstrate the necessity of incorporating drift-aware mechanisms into AI-driven network traffic classification systems.展开更多
With the continuous progress of automatic driving technology,automatic driving technology standards are gradually affecting the determination of criminal responsibility for traffic accidents in China.At present,the ch...With the continuous progress of automatic driving technology,automatic driving technology standards are gradually affecting the determination of criminal responsibility for traffic accidents in China.At present,the characteristics and tendency of China's automatic driving technology standards present the situation of high policy relevance coexisting with low normative binding,professionalism coexist with barriers,forefront coexist with ambiguity.Therefore,challenges are presented both theoretically and practically on the determination of criminal responsibility based on automatic driving technology standard..In this regard,the misunderstanding should be clarified in theory:The legal order under the automatic driving technology standard has constitutionality and systematic,and there is a balance between the frontier of automatic driving technology development and the lagging of criminal law.The automatic driving technology risk level system should be built to clarify the boundary of the effectiveness of criminal law norms,seeking fora breakthrough in the application of the establishment of a comprehensive judgment system of the risks and accidents and the system of evidence to prove the system,which clarifies the determination of criminal responsibility under the automatic driving technology standard.This essay hopes to pursue breakthroughs in the application-to establish a comprehensive judgment system of risks and accidents as well as an evidence proof system,so as to clarify the determination of criminal responsibility under automatic driving technology standards.展开更多
This paper investigates the traffic offloading optimization challenge in Space-Air-Ground Integrated Networks(SAGIN)through a novel Recursive Multi-Agent Proximal Policy Optimization(RMAPPO)algorithm.The exponential g...This paper investigates the traffic offloading optimization challenge in Space-Air-Ground Integrated Networks(SAGIN)through a novel Recursive Multi-Agent Proximal Policy Optimization(RMAPPO)algorithm.The exponential growth of mobile devices and data traffic has substantially increased network congestion,particularly in urban areas and regions with limited terrestrial infrastructure.Our approach jointly optimizes unmanned aerial vehicle(UAV)trajectories and satellite-assisted offloading strategies to simultaneously maximize data throughput,minimize energy consumption,and maintain equitable resource distribution.The proposed RMAPPO framework incorporates recurrent neural networks(RNNs)to model temporal dependencies in UAV mobility patterns and utilizes a decentralized multi-agent reinforcement learning architecture to reduce communication overhead while improving system robustness.The proposed RMAPPO algorithm was evaluated through simulation experiments,with the results indicating that it significantly enhances the cumulative traffic offloading rate of nodes and reduces the energy consumption of UAVs.展开更多
Urban traffic generates massive and diverse data,yet most systems remain fragmented.Current approaches to congestion management suffer from weak data consistency and poor scalability.This study addresses this gap by p...Urban traffic generates massive and diverse data,yet most systems remain fragmented.Current approaches to congestion management suffer from weak data consistency and poor scalability.This study addresses this gap by proposing the Urban Traffic Congestion Unified Metadata Model(UTC-UMM).The goal is to provide a standardized and extensible framework for describing,extracting,and storing multisource traffic data in smart cities.The model defines a two-tier specification that organizes nine core traffic resource classes.It employs an eXtensible Markup Language(XML)Schema that connects general elements with resource-specific elements.This design ensures both syntactic and semantic interoperability across siloed datasets.Extension principles allow new elements or constraints to be introducedwithout breaking backward compatibility.Adistributed pipeline is implemented usingHadoop Distributed File System(HDFS)and HBase.It integrates computer vision for video and natural language processing for text to automate metadata extraction.Optimized row-key designs enable low-latency queries.Performance is tested with the Yahoo!Cloud Serving Benchmark(YCSB),which shows linear scalability and high throughput.The results demonstrate that UTC-UMM can unify heterogeneous traffic data while supporting real-time analytics.The discussion highlights its potential to improve data reuse,portability,and scalability in urban congestion studies.Future research will explore integration with association rulemining and advanced knowledge representation to capture richer spatiotemporal traffic patterns.展开更多
Traffic sign detection is an important part of autonomous driving,and its recognition accuracy and speed are directly related to road traffic safety.Although convolutional neural networks(CNNs)have made certain breakt...Traffic sign detection is an important part of autonomous driving,and its recognition accuracy and speed are directly related to road traffic safety.Although convolutional neural networks(CNNs)have made certain breakthroughs in this field,in the face of complex scenes,such as image blur and target occlusion,the traffic sign detection continues to exhibit limited accuracy,accompanied by false positives and missed detections.To address the above problems,a traffic sign detection algorithm,You Only Look Once-based Skip Dynamic Way(YOLO-SDW)based on You Only Look Once version 8 small(YOLOv8s),is proposed.Firstly,a Skip Connection Reconstruction(SCR)module is introduced to efficiently integrate fine-grained feature information and enhance the detection accuracy of the algorithm in complex scenes.Secondly,a C2f module based on Dynamic Snake Convolution(C2f-DySnake)is proposed to dynamically adjust the receptive field information,improve the algorithm’s feature extraction ability for blurred or occluded targets,and reduce the occurrence of false detections and missed detections.Finally,the Wise Powerful IoU v2(WPIoUv2)loss function is proposed to further improve the detection accuracy of the algorithm.Experimental results show that the average precision mAP@0.5 of YOLO-SDW on the TT100K dataset is 89.2%,and mAP@0.5:0.95 is 68.5%,which is 4%and 3.3%higher than the YOLOv8s baseline,respectively.YOLO-SDW ensures real-time performance while having higher accuracy.展开更多
With network attack technology continuing to develop,traditional anomaly traffic detection methods that rely on feature engineering are increasingly insufficient in efficiency and accuracy.Graph Neural Network(GNN),a ...With network attack technology continuing to develop,traditional anomaly traffic detection methods that rely on feature engineering are increasingly insufficient in efficiency and accuracy.Graph Neural Network(GNN),a promising Deep Learning(DL)approach,has proven to be highly effective in identifying intricate patterns in graph⁃structured data and has already found wide applications in the field of network security.In this paper,we propose a hybrid Graph Convolutional Network(GCN)⁃GraphSAGE model for Anomaly Traffic Detection,namely HGS⁃ATD,which aims to improve the accuracy of anomaly traffic detection by leveraging edge feature learning to better capture the relationships between network entities.We validate the HGS⁃ATD model on four publicly available datasets,including NF⁃UNSW⁃NB15⁃v2.The experimental results show that the enhanced hybrid model is 5.71%to 10.25%higher than the baseline model in terms of accuracy,and the F1⁃score is 5.53%to 11.63%higher than the baseline model,proving that the model can effectively distinguish normal traffic from attack traffic and accurately classify various types of attacks.展开更多
This paper presents a unified Unmanned Aerial Vehicle-based(UAV-based)traffic monitoring framework that integrates vehicle detection,tracking,counting,motion prediction,and classification in a modular and co-optimized...This paper presents a unified Unmanned Aerial Vehicle-based(UAV-based)traffic monitoring framework that integrates vehicle detection,tracking,counting,motion prediction,and classification in a modular and co-optimized pipeline.Unlike prior works that address these tasks in isolation,our approach combines You Only Look Once(YOLO)v10 detection,ByteTrack tracking,optical-flow density estimation,Long Short-Term Memory-based(LSTM-based)trajectory forecasting,and hybrid Speeded-Up Robust Feature(SURF)+Gray-Level Co-occurrence Matrix(GLCM)feature engineering with VGG16 classification.Upon the validation across datasets(UAVDT and UAVID)our framework achieved a detection accuracy of 94.2%,and 92.3%detection accuracy when conducting a real-time UAV field validation.Our comprehensive evaluations,including multi-metric analyses,ablation studies,and cross-dataset validations,confirm the framework’s accuracy,efficiency,and generalizability.These results highlight the novelty of integrating complementary methods into a single framework,offering a practical solution for accurate and efficient UAV-based traffic monitoring.展开更多
With the increasing emphasis on personal information protection,encryption through security protocols has emerged as a critical requirement in data transmission and reception processes.Nevertheless,IoT ecosystems comp...With the increasing emphasis on personal information protection,encryption through security protocols has emerged as a critical requirement in data transmission and reception processes.Nevertheless,IoT ecosystems comprise heterogeneous networks where outdated systems coexist with the latest devices,spanning a range of devices from non-encrypted ones to fully encrypted ones.Given the limited visibility into payloads in this context,this study investigates AI-based attack detection methods that leverage encrypted traffic metadata,eliminating the need for decryption and minimizing system performance degradation—especially in light of these heterogeneous devices.Using the UNSW-NB15 and CICIoT-2023 dataset,encrypted and unencrypted traffic were categorized according to security protocol,and AI-based intrusion detection experiments were conducted for each traffic type based on metadata.To mitigate the problem of class imbalance,eight different data sampling techniques were applied.The effectiveness of these sampling techniques was then comparatively analyzed using two ensemble models and three Deep Learning(DL)models from various perspectives.The experimental results confirmed that metadata-based attack detection is feasible using only encrypted traffic.In the UNSW-NB15 dataset,the f1-score of encrypted traffic was approximately 0.98,which is 4.3%higher than that of unencrypted traffic(approximately 0.94).In addition,analysis of the encrypted traffic in the CICIoT-2023 dataset using the same method showed a significantly lower f1-score of roughly 0.43,indicating that the quality of the dataset and the preprocessing approach have a substantial impact on detection performance.Furthermore,when data sampling techniques were applied to encrypted traffic,the recall in the UNSWNB15(Encrypted)dataset improved by up to 23.0%,and in the CICIoT-2023(Encrypted)dataset by 20.26%,showing a similar level of improvement.Notably,in CICIoT-2023,f1-score and Receiver Operation Characteristic-Area Under the Curve(ROC-AUC)increased by 59.0%and 55.94%,respectively.These results suggest that data sampling can have a positive effect even in encrypted environments.However,the extent of the improvement may vary depending on data quality,model architecture,and sampling strategy.展开更多
Semantic segmentation for mixed scenes of aerial remote sensing and road traffic is one of the key technologies for visual perception of flying cars.The State-of-the-Art(SOTA)semantic segmentation methods have made re...Semantic segmentation for mixed scenes of aerial remote sensing and road traffic is one of the key technologies for visual perception of flying cars.The State-of-the-Art(SOTA)semantic segmentation methods have made remarkable achievements in both fine-grained segmentation and real-time performance.However,when faced with the huge differences in scale and semantic categories brought about by the mixed scenes of aerial remote sensing and road traffic,they still face great challenges and there is little related research.Addressing the above issue,this paper proposes a semantic segmentation model specifically for mixed datasets of aerial remote sensing and road traffic scenes.First,a novel decoding-recoding multi-scale feature iterative refinement structure is proposed,which utilizes the re-integration and continuous enhancement of multi-scale information to effectively deal with the huge scale differences between cross-domain scenes,while using a fully convolutional structure to ensure the lightweight and real-time requirements.Second,a welldesigned cross-window attention mechanism combined with a global information integration decoding block forms an enhanced global context perception,which can effectively capture the long-range dependencies and multi-scale global context information of different scenes,thereby achieving fine-grained semantic segmentation.The proposed method is tested on a large-scale mixed dataset of aerial remote sensing and road traffic scenes.The results confirm that it can effectively deal with the problem of large-scale differences in cross-domain scenes.Its segmentation accuracy surpasses that of the SOTA methods,which meets the real-time requirements.展开更多
Traffic flow prediction constitutes a fundamental component of Intelligent Transportation Systems(ITS),playing a pivotal role in mitigating congestion,enhancing route optimization,and improving the utilization efficie...Traffic flow prediction constitutes a fundamental component of Intelligent Transportation Systems(ITS),playing a pivotal role in mitigating congestion,enhancing route optimization,and improving the utilization efficiency of roadway infrastructure.However,existingmethods struggle in complex traffic scenarios due to static spatio-temporal embedding,restricted multi-scale temporal modeling,and weak representation of local spatial interactions.This study proposes Bi-STAT+,an enhanced bidirectional spatio-temporal attention framework to address existing limitations through three principal contributions:(1)an adaptive spatio-temporal embedding module that dynamically adjusts embeddings to capture complex traffic variations;(2)frequency-domain analysis in the temporal dimension for simultaneous high-frequency details and low-frequency trend extraction;and(3)an agent attention mechanism in the spatial dimension that enhances local feature extraction through dynamic weight allocation.Extensive experiments were performed on four distinct datasets,including two publicly benchmark datasets(PEMS04 and PEMS08)and two private datasets collected from Baotou and Chengdu,China.The results demonstrate that Bi-STAT+consistently outperforms existing methods in terms of MAE,RMSE,and MAPE,while maintaining strong robustness against missing data and noise.Furthermore,the results highlight that prediction accuracy improves significantly with higher sampling rates,providing crucial insights for optimizing real-world deployment scenarios.展开更多
Reliable detection of traffic signs and lights(TSLs)at long range and under varying illumination is essen-tial for improving the perception and safety of autonomous driving systems(ADS).Traditional object detection mo...Reliable detection of traffic signs and lights(TSLs)at long range and under varying illumination is essen-tial for improving the perception and safety of autonomous driving systems(ADS).Traditional object detection models often exhibit significant performance degradation in real-world environments characterized by high dynamic range and complex lighting conditions.To overcome these limitations,this research presents FED-YOLOv10s,an improved and lightweight object detection framework based on You Only look Once v10(YOLOv10).The proposed model integrates a C2f-Faster block derived from FasterNet to reduce parameters and floating-point operations,an Efficient Multiscale Attention(EMA)mechanism to improve TSL-invariant feature extraction,and a deformable Convolution Networks v4(DCNv4)module to enhance multiscale spatial adaptability.Experimental findings demonstrate that the proposed architecture achieves an optimal balance between computational efficiency and detection accuracy,attaining an F1-score of 91.8%,and mAP@0.5 of 95.1%,while reducing parameters to 8.13 million.Comparative analyses across multiple traffic sign detection benchmarks demonstrate that FED-YOLOv10s outperforms state-of-the-art models in precision,recall,and mAP.These results highlight FED-YOLOv10s as a robust,efficient,and deployable solution for intelligent traffic perception in ADS.展开更多
Lizards usually exhibit frequent turnovers and a much greater diversity of sex determination mechanisms compared to birds and mammals,with the conserved ZW sex chromosomes of anguimorph lizards originating over 115 mi...Lizards usually exhibit frequent turnovers and a much greater diversity of sex determination mechanisms compared to birds and mammals,with the conserved ZW sex chromosomes of anguimorph lizards originating over 115 million years ago a seeming exception.We previously discovered in an anguimorph lizard Varanus acanthurus(Vac)whose entire chrW,but not chrZ is homologous to part of the chr2 by cytogenetic mapping,suggesting its complex history of sex chromosome evolution yet to be elucidated.To address this,we assemble a chromosome-level genome,and provide evidence that the Vac sex chromosome pair has undergone at least two times of recombination loss,producing a pattern of evolutionary strata like that of birds and mammals.Comparison to other lizard genomes date the stepwise propagation of specific retrotransposon subfamilies enriched near the duplicated gene pairs on the chrW and chr2 to the varanid ancestor.These retrotransposons probably have mediated the recruitment and amplification of autosomal genes on the chrW,including members of a large vomeronasal chemosensory receptor gene family V2R.Our results suggest that the W or Y chromosome as a refugium of repetitive elements,may recurrently recruit short-lived functional genes responsible for sexual dimorphisms during its long-term course of degeneration.展开更多
Reliable traffic flow prediction is crucial for mitigating urban congestion.This paper proposes Attentionbased spatiotemporal Interactive Dynamic Graph Convolutional Network(AIDGCN),a novel architecture integrating In...Reliable traffic flow prediction is crucial for mitigating urban congestion.This paper proposes Attentionbased spatiotemporal Interactive Dynamic Graph Convolutional Network(AIDGCN),a novel architecture integrating Interactive Dynamic Graph Convolution Network(IDGCN)with Temporal Multi-Head Trend-Aware Attention.Its core innovation lies in IDGCN,which uniquely splits sequences into symmetric intervals for interactive feature sharing via dynamic graphs,and a novel attention mechanism incorporating convolutional operations to capture essential local traffic trends—addressing a critical gap in standard attention for continuous data.For 15-and 60-min forecasting on METR-LA,AIDGCN achieves MAEs of 0.75%and 0.39%,and RMSEs of 1.32%and 0.14%,respectively.In the 60-min long-term forecasting of the PEMS-BAY dataset,the AIDGCN out-performs the MRA-BGCN method by 6.28%,4.93%,and 7.17%in terms of MAE,RMSE,and MAPE,respectively.Experimental results demonstrate the superiority of our pro-posed model over state-of-the-art methods.展开更多
Alzheimer’s Disease(AD)is a progressive neurodegenerative disorder that significantly affects cognitive function,making early and accurate diagnosis essential.Traditional Deep Learning(DL)-based approaches often stru...Alzheimer’s Disease(AD)is a progressive neurodegenerative disorder that significantly affects cognitive function,making early and accurate diagnosis essential.Traditional Deep Learning(DL)-based approaches often struggle with low-contrast MRI images,class imbalance,and suboptimal feature extraction.This paper develops a Hybrid DL system that unites MobileNetV2 with adaptive classification methods to boost Alzheimer’s diagnosis by processing MRI scans.Image enhancement is done using Contrast-Limited Adaptive Histogram Equalization(CLAHE)and Enhanced Super-Resolution Generative Adversarial Networks(ESRGAN).A classification robustness enhancement system integrates class weighting techniques and a Matthews Correlation Coefficient(MCC)-based evaluation method into the design.The trained and validated model gives a 98.88%accuracy rate and 0.9614 MCC score.We also performed a 10-fold cross-validation experiment with an average accuracy of 96.52%(±1.51),a loss of 0.1671,and an MCC score of 0.9429 across folds.The proposed framework outperforms the state-of-the-art models with a 98%weighted F1-score while decreasing misdiagnosis results for every AD stage.The model demonstrates apparent separation abilities between AD progression stages according to the results of the confusion matrix analysis.These results validate the effectiveness of hybrid DL models with adaptive preprocessing for early and reliable Alzheimer’s diagnosis,contributing to improved computer-aided diagnosis(CAD)systems in clinical practice.展开更多
The exponential growth of the Internet of Things(IoT)has introduced significant security challenges,with zero-day attacks emerging as one of the most critical and challenging threats.Traditional Machine Learning(ML)an...The exponential growth of the Internet of Things(IoT)has introduced significant security challenges,with zero-day attacks emerging as one of the most critical and challenging threats.Traditional Machine Learning(ML)and Deep Learning(DL)techniques have demonstrated promising early detection capabilities.However,their effectiveness is limited when handling the vast volumes of IoT-generated data due to scalability constraints,high computational costs,and the costly time-intensive process of data labeling.To address these challenges,this study proposes a Federated Learning(FL)framework that leverages collaborative and hybrid supervised learning to enhance cyber threat detection in IoT networks.By employing Deep Neural Networks(DNNs)and decentralized model training,the approach reduces computational complexity while improving detection accuracy.The proposed model demonstrates robust performance,achieving accuracies of 94.34%,99.95%,and 87.94%on the publicly available kitsune,Bot-IoT,and UNSW-NB15 datasets,respectively.Furthermore,its ability to detect zero-day attacks is validated through evaluations on two additional benchmark datasets,TON-IoT and IoT-23,using a Deep Federated Learning(DFL)framework,underscoring the generalization and effectiveness of the model in heterogeneous and decentralized IoT environments.Experimental results demonstrate superior performance over existing methods,establishing the proposed framework as an efficient and scalable solution for IoT security.展开更多
Accurate and real-time traffic-sign detection is a cornerstone of Advanced Driver-Assistance Systems(ADAS)and autonomous vehicles.However,existing one-stage detectors miss distant signs,and two-stage pipelines are imp...Accurate and real-time traffic-sign detection is a cornerstone of Advanced Driver-Assistance Systems(ADAS)and autonomous vehicles.However,existing one-stage detectors miss distant signs,and two-stage pipelines are impractical for embedded deployment.To address this issue,we present YOLO-SMM,a lightweight two-stage framework.This framework is designed to augment the YOLOv8 baseline with three targeted modules.(1)SlimNeck replaces PAN/FPN with a CSP-OSA/GSConv fusion block,reducing parameters and FLOPs without compromising multi-scale detail.(2)The MCA model introduces row-and column-aware weights to selectively amplify small sign regions in cluttered scenes.(3)MPDIoU augments CIoU loss with a corner-distance term,supplying stable gradients for sub-20-pixel boxes and tightening localization.An evaluation of YOLO-SMMon the German Traffic Sign Recognition Benchmark(GTSRB)revealed that it attained 96.3% mAP50 and 93.1% mAP50-90 at a rate of 90.6 frames per second(FPS).This represents an improvement of+1.0% over previous performance benchmarks.Them APat 64×64 resolution was found to be 50% of the maximum attainable value,with an FPS of+8.3 when compared to YOLOv8.This result indicates superior performance in terms of accuracy and speed compared to YOLOv7,YOLOv5,RetinaNet,EfficientDet,and Faster R-CNN,all of which were operated under equivalent conditions.展开更多
基金Supported by the National High Technology Research and Development Program of China(″863″Program)(2010AAxxx404)~~
文摘Due to the diversified demands of quality of service（QoS） in volume multimedia application, QoS routings for multiservice are becoming a research hotspot in low earth orbit（LEO） satellite networks. A novel QoS satellite routing algorithm for multi-class traffic is proposed. The goal of the routing algorithm is to provide the distinct QoS for different traffic classes and improve the utilization of network resources. Traffic is classified into three classes and allocated priorities based on their QoS requirements, respectively. A priority queuing mechanism guarantees the algorithm to work better for high-priority classes. In order to control the congestion, a blocking probability analysis model is built up based on the Markov process theory. Finally, according to the classification link-cost metrics, routings for different classes are calculated with the distinct QoS requirments and the status of network resource. Simulations verify the performance of the routing algorithm at different time and in different regions, and results demonstrate that the algorithm has great advantages in terms of the average delay and the blocking probability. Meanwhile, the robustness issue is also discussed.
基金Project supported by the National Natural Science Foundation of China (Nos. 10472064,10371118)the Post-Doctoral Science Foundation of China (No. 2003034254)the Special Fund for PhD Program of Education Ministry of China (No. 20040280014)
文摘A strict proof of the hyperbolicity of the multi-class LWR ( Lighthill-Whitham-Richards) traffic flow model, as well as the descriptions on those nonlinear waves characterized in the traffic flow problems were given. They were mainly about the monotonicity of densities across shocks and in rarefactions. As the system had no characteristic decomposition explicitly, a high resolution and higher order accuracy WENO( weighted essentially non-oscillatory) scheme was introduced to the numerical simulation, which coincides with the analytical description.
基金supported by the Korea Institute of Police Technology(No.:RS-2024-00405603).
文摘Autonomous vehicles operate without direct human intervention,which introduces safety risks that differ from those of conventional vehicles.Although many studies have examined safety issues related to autonomous driving,high-risk situations have often been defined using single indicators,making it difficult to capture the complex and evolving nature of accident risk.To address this limitation,this study proposes a structured framework for defining and analyzing high-risk situations throughout the traffic accident process.High-risk situations are described using three complementary indicators:accident likelihood,accident severity,and accident duration.These indicators explain how risk emerges,increases,and persists over time.Based on this concept,a framework for traffic accident visualization analysis is developed to support phase-specific risk assessment and visualization.The framework combines accident-phase information with factor-level risk contributions,allowing systematic identification of key factors and their interactions across different accident stages.Using combinations of the three indicators,high-risk situations are classified into twenty-seven distinct types,providing a clear typology for complex accident scenarios involving autonomous vehicles.The applicability of the proposed framework is demonstrated through two representative accident scenarioswith different risk characteristics.The results showthat the framework effectively captures interactions among multiple risk factors,explains how risk levels change from pre-crash to post-crash phases,and identifies contributing factors that are difficult to detect using conventional traffic accident investigation methods.Overall,the proposed framework offers a practical basis for autonomous vehicle accident analysis,safety evaluation,and policy-related decision-making.
基金supported by the National Natural Science Foundation of China 62402171.
文摘As a fundamental component in computer vision,edges can be categorized into four types based on discontinuities in reflectance,illumination,surface normal,or depth.While deep CNNs have significantly advanced generic edge detection,real-time multi-class semantic edge detection under resource constraints remains challenging.To address this,we propose a lightweight framework based on PiDiNet that enables fine-grained semantic edge detection.Our model simultaneously predicts background and four edge categories from full-resolution inputs,balancing accuracy and efficiency.Key contributions include:a multi-channel output structure expanding binary edge prediction to five classes,supported by a deep supervision mechanism;a dynamic class-balancing strategy combining adaptive weighting with physical priors to handle extreme class imbalance;and maintained architectural efficiency enabling real-time inference.Extensive evaluations on BSDS-RIND show our approach achieves accuracy competitive with state-of-the-art methods while operating in real time.
基金supported by ZTE Industry-University-Institute Cooperation Funds under Grant No.HC-CN-20220607009。
文摘With the evolution of next-generation network technologies,the complexity of network management has significantly increased,and the means of network attacks are diversified,bringing new challenges to network traffic classification.This paper presents a general AIdriven network traffic classification workflow and elaborates on a traffic data and feature engineering framework.Most importantly,it analyzes the concept and causes of data distribution shifts in ne twork traffic,proposing detection methods and countermeasures.Experimental results on real traffic collected at different time intervals show that application evolution can induce data distribution shifts,which in turn lead to a noticeable degradation in traffic classification performance.Comparative drift detection experiments further confirm that such shifts are more evident over long-term intervals,while short-term traffic remains relatively stable.These findings demonstrate the necessity of incorporating drift-aware mechanisms into AI-driven network traffic classification systems.
基金The National Social Science Foundation Youth Project of China:Research on the collaborative govemance path of administrative law and criminal law against dangerous driving behaviors in the digital-intelligent society(25CFX108)。
文摘With the continuous progress of automatic driving technology,automatic driving technology standards are gradually affecting the determination of criminal responsibility for traffic accidents in China.At present,the characteristics and tendency of China's automatic driving technology standards present the situation of high policy relevance coexisting with low normative binding,professionalism coexist with barriers,forefront coexist with ambiguity.Therefore,challenges are presented both theoretically and practically on the determination of criminal responsibility based on automatic driving technology standard..In this regard,the misunderstanding should be clarified in theory:The legal order under the automatic driving technology standard has constitutionality and systematic,and there is a balance between the frontier of automatic driving technology development and the lagging of criminal law.The automatic driving technology risk level system should be built to clarify the boundary of the effectiveness of criminal law norms,seeking fora breakthrough in the application of the establishment of a comprehensive judgment system of the risks and accidents and the system of evidence to prove the system,which clarifies the determination of criminal responsibility under the automatic driving technology standard.This essay hopes to pursue breakthroughs in the application-to establish a comprehensive judgment system of risks and accidents as well as an evidence proof system,so as to clarify the determination of criminal responsibility under automatic driving technology standards.
文摘This paper investigates the traffic offloading optimization challenge in Space-Air-Ground Integrated Networks(SAGIN)through a novel Recursive Multi-Agent Proximal Policy Optimization(RMAPPO)algorithm.The exponential growth of mobile devices and data traffic has substantially increased network congestion,particularly in urban areas and regions with limited terrestrial infrastructure.Our approach jointly optimizes unmanned aerial vehicle(UAV)trajectories and satellite-assisted offloading strategies to simultaneously maximize data throughput,minimize energy consumption,and maintain equitable resource distribution.The proposed RMAPPO framework incorporates recurrent neural networks(RNNs)to model temporal dependencies in UAV mobility patterns and utilizes a decentralized multi-agent reinforcement learning architecture to reduce communication overhead while improving system robustness.The proposed RMAPPO algorithm was evaluated through simulation experiments,with the results indicating that it significantly enhances the cumulative traffic offloading rate of nodes and reduces the energy consumption of UAVs.
基金supported by the National Natural Science Foundation of China(Grant No.62172033).
文摘Urban traffic generates massive and diverse data,yet most systems remain fragmented.Current approaches to congestion management suffer from weak data consistency and poor scalability.This study addresses this gap by proposing the Urban Traffic Congestion Unified Metadata Model(UTC-UMM).The goal is to provide a standardized and extensible framework for describing,extracting,and storing multisource traffic data in smart cities.The model defines a two-tier specification that organizes nine core traffic resource classes.It employs an eXtensible Markup Language(XML)Schema that connects general elements with resource-specific elements.This design ensures both syntactic and semantic interoperability across siloed datasets.Extension principles allow new elements or constraints to be introducedwithout breaking backward compatibility.Adistributed pipeline is implemented usingHadoop Distributed File System(HDFS)and HBase.It integrates computer vision for video and natural language processing for text to automate metadata extraction.Optimized row-key designs enable low-latency queries.Performance is tested with the Yahoo!Cloud Serving Benchmark(YCSB),which shows linear scalability and high throughput.The results demonstrate that UTC-UMM can unify heterogeneous traffic data while supporting real-time analytics.The discussion highlights its potential to improve data reuse,portability,and scalability in urban congestion studies.Future research will explore integration with association rulemining and advanced knowledge representation to capture richer spatiotemporal traffic patterns.
基金funded by Key research and development Program of Henan Province(No.251111211200)National Natural Science Foundation of China(Grant No.U2004163).
文摘Traffic sign detection is an important part of autonomous driving,and its recognition accuracy and speed are directly related to road traffic safety.Although convolutional neural networks(CNNs)have made certain breakthroughs in this field,in the face of complex scenes,such as image blur and target occlusion,the traffic sign detection continues to exhibit limited accuracy,accompanied by false positives and missed detections.To address the above problems,a traffic sign detection algorithm,You Only Look Once-based Skip Dynamic Way(YOLO-SDW)based on You Only Look Once version 8 small(YOLOv8s),is proposed.Firstly,a Skip Connection Reconstruction(SCR)module is introduced to efficiently integrate fine-grained feature information and enhance the detection accuracy of the algorithm in complex scenes.Secondly,a C2f module based on Dynamic Snake Convolution(C2f-DySnake)is proposed to dynamically adjust the receptive field information,improve the algorithm’s feature extraction ability for blurred or occluded targets,and reduce the occurrence of false detections and missed detections.Finally,the Wise Powerful IoU v2(WPIoUv2)loss function is proposed to further improve the detection accuracy of the algorithm.Experimental results show that the average precision mAP@0.5 of YOLO-SDW on the TT100K dataset is 89.2%,and mAP@0.5:0.95 is 68.5%,which is 4%and 3.3%higher than the YOLOv8s baseline,respectively.YOLO-SDW ensures real-time performance while having higher accuracy.
基金National Natural Science Foundation of China(Grant No.62103434)National Science Fund for Distinguished Young Scholars(Grant No.62176263).
文摘With network attack technology continuing to develop,traditional anomaly traffic detection methods that rely on feature engineering are increasingly insufficient in efficiency and accuracy.Graph Neural Network(GNN),a promising Deep Learning(DL)approach,has proven to be highly effective in identifying intricate patterns in graph⁃structured data and has already found wide applications in the field of network security.In this paper,we propose a hybrid Graph Convolutional Network(GCN)⁃GraphSAGE model for Anomaly Traffic Detection,namely HGS⁃ATD,which aims to improve the accuracy of anomaly traffic detection by leveraging edge feature learning to better capture the relationships between network entities.We validate the HGS⁃ATD model on four publicly available datasets,including NF⁃UNSW⁃NB15⁃v2.The experimental results show that the enhanced hybrid model is 5.71%to 10.25%higher than the baseline model in terms of accuracy,and the F1⁃score is 5.53%to 11.63%higher than the baseline model,proving that the model can effectively distinguish normal traffic from attack traffic and accurately classify various types of attacks.
基金supported by the IITP(Institute of Information&Communications Technology Planning&Evaluation)-ICAN(ICT Challenge and Advanced Network of HRD)(IITP-2025-RS-2022-00156326,50)grant funded by theKorea government(Ministry of Science and ICT)supported and funded by Princess Nourah bint Abdulrahman University Researchers Supporting Project number(PNURSP2025R410)Princess Nourah bint Abdulrahman University,Riyadh,Saudi Arabia。
文摘This paper presents a unified Unmanned Aerial Vehicle-based(UAV-based)traffic monitoring framework that integrates vehicle detection,tracking,counting,motion prediction,and classification in a modular and co-optimized pipeline.Unlike prior works that address these tasks in isolation,our approach combines You Only Look Once(YOLO)v10 detection,ByteTrack tracking,optical-flow density estimation,Long Short-Term Memory-based(LSTM-based)trajectory forecasting,and hybrid Speeded-Up Robust Feature(SURF)+Gray-Level Co-occurrence Matrix(GLCM)feature engineering with VGG16 classification.Upon the validation across datasets(UAVDT and UAVID)our framework achieved a detection accuracy of 94.2%,and 92.3%detection accuracy when conducting a real-time UAV field validation.Our comprehensive evaluations,including multi-metric analyses,ablation studies,and cross-dataset validations,confirm the framework’s accuracy,efficiency,and generalizability.These results highlight the novelty of integrating complementary methods into a single framework,offering a practical solution for accurate and efficient UAV-based traffic monitoring.
基金supported by the Institute of Information&Communications Technology Planning&Evaluation(IITP)grant funded by the Korea government(MSIT)(No.RS-2023-00235509Development of security monitoring technology based network behavior against encrypted cyber threats in ICT convergence environment).
文摘With the increasing emphasis on personal information protection,encryption through security protocols has emerged as a critical requirement in data transmission and reception processes.Nevertheless,IoT ecosystems comprise heterogeneous networks where outdated systems coexist with the latest devices,spanning a range of devices from non-encrypted ones to fully encrypted ones.Given the limited visibility into payloads in this context,this study investigates AI-based attack detection methods that leverage encrypted traffic metadata,eliminating the need for decryption and minimizing system performance degradation—especially in light of these heterogeneous devices.Using the UNSW-NB15 and CICIoT-2023 dataset,encrypted and unencrypted traffic were categorized according to security protocol,and AI-based intrusion detection experiments were conducted for each traffic type based on metadata.To mitigate the problem of class imbalance,eight different data sampling techniques were applied.The effectiveness of these sampling techniques was then comparatively analyzed using two ensemble models and three Deep Learning(DL)models from various perspectives.The experimental results confirmed that metadata-based attack detection is feasible using only encrypted traffic.In the UNSW-NB15 dataset,the f1-score of encrypted traffic was approximately 0.98,which is 4.3%higher than that of unencrypted traffic(approximately 0.94).In addition,analysis of the encrypted traffic in the CICIoT-2023 dataset using the same method showed a significantly lower f1-score of roughly 0.43,indicating that the quality of the dataset and the preprocessing approach have a substantial impact on detection performance.Furthermore,when data sampling techniques were applied to encrypted traffic,the recall in the UNSWNB15(Encrypted)dataset improved by up to 23.0%,and in the CICIoT-2023(Encrypted)dataset by 20.26%,showing a similar level of improvement.Notably,in CICIoT-2023,f1-score and Receiver Operation Characteristic-Area Under the Curve(ROC-AUC)increased by 59.0%and 55.94%,respectively.These results suggest that data sampling can have a positive effect even in encrypted environments.However,the extent of the improvement may vary depending on data quality,model architecture,and sampling strategy.
基金supported by the National Key Research and Development of China(No.2022YFB2503400).
文摘Semantic segmentation for mixed scenes of aerial remote sensing and road traffic is one of the key technologies for visual perception of flying cars.The State-of-the-Art(SOTA)semantic segmentation methods have made remarkable achievements in both fine-grained segmentation and real-time performance.However,when faced with the huge differences in scale and semantic categories brought about by the mixed scenes of aerial remote sensing and road traffic,they still face great challenges and there is little related research.Addressing the above issue,this paper proposes a semantic segmentation model specifically for mixed datasets of aerial remote sensing and road traffic scenes.First,a novel decoding-recoding multi-scale feature iterative refinement structure is proposed,which utilizes the re-integration and continuous enhancement of multi-scale information to effectively deal with the huge scale differences between cross-domain scenes,while using a fully convolutional structure to ensure the lightweight and real-time requirements.Second,a welldesigned cross-window attention mechanism combined with a global information integration decoding block forms an enhanced global context perception,which can effectively capture the long-range dependencies and multi-scale global context information of different scenes,thereby achieving fine-grained semantic segmentation.The proposed method is tested on a large-scale mixed dataset of aerial remote sensing and road traffic scenes.The results confirm that it can effectively deal with the problem of large-scale differences in cross-domain scenes.Its segmentation accuracy surpasses that of the SOTA methods,which meets the real-time requirements.
基金partly supported by the Youth Foundation of the Inner Mongolia Natural Science Foundation[grant number 2024QN06017 and 2025MS06022]the Basic Scientific Research Business Fee Project for Universities in Inner Mongolia[grant numbers 2023XKJX019 and 2023XKJX024]the Central Guidance on Local Science and Technology Development Fund through[grant number 2024ZY0084].
文摘Traffic flow prediction constitutes a fundamental component of Intelligent Transportation Systems(ITS),playing a pivotal role in mitigating congestion,enhancing route optimization,and improving the utilization efficiency of roadway infrastructure.However,existingmethods struggle in complex traffic scenarios due to static spatio-temporal embedding,restricted multi-scale temporal modeling,and weak representation of local spatial interactions.This study proposes Bi-STAT+,an enhanced bidirectional spatio-temporal attention framework to address existing limitations through three principal contributions:(1)an adaptive spatio-temporal embedding module that dynamically adjusts embeddings to capture complex traffic variations;(2)frequency-domain analysis in the temporal dimension for simultaneous high-frequency details and low-frequency trend extraction;and(3)an agent attention mechanism in the spatial dimension that enhances local feature extraction through dynamic weight allocation.Extensive experiments were performed on four distinct datasets,including two publicly benchmark datasets(PEMS04 and PEMS08)and two private datasets collected from Baotou and Chengdu,China.The results demonstrate that Bi-STAT+consistently outperforms existing methods in terms of MAE,RMSE,and MAPE,while maintaining strong robustness against missing data and noise.Furthermore,the results highlight that prediction accuracy improves significantly with higher sampling rates,providing crucial insights for optimizing real-world deployment scenarios.
基金funded by the Deanship of Scientific Research(DSR)at King Abdulaziz University,Jeddah,Saudi Arabia under Grant No.IPP:172-830-2025.
文摘Reliable detection of traffic signs and lights(TSLs)at long range and under varying illumination is essen-tial for improving the perception and safety of autonomous driving systems(ADS).Traditional object detection models often exhibit significant performance degradation in real-world environments characterized by high dynamic range and complex lighting conditions.To overcome these limitations,this research presents FED-YOLOv10s,an improved and lightweight object detection framework based on You Only look Once v10(YOLOv10).The proposed model integrates a C2f-Faster block derived from FasterNet to reduce parameters and floating-point operations,an Efficient Multiscale Attention(EMA)mechanism to improve TSL-invariant feature extraction,and a deformable Convolution Networks v4(DCNv4)module to enhance multiscale spatial adaptability.Experimental findings demonstrate that the proposed architecture achieves an optimal balance between computational efficiency and detection accuracy,attaining an F1-score of 91.8%,and mAP@0.5 of 95.1%,while reducing parameters to 8.13 million.Comparative analyses across multiple traffic sign detection benchmarks demonstrate that FED-YOLOv10s outperforms state-of-the-art models in precision,recall,and mAP.These results highlight FED-YOLOv10s as a robust,efficient,and deployable solution for intelligent traffic perception in ADS.
基金supported by the National Key Research and Development Program of China(2023YFA1800500,2024YFA1802500)the National Natural Science Foundation of China(32170415)+1 种基金supported by the Australian Government Research Training Program(RTP)stipend scholarship.supported by the Australian Research Council Discovery Project grant(ARC DP200101406).
文摘Lizards usually exhibit frequent turnovers and a much greater diversity of sex determination mechanisms compared to birds and mammals,with the conserved ZW sex chromosomes of anguimorph lizards originating over 115 million years ago a seeming exception.We previously discovered in an anguimorph lizard Varanus acanthurus(Vac)whose entire chrW,but not chrZ is homologous to part of the chr2 by cytogenetic mapping,suggesting its complex history of sex chromosome evolution yet to be elucidated.To address this,we assemble a chromosome-level genome,and provide evidence that the Vac sex chromosome pair has undergone at least two times of recombination loss,producing a pattern of evolutionary strata like that of birds and mammals.Comparison to other lizard genomes date the stepwise propagation of specific retrotransposon subfamilies enriched near the duplicated gene pairs on the chrW and chr2 to the varanid ancestor.These retrotransposons probably have mediated the recruitment and amplification of autosomal genes on the chrW,including members of a large vomeronasal chemosensory receptor gene family V2R.Our results suggest that the W or Y chromosome as a refugium of repetitive elements,may recurrently recruit short-lived functional genes responsible for sexual dimorphisms during its long-term course of degeneration.
文摘Reliable traffic flow prediction is crucial for mitigating urban congestion.This paper proposes Attentionbased spatiotemporal Interactive Dynamic Graph Convolutional Network(AIDGCN),a novel architecture integrating Interactive Dynamic Graph Convolution Network(IDGCN)with Temporal Multi-Head Trend-Aware Attention.Its core innovation lies in IDGCN,which uniquely splits sequences into symmetric intervals for interactive feature sharing via dynamic graphs,and a novel attention mechanism incorporating convolutional operations to capture essential local traffic trends—addressing a critical gap in standard attention for continuous data.For 15-and 60-min forecasting on METR-LA,AIDGCN achieves MAEs of 0.75%and 0.39%,and RMSEs of 1.32%and 0.14%,respectively.In the 60-min long-term forecasting of the PEMS-BAY dataset,the AIDGCN out-performs the MRA-BGCN method by 6.28%,4.93%,and 7.17%in terms of MAE,RMSE,and MAPE,respectively.Experimental results demonstrate the superiority of our pro-posed model over state-of-the-art methods.
基金funded by the Deanship of Graduate Studies and Scientific Research at Jouf University under grant No.(DGSSR-2025-02-01295).
文摘Alzheimer’s Disease(AD)is a progressive neurodegenerative disorder that significantly affects cognitive function,making early and accurate diagnosis essential.Traditional Deep Learning(DL)-based approaches often struggle with low-contrast MRI images,class imbalance,and suboptimal feature extraction.This paper develops a Hybrid DL system that unites MobileNetV2 with adaptive classification methods to boost Alzheimer’s diagnosis by processing MRI scans.Image enhancement is done using Contrast-Limited Adaptive Histogram Equalization(CLAHE)and Enhanced Super-Resolution Generative Adversarial Networks(ESRGAN).A classification robustness enhancement system integrates class weighting techniques and a Matthews Correlation Coefficient(MCC)-based evaluation method into the design.The trained and validated model gives a 98.88%accuracy rate and 0.9614 MCC score.We also performed a 10-fold cross-validation experiment with an average accuracy of 96.52%(±1.51),a loss of 0.1671,and an MCC score of 0.9429 across folds.The proposed framework outperforms the state-of-the-art models with a 98%weighted F1-score while decreasing misdiagnosis results for every AD stage.The model demonstrates apparent separation abilities between AD progression stages according to the results of the confusion matrix analysis.These results validate the effectiveness of hybrid DL models with adaptive preprocessing for early and reliable Alzheimer’s diagnosis,contributing to improved computer-aided diagnosis(CAD)systems in clinical practice.
基金supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project Number(PNURSP2025R97)Princess Nourah bint Abdulrahman University,Riyadh,Saudi Arabia.
文摘The exponential growth of the Internet of Things(IoT)has introduced significant security challenges,with zero-day attacks emerging as one of the most critical and challenging threats.Traditional Machine Learning(ML)and Deep Learning(DL)techniques have demonstrated promising early detection capabilities.However,their effectiveness is limited when handling the vast volumes of IoT-generated data due to scalability constraints,high computational costs,and the costly time-intensive process of data labeling.To address these challenges,this study proposes a Federated Learning(FL)framework that leverages collaborative and hybrid supervised learning to enhance cyber threat detection in IoT networks.By employing Deep Neural Networks(DNNs)and decentralized model training,the approach reduces computational complexity while improving detection accuracy.The proposed model demonstrates robust performance,achieving accuracies of 94.34%,99.95%,and 87.94%on the publicly available kitsune,Bot-IoT,and UNSW-NB15 datasets,respectively.Furthermore,its ability to detect zero-day attacks is validated through evaluations on two additional benchmark datasets,TON-IoT and IoT-23,using a Deep Federated Learning(DFL)framework,underscoring the generalization and effectiveness of the model in heterogeneous and decentralized IoT environments.Experimental results demonstrate superior performance over existing methods,establishing the proposed framework as an efficient and scalable solution for IoT security.
基金supported by University of Malaya and Ministry of High Education-Malaysia via Fundamental Research Grant Scheme No.FRGS/1/2023/TK10/UM/02/3.
文摘Accurate and real-time traffic-sign detection is a cornerstone of Advanced Driver-Assistance Systems(ADAS)and autonomous vehicles.However,existing one-stage detectors miss distant signs,and two-stage pipelines are impractical for embedded deployment.To address this issue,we present YOLO-SMM,a lightweight two-stage framework.This framework is designed to augment the YOLOv8 baseline with three targeted modules.(1)SlimNeck replaces PAN/FPN with a CSP-OSA/GSConv fusion block,reducing parameters and FLOPs without compromising multi-scale detail.(2)The MCA model introduces row-and column-aware weights to selectively amplify small sign regions in cluttered scenes.(3)MPDIoU augments CIoU loss with a corner-distance term,supplying stable gradients for sub-20-pixel boxes and tightening localization.An evaluation of YOLO-SMMon the German Traffic Sign Recognition Benchmark(GTSRB)revealed that it attained 96.3% mAP50 and 93.1% mAP50-90 at a rate of 90.6 frames per second(FPS).This represents an improvement of+1.0% over previous performance benchmarks.Them APat 64×64 resolution was found to be 50% of the maximum attainable value,with an FPS of+8.3 when compared to YOLOv8.This result indicates superior performance in terms of accuracy and speed compared to YOLOv7,YOLOv5,RetinaNet,EfficientDet,and Faster R-CNN,all of which were operated under equivalent conditions.
