Pub Date : 2025-03-06DOI: 10.1109/ACCESS.2025.3548726
Linjin Wang;Jiangtao He;Guohong Geng;Lisha Zhong;Xinwei Li
Accurate segmentation of hippocampal subfields in MRI scans is crucial for aiding in the diagnosis of various neurological diseases and for monitoring brain states. However, due to limitations of imaging systems and the inherent complexity of hippocampal subfield delineation, achieving accurate hippocampal subfield delineation on routine 3T MRI is highly challenging. In this paper, we propose a novel Guided Learning Network (GLNet) that leverages 7T MRI to enhance the accuracy of hippocampal subfield segmentation on routine 3T MRI. GLNet aligns and learns shared features between 3T MRI and 7T MRI through a modeling approach based on domain-specific and domain-shared feature learning, leveraging the features of 7T MRI to guide learning for 3T MRI features. In this process, we also introduce a Multi-Feature Attention Fusion (MFAF) block to integrate both specific and shared features from each modality. By leveraging an attention mechanism, MFAF adaptively focuses on relevant information between the specific and shared features within the same modality, thereby reducing the impact of irrelevant information. Additionally, we further proposed an Online Knowledge Distillation (OLKD) method to distill detailed knowledge from 7T MRI into 3T MRI, enhancing the feature representation capability and robustness of the 3T MRI segmentation model. Our method was validated on PAIRED 3T-7T HIPPOCAMPAL SUBFIELD DATASET, and the experimental results demonstrate that GLNet outperforms other competitive methods.
{"title":"A 7T MRI-Guided Learning Method for Automatic Hippocampal Subfield Segmentation on Routine 3T MRI","authors":"Linjin Wang;Jiangtao He;Guohong Geng;Lisha Zhong;Xinwei Li","doi":"10.1109/ACCESS.2025.3548726","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3548726","url":null,"abstract":"Accurate segmentation of hippocampal subfields in MRI scans is crucial for aiding in the diagnosis of various neurological diseases and for monitoring brain states. However, due to limitations of imaging systems and the inherent complexity of hippocampal subfield delineation, achieving accurate hippocampal subfield delineation on routine 3T MRI is highly challenging. In this paper, we propose a novel Guided Learning Network (GLNet) that leverages 7T MRI to enhance the accuracy of hippocampal subfield segmentation on routine 3T MRI. GLNet aligns and learns shared features between 3T MRI and 7T MRI through a modeling approach based on domain-specific and domain-shared feature learning, leveraging the features of 7T MRI to guide learning for 3T MRI features. In this process, we also introduce a Multi-Feature Attention Fusion (MFAF) block to integrate both specific and shared features from each modality. By leveraging an attention mechanism, MFAF adaptively focuses on relevant information between the specific and shared features within the same modality, thereby reducing the impact of irrelevant information. Additionally, we further proposed an Online Knowledge Distillation (OLKD) method to distill detailed knowledge from 7T MRI into 3T MRI, enhancing the feature representation capability and robustness of the 3T MRI segmentation model. Our method was validated on PAIRED 3T-7T HIPPOCAMPAL SUBFIELD DATASET, and the experimental results demonstrate that GLNet outperforms other competitive methods.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"42428-42440"},"PeriodicalIF":3.4,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10915623","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143594276","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Advancements in vehicle technology such as collision warning systems are essential for improving driver alertness and decision-making to enhance road safety. This study evaluates the effectiveness of 2-level and 3-level graded collision warning systems on driver performance under various driving conditions. Forty participants participated in a controlled driving simulator study using a within-between-participant design to examine the impact of these warning systems on response time, collision frequency, physiological responses, and visual attention dynamics. The 3-level system providing graduated alerts through visual, haptic, and auditory cues, showed a significant reduction in response times and collision frequencies compared to the 2-level system. This improvement likely results from the multi-sensory approach that supports cognitive load theory and facilitates hazard detection. Although physiological measures such as Electrodermal Activity and Heart Rate did not show significant differences between the systems, the 3-level system produced more consistent responses suggesting a stable emotional state and reduced stress. Eye-tracking data indicated that the 3-level system improved sustained visual attention and reduced distraction. Subjective evaluations favored auditory warnings emphasizing the importance of user-friendly and intuitive systems. The findings demonstrate the potential of multi-level warning systems to enhance driver safety and performance in high-risk scenarios and suggest the need for customizable systems to accommodate individual differences in cognitive load management. These insights can inform developing advanced collision warning systems that mitigate risks associated with distracted driving and promote a safer driving environment.
{"title":"Comparison of Two-Level and Three-Level Graded Collision Warning Systems Under Distracted Driving Conditions","authors":"Khatereh Shariatmadari;Siby Samuel;Shi Cao;Amandeep Singh","doi":"10.1109/ACCESS.2025.3549290","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3549290","url":null,"abstract":"Advancements in vehicle technology such as collision warning systems are essential for improving driver alertness and decision-making to enhance road safety. This study evaluates the effectiveness of 2-level and 3-level graded collision warning systems on driver performance under various driving conditions. Forty participants participated in a controlled driving simulator study using a within-between-participant design to examine the impact of these warning systems on response time, collision frequency, physiological responses, and visual attention dynamics. The 3-level system providing graduated alerts through visual, haptic, and auditory cues, showed a significant reduction in response times and collision frequencies compared to the 2-level system. This improvement likely results from the multi-sensory approach that supports cognitive load theory and facilitates hazard detection. Although physiological measures such as Electrodermal Activity and Heart Rate did not show significant differences between the systems, the 3-level system produced more consistent responses suggesting a stable emotional state and reduced stress. Eye-tracking data indicated that the 3-level system improved sustained visual attention and reduced distraction. Subjective evaluations favored auditory warnings emphasizing the importance of user-friendly and intuitive systems. The findings demonstrate the potential of multi-level warning systems to enhance driver safety and performance in high-risk scenarios and suggest the need for customizable systems to accommodate individual differences in cognitive load management. These insights can inform developing advanced collision warning systems that mitigate risks associated with distracted driving and promote a safer driving environment.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"43818-43829"},"PeriodicalIF":3.4,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10916632","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143621736","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-03-06DOI: 10.1109/ACCESS.2025.3548864
Kapeel Kumar;Qasim Ali;Sheeraz Ahmed;Muhammad Humza;Abdul Aziz Memon;Mudassir Raza Siddiqi
This paper presents the design and performance analysis of a brushless wound rotor vernier machine (BLWRVM) for variable speed applications, particularly in the context of domestic appliances like automatic washing machines which have two cycles of operation: the wash cycle at low-speed and dry-cycle at high-speed. The proposed BLWRVM utilizes a different winding configuration that enhances the performance compared to the existing reference machine. The machine topology uses a dual winding setup on the stator, a 3-phase main ABC1 winding and a 1-phase additional A2 winding on the stator, specifically engineered to generate both fundamental and subharmonic components of MMF in the airgap of the machine. In order to achieve the brushless operation, the additional excitation winding is placed on the rotor which is connected through a rectifier to the field winding. The voltages are induced in this excitation winding using the MMF’s sub-harmonic component. These AC-induced voltages are rectified by a rotating rectifier placed on the rotor periphery and are then supplied to the main field winding. A 4-pole, 24-slot stator and 44-pole 44-slot outer rotor machine is designed. The fill factor of the machine’s stator slot was then increased to 35% and the rotor’s slot fill factor was increased to 30%. Furthermore, different skew angles of the rotor were analyzed to find the optimal skew angle that yields minimum torque ripple. The performance characteristics were compared with a reference BLWRVM. To validate the design a 2D-FEA was carried out to find out the performing characteristics of the proposed machine and reference machines using ANSYS Electromagnetic suite version 2022 R1.
{"title":"Performance Improvement of Brushless Wound Rotor Vernier Motor for Washing Machine Applications","authors":"Kapeel Kumar;Qasim Ali;Sheeraz Ahmed;Muhammad Humza;Abdul Aziz Memon;Mudassir Raza Siddiqi","doi":"10.1109/ACCESS.2025.3548864","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3548864","url":null,"abstract":"This paper presents the design and performance analysis of a brushless wound rotor vernier machine (BLWRVM) for variable speed applications, particularly in the context of domestic appliances like automatic washing machines which have two cycles of operation: the wash cycle at low-speed and dry-cycle at high-speed. The proposed BLWRVM utilizes a different winding configuration that enhances the performance compared to the existing reference machine. The machine topology uses a dual winding setup on the stator, a 3-phase main ABC1 winding and a 1-phase additional A2 winding on the stator, specifically engineered to generate both fundamental and subharmonic components of MMF in the airgap of the machine. In order to achieve the brushless operation, the additional excitation winding is placed on the rotor which is connected through a rectifier to the field winding. The voltages are induced in this excitation winding using the MMF’s sub-harmonic component. These AC-induced voltages are rectified by a rotating rectifier placed on the rotor periphery and are then supplied to the main field winding. A 4-pole, 24-slot stator and 44-pole 44-slot outer rotor machine is designed. The fill factor of the machine’s stator slot was then increased to 35% and the rotor’s slot fill factor was increased to 30%. Furthermore, different skew angles of the rotor were analyzed to find the optimal skew angle that yields minimum torque ripple. The performance characteristics were compared with a reference BLWRVM. To validate the design a 2D-FEA was carried out to find out the performing characteristics of the proposed machine and reference machines using ANSYS Electromagnetic suite version 2022 R1.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"44739-44749"},"PeriodicalIF":3.4,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10915647","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143621693","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-03-06DOI: 10.1109/ACCESS.2025.3549357
S. J. Dat Tran
The Hodgkin and Huxley neuron model describes the complex behavior of biological neurons. However, due to the complexity of these computations, the Hodgkin and Huxley models are impractical for use in large-scale networks. In contrast, Izhikevich introduced a simpler model capable of producing various firing patterns typical of cortical neurons. This study proposes a novel model of memcapacitive-based neurons that offers a potential implementation of spiking neurons with energy efficiency due to the inherent storage nature of memcapacitive devices. The findings demonstrate that memcapacitive neurons can produce 23 firing patterns similar to Izhikevich neurons but at significantly higher firing rates. Memcapacitive neurons exhibit firing patterns associated with excitatory, inhibitory, and thalamocortical neurons. Similar to Izhikevich neurons, pulse-coupled neural networks of memcapacitive neurons display collective behaviors, such as synchronous and asynchronous responses, which are common in the biological brain. Compared to Hopfield and Izhikevich networks in content-addressable memory applications, memcapacitive networks successfully retrieved correct memory patterns with high accuracy, even for distorted inputs of up to 40%. The simulation results illustrate that the novel model of the memcapacitive spiking neuron offers a potential advancement in implementing artificial spiking neurons with high energy efficiency, bringing a step closer to mimicking biological neurons.
{"title":"Memcapacitive Spiking Neurons and Associative Memory Application","authors":"S. J. Dat Tran","doi":"10.1109/ACCESS.2025.3549357","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3549357","url":null,"abstract":"The Hodgkin and Huxley neuron model describes the complex behavior of biological neurons. However, due to the complexity of these computations, the Hodgkin and Huxley models are impractical for use in large-scale networks. In contrast, Izhikevich introduced a simpler model capable of producing various firing patterns typical of cortical neurons. This study proposes a novel model of memcapacitive-based neurons that offers a potential implementation of spiking neurons with energy efficiency due to the inherent storage nature of memcapacitive devices. The findings demonstrate that memcapacitive neurons can produce 23 firing patterns similar to Izhikevich neurons but at significantly higher firing rates. Memcapacitive neurons exhibit firing patterns associated with excitatory, inhibitory, and thalamocortical neurons. Similar to Izhikevich neurons, pulse-coupled neural networks of memcapacitive neurons display collective behaviors, such as synchronous and asynchronous responses, which are common in the biological brain. Compared to Hopfield and Izhikevich networks in content-addressable memory applications, memcapacitive networks successfully retrieved correct memory patterns with high accuracy, even for distorted inputs of up to 40%. The simulation results illustrate that the novel model of the memcapacitive spiking neuron offers a potential advancement in implementing artificial spiking neurons with high energy efficiency, bringing a step closer to mimicking biological neurons.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"43933-43946"},"PeriodicalIF":3.4,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10916667","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143619949","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-03-06DOI: 10.1109/ACCESS.2025.3549417
Maria Trigka;Elias Dritsas
Cardiovascular disease (CVD) ranks among the top causes of mortality globally, underscoring the urgent necessity for advanced predictive models to enhance early detection and preventative measures. In this direction, this study investigates the performance of five well-established deep learning (DL) models, namely Multilayer Perceptron (MLP), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Autoencoder in predicting CVD using a diverse patient dataset. To tackle the prevalent class imbalance issue in medical datasets, we introduce an enhanced Synthetic Minority Over-sampling Technique (SMOTE). This innovative technique enhances traditional SMOTE by incorporating feature correlations to produce more realistic synthetic samples. We compare model performance across three scenarios: without SMOTE, with traditional SMOTE, and with enhanced SMOTE, using metrics such as Accuracy, Precision, Recall, F1-Score, and Area Under the Curve (AUC). Our results show that the enhanced SMOTE significantly improves model performance, especially in recall and AUC-ROC. Notably, the CNN model with enhanced SMOTE prevailed, achieving the highest overall performance with an AUC of 0.90, an Accuracy of 0.91, a Precision of 0.89, a Recall of 0.86, and an F1-Score equal to 0.87, making it the most effective model in this study. This research highlights the potential of the enhanced SMOTE in developing robust predictive models for CVD, with broader implications for healthcare analytics.
{"title":"Improving Cardiovascular Disease Prediction With Deep Learning and Correlation-Aware SMOTE","authors":"Maria Trigka;Elias Dritsas","doi":"10.1109/ACCESS.2025.3549417","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3549417","url":null,"abstract":"Cardiovascular disease (CVD) ranks among the top causes of mortality globally, underscoring the urgent necessity for advanced predictive models to enhance early detection and preventative measures. In this direction, this study investigates the performance of five well-established deep learning (DL) models, namely Multilayer Perceptron (MLP), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Autoencoder in predicting CVD using a diverse patient dataset. To tackle the prevalent class imbalance issue in medical datasets, we introduce an enhanced Synthetic Minority Over-sampling Technique (SMOTE). This innovative technique enhances traditional SMOTE by incorporating feature correlations to produce more realistic synthetic samples. We compare model performance across three scenarios: without SMOTE, with traditional SMOTE, and with enhanced SMOTE, using metrics such as Accuracy, Precision, Recall, F1-Score, and Area Under the Curve (AUC). Our results show that the enhanced SMOTE significantly improves model performance, especially in recall and AUC-ROC. Notably, the CNN model with enhanced SMOTE prevailed, achieving the highest overall performance with an AUC of 0.90, an Accuracy of 0.91, a Precision of 0.89, a Recall of 0.86, and an F1-Score equal to 0.87, making it the most effective model in this study. This research highlights the potential of the enhanced SMOTE in developing robust predictive models for CVD, with broader implications for healthcare analytics.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"44590-44606"},"PeriodicalIF":3.4,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10916648","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143621555","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-03-06DOI: 10.1109/ACCESS.2025.3548702
Chaejin Lim;Junhee Hyeon;Kiseong Lee;Dongil Han
Many existing deep learning models suffer from a fundamental limitation, where they often misclassify Out-of-Distribution (OOD) data as In-Distribution (ID) data. OOD data represent data patterns that differ from the training distribution, such as images of unseen classes or different domains. This misclassification occurs because deep learning models are inherently designed to classify inputs into one of their known categories. To address this limitation, we propose Feature-Label Negative Sampling (FLaNS), which exploits the observation that OOD data inherently exhibit mismatches between features and their assigned labels. Our method constructs negative data by deliberately creating feature-label mismatches from ID data, which naturally deviate from the learned ID distribution. These synthetic data enable the Support Vector Machine (SVM) to learn decision boundaries that discriminate between ID and OOD data without requiring actual OOD data during training. We evaluated our method on the NCT-CRC-HE and CIFAR-100 datasets using two backbone models. Our approach demonstrates stable and reliable performance in both Receiver Operating Characteristics (ROC) Curve Area (AUROC) and False Positive Rate at 95% True Positive Rate (FPR95) metrics compared to existing methods. This research enhances OOD detection by introducing a method that leverages feature-label mismatches in ID data, improving detection performance without needing diverse OOD data, which are often difficult to collect.
{"title":"FLaNS: Feature-Label Negative Sampling for Out-of-Distribution Detection","authors":"Chaejin Lim;Junhee Hyeon;Kiseong Lee;Dongil Han","doi":"10.1109/ACCESS.2025.3548702","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3548702","url":null,"abstract":"Many existing deep learning models suffer from a fundamental limitation, where they often misclassify Out-of-Distribution (OOD) data as In-Distribution (ID) data. OOD data represent data patterns that differ from the training distribution, such as images of unseen classes or different domains. This misclassification occurs because deep learning models are inherently designed to classify inputs into one of their known categories. To address this limitation, we propose Feature-Label Negative Sampling (FLaNS), which exploits the observation that OOD data inherently exhibit mismatches between features and their assigned labels. Our method constructs negative data by deliberately creating feature-label mismatches from ID data, which naturally deviate from the learned ID distribution. These synthetic data enable the Support Vector Machine (SVM) to learn decision boundaries that discriminate between ID and OOD data without requiring actual OOD data during training. We evaluated our method on the NCT-CRC-HE and CIFAR-100 datasets using two backbone models. Our approach demonstrates stable and reliable performance in both Receiver Operating Characteristics (ROC) Curve Area (AUROC) and False Positive Rate at 95% True Positive Rate (FPR95) metrics compared to existing methods. This research enhances OOD detection by introducing a method that leverages feature-label mismatches in ID data, improving detection performance without needing diverse OOD data, which are often difficult to collect.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"43878-43888"},"PeriodicalIF":3.4,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10915589","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143621633","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-03-06DOI: 10.1109/ACCESS.2025.3548785
Huan Li;Hongyuan Yuan;Shi He;Yu Zhou
A battery management system design and test scheme are proposed to meet the test requirements for high-precision state-of-energy (SOE) calculation in energy storage systems and enable joint debugging with the power system. The proposed system’s hardware and software design, along with the testing scheme using RTDS as the main control logic, are presented. In this design, the SOE value of the single unit can be calculated using the extended Kalman algorithm and other correction methods in the energy storage system with multiple strings. The experimental results show that the SOE error is 1.5% under the condition of constant power test at room temperature. Additionally, the RTDS battery model simulates battery data, and the high-pressure box control logic is integrated into RTDS to enable direct control of the RTDS battery cluster output voltage via the battery management system.
{"title":"Design of Battery Management System for Grid Energy Storage Based on FreeRTOS and RTDS Testing System","authors":"Huan Li;Hongyuan Yuan;Shi He;Yu Zhou","doi":"10.1109/ACCESS.2025.3548785","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3548785","url":null,"abstract":"A battery management system design and test scheme are proposed to meet the test requirements for high-precision state-of-energy (SOE) calculation in energy storage systems and enable joint debugging with the power system. The proposed system’s hardware and software design, along with the testing scheme using RTDS as the main control logic, are presented. In this design, the SOE value of the single unit can be calculated using the extended Kalman algorithm and other correction methods in the energy storage system with multiple strings. The experimental results show that the SOE error is 1.5% under the condition of constant power test at room temperature. Additionally, the RTDS battery model simulates battery data, and the high-pressure box control logic is integrated into RTDS to enable direct control of the RTDS battery cluster output voltage via the battery management system.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"43414-43423"},"PeriodicalIF":3.4,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10915661","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143621859","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-03-06DOI: 10.1109/ACCESS.2025.3548854
Shu-Ming Tseng;Sz-Tze Wen;Chao Fang;Mehdi Norouzi
The massive connectivity trend was set to shape B5G/6G networks. Device-to-device (D2D) communications play a crucial role in massive connectivity in the context of Internet of Things (IoT) applications. Recently, a heterogeneous interference graph neural network (HIGNN) was proposed for resource allocation in heterogeneous networks. The HIGNN captured the spatial information hidden in heterogeneous network topology and was scalable. However, existing methods primarily focused on resource allocation at the physical layer only and did not adequately address the cross-layer issues involved in video transmission. Therefore, in this paper, we propose the video-optimized heterogeneous interference graph neural network (VD-HIGNN) as a cross-layer D2D resource allocation method for video transmission, which introduces the following contributions: 1) joint source encoder rate and beamforming/power control, 2) incorporating video rate distortion function parameters from the application layer into the node features, and 3) changing the loss function from data rate to Peak-Signal-to-Noise-Ratio (PSNR), a function of video rate distortion and a metric of video quality. Simulation results demonstrate that our proposed VD-HIGNN outperforms two physical layer baseline schemes: the iterative fractional programming method by 0.53 dB and HIGNN by approximately 2 dB for video transmission. Moreover, when scaled to larger problems with 2-12 times the number of nodes within a fixed area size, the VD-HIGNN achieves 94% or more of the performance of a retrained model, showcasing its scalability and generalization ability.
{"title":"Cross Layer Power Allocation by Graph Neural Networks in Heterogeneous D2D Video Communications","authors":"Shu-Ming Tseng;Sz-Tze Wen;Chao Fang;Mehdi Norouzi","doi":"10.1109/ACCESS.2025.3548854","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3548854","url":null,"abstract":"The massive connectivity trend was set to shape B5G/6G networks. Device-to-device (D2D) communications play a crucial role in massive connectivity in the context of Internet of Things (IoT) applications. Recently, a heterogeneous interference graph neural network (HIGNN) was proposed for resource allocation in heterogeneous networks. The HIGNN captured the spatial information hidden in heterogeneous network topology and was scalable. However, existing methods primarily focused on resource allocation at the physical layer only and did not adequately address the cross-layer issues involved in video transmission. Therefore, in this paper, we propose the video-optimized heterogeneous interference graph neural network (VD-HIGNN) as a cross-layer D2D resource allocation method for video transmission, which introduces the following contributions: 1) joint source encoder rate and beamforming/power control, 2) incorporating video rate distortion function parameters from the application layer into the node features, and 3) changing the loss function from data rate to Peak-Signal-to-Noise-Ratio (PSNR), a function of video rate distortion and a metric of video quality. Simulation results demonstrate that our proposed VD-HIGNN outperforms two physical layer baseline schemes: the iterative fractional programming method by 0.53 dB and HIGNN by approximately 2 dB for video transmission. Moreover, when scaled to larger problems with 2-12 times the number of nodes within a fixed area size, the VD-HIGNN achieves 94% or more of the performance of a retrained model, showcasing its scalability and generalization ability.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"44484-44496"},"PeriodicalIF":3.4,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10915671","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143621704","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-03-06DOI: 10.1109/ACCESS.2025.3548881
Shuting Guo;Daniel N. Aloi;Jia Li;Hongmei Zhao
Motivated by the evolution of vehicular technologies and applications, the vehicle-to-everything (V2X) communications can be realized by the dedicated short-range communications (DSRC) and cellular V2X (C-V2X) which are undergoing continuous and widespread development. As the first standardized DSRC technology, IEEE 802.11p, that has been studied with large-scale field trials performed worldwide, is more mature and robust than the other V2X technologies. The main contributions of the proposed work over previous work are listed as follows. Firstly, to break the limitation of partial physical layer (PHY) evaluation, extensive PHY metrics, which include the packet error rate (PER), packet reception ratio (PRR), output packet inter-arrival time (IAT), and output effective data rate, are adequately employed to fulfill complete PHY evaluation. Secondly, to avoid incomplete analysis on antenna configurations, various multi-antenna configurations, containing the multiple-input multiple-output (MIMO), single-input multiple-output (SIMO), and multiple-input single-output (MISO) systems, are involved together with the single-input single-output (SISO) configuration to realize comprehensive analysis on diverse antenna configurations. Finally, to overcome unobvious exhibition on effect of parameters on PHY performance, considerably different packet sizes and modulation and coding schemes (MCSs) are investigated under the urban non-line-of-sight (NLOS) and highway NLOS scenarios to disclose the deep impact of each parameter. Important conclusions from a thorough MATLAB-based PHY simulation are summarized as follows. Firstly, in comparison with the SISO system, the multi-antenna systems are more favorable in reducing the PER, increasing the PRR and transmission coverage, decreasing the output packet IAT, and elevating the output effective data rate, below the signal-to-noise ratio (SNR) threshold and above the distance threshold. Secondly, the packet size and the MCS should be determined suitably to adapt to the high-reliability, low-latency, or high-throughput requirement in different applications. Finally, compared to the highway NLOS scenario with higher Doppler effect, the urban NLOS scenario is more tolerant to the larger packet and higher MCS in the vehicle-to-vehicle (V2V) communications with its lower PER, larger PRR and transmission coverage, smaller output packet IAT, and higher output effective data rate.
{"title":"Physical Layer Evaluation on IEEE 802.11p With Different Configurations in NLOS Scenarios for V2V Communications","authors":"Shuting Guo;Daniel N. Aloi;Jia Li;Hongmei Zhao","doi":"10.1109/ACCESS.2025.3548881","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3548881","url":null,"abstract":"Motivated by the evolution of vehicular technologies and applications, the vehicle-to-everything (V2X) communications can be realized by the dedicated short-range communications (DSRC) and cellular V2X (C-V2X) which are undergoing continuous and widespread development. As the first standardized DSRC technology, IEEE 802.11p, that has been studied with large-scale field trials performed worldwide, is more mature and robust than the other V2X technologies. The main contributions of the proposed work over previous work are listed as follows. Firstly, to break the limitation of partial physical layer (PHY) evaluation, extensive PHY metrics, which include the packet error rate (PER), packet reception ratio (PRR), output packet inter-arrival time (IAT), and output effective data rate, are adequately employed to fulfill complete PHY evaluation. Secondly, to avoid incomplete analysis on antenna configurations, various multi-antenna configurations, containing the multiple-input multiple-output (MIMO), single-input multiple-output (SIMO), and multiple-input single-output (MISO) systems, are involved together with the single-input single-output (SISO) configuration to realize comprehensive analysis on diverse antenna configurations. Finally, to overcome unobvious exhibition on effect of parameters on PHY performance, considerably different packet sizes and modulation and coding schemes (MCSs) are investigated under the urban non-line-of-sight (NLOS) and highway NLOS scenarios to disclose the deep impact of each parameter. Important conclusions from a thorough MATLAB-based PHY simulation are summarized as follows. Firstly, in comparison with the SISO system, the multi-antenna systems are more favorable in reducing the PER, increasing the PRR and transmission coverage, decreasing the output packet IAT, and elevating the output effective data rate, below the signal-to-noise ratio (SNR) threshold and above the distance threshold. Secondly, the packet size and the MCS should be determined suitably to adapt to the high-reliability, low-latency, or high-throughput requirement in different applications. Finally, compared to the highway NLOS scenario with higher Doppler effect, the urban NLOS scenario is more tolerant to the larger packet and higher MCS in the vehicle-to-vehicle (V2V) communications with its lower PER, larger PRR and transmission coverage, smaller output packet IAT, and higher output effective data rate.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"44428-44444"},"PeriodicalIF":3.4,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10915594","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143621630","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-03-06DOI: 10.1109/ACCESS.2025.3548649
Sonalika Mishra;Preeti Ranjan Sahu;Ramesh Chandra Prusty;Sidhartha Panda;Taha Selim Ustun;Ahmet Onen
The frequency control of an islanded microgrid (MG) is a challenging task due to the lack of system inertia as it is highly penetrated with renewable energy sources (RESs). Current work suggests overcoming this issue with an energy storage system (ESS)-based virtual inertia (VI) approach by providing appropriate proportional damping instead of a fixed value. In this study to overcome the frequency control issue, a fuzzy-based self-adaptive enhanced VI controller (SAEVIC) coordinated with electric vehicles (EV) is proposed. The controller is proposed to stabilize the system frequency and balance state of charge (SOC) of plugged-in electric vehicles (EVs). The performance of the proposed controller is justified in terms of frequency control over with/without conventional VI control, conventional enhanced VI control, and self-adaptive VI control. The system frequency and SOC signal are considered for the control action of the proposed controller. The impact of EV integration on the system frequency dynamics is tested. The validation of the proposed controller is carried out with a system injected with stochastic disturbances, high and low levels of renewable energies, denial of service attacks on renewable energy, and disturbed operating conditions with varied internal parameters. It is noticed that with the SAEVIC approach, the overshoot (OS)-11.40%, undershoot (US)- 46.46%, settling time (ST)-98.6% and fitness value-10.27% are decreased as compared to conventional enhanced VI approach under Stochastic variations of wind, PV, and multi-step load disturbance of MG system.
{"title":"Application of Enhanced Self-Adaptive Virtual Inertia Control for Efficient Frequency Control of Renewable Energy-Based Microgrid System Integrated With Electric Vehicles","authors":"Sonalika Mishra;Preeti Ranjan Sahu;Ramesh Chandra Prusty;Sidhartha Panda;Taha Selim Ustun;Ahmet Onen","doi":"10.1109/ACCESS.2025.3548649","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3548649","url":null,"abstract":"The frequency control of an islanded microgrid (MG) is a challenging task due to the lack of system inertia as it is highly penetrated with renewable energy sources (RESs). Current work suggests overcoming this issue with an energy storage system (ESS)-based virtual inertia (VI) approach by providing appropriate proportional damping instead of a fixed value. In this study to overcome the frequency control issue, a fuzzy-based self-adaptive enhanced VI controller (SAEVIC) coordinated with electric vehicles (EV) is proposed. The controller is proposed to stabilize the system frequency and balance state of charge (SOC) of plugged-in electric vehicles (EVs). The performance of the proposed controller is justified in terms of frequency control over with/without conventional VI control, conventional enhanced VI control, and self-adaptive VI control. The system frequency and SOC signal are considered for the control action of the proposed controller. The impact of EV integration on the system frequency dynamics is tested. The validation of the proposed controller is carried out with a system injected with stochastic disturbances, high and low levels of renewable energies, denial of service attacks on renewable energy, and disturbed operating conditions with varied internal parameters. It is noticed that with the SAEVIC approach, the overshoot (OS)-11.40%, undershoot (US)- 46.46%, settling time (ST)-98.6% and fitness value-10.27% are decreased as compared to conventional enhanced VI approach under Stochastic variations of wind, PV, and multi-step load disturbance of MG system.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"43520-43531"},"PeriodicalIF":3.4,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10914575","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143621640","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：