Pub Date : 2024-09-11DOI: 10.26599/TST.2024.9010071
Shabeer Ahmad;Jinling Zhang;Ali Nauman;Adil Khan;Khizar Abbas;Babar Hayat
The rise of innovative applications, like online gaming, smart healthcare, and Internet of Things (IoT) services, has increased demand for high data rates and seamless connectivity, posing challenges for Beyond 5G (B5G) networks. There is a need for cost-effective solutions to enhance spectral efficiency in densely populated areas, ensuring higher data rates and uninterrupted connectivity while minimizing costs. Unmanned Aerial Vehicles (UAVs) as Aerial Base Stations (ABSs) offer a promising and cost-effective solution to boost network capacity, especially during emergencies and high-data-rate demands. Nevertheless, integrating UAVs into the B5G networks presents new challenges, including resource scarcity, energy efficiency, resource allocation, optimal power transmission control, and maximizing overall throughput. This paper presents a UAV-assisted B5G communication system where UAVs act as ABSs, and introduces the Deep Reinforcement Learning (DRL) based Energy Efficient Resource Allocation (Deep-EERA) mechanism. An efficient DRL-based Deep Deterministic Policy Gradient (DDPG) mechanism is introduced for optimal resource allocation with the twin goals of energy efficiency and average throughput maximization. The proposed Deep-EERA method learns optimal policies to conserve energy and enhance throughput within the dynamic and complex UAV-empowered B5G environment. Through extensive simulations, we validate the performance of the proposed approach, demonstrating that it outperforms other baseline methods in energy efficiency and throughput maximization.
{"title":"Deep-EERA: DRL-Based Energy-Efficient Resource Allocation in UAV-Empowered Beyond 5G Networks","authors":"Shabeer Ahmad;Jinling Zhang;Ali Nauman;Adil Khan;Khizar Abbas;Babar Hayat","doi":"10.26599/TST.2024.9010071","DOIUrl":"https://doi.org/10.26599/TST.2024.9010071","url":null,"abstract":"The rise of innovative applications, like online gaming, smart healthcare, and Internet of Things (IoT) services, has increased demand for high data rates and seamless connectivity, posing challenges for Beyond 5G (B5G) networks. There is a need for cost-effective solutions to enhance spectral efficiency in densely populated areas, ensuring higher data rates and uninterrupted connectivity while minimizing costs. Unmanned Aerial Vehicles (UAVs) as Aerial Base Stations (ABSs) offer a promising and cost-effective solution to boost network capacity, especially during emergencies and high-data-rate demands. Nevertheless, integrating UAVs into the B5G networks presents new challenges, including resource scarcity, energy efficiency, resource allocation, optimal power transmission control, and maximizing overall throughput. This paper presents a UAV-assisted B5G communication system where UAVs act as ABSs, and introduces the Deep Reinforcement Learning (DRL) based Energy Efficient Resource Allocation (Deep-EERA) mechanism. An efficient DRL-based Deep Deterministic Policy Gradient (DDPG) mechanism is introduced for optimal resource allocation with the twin goals of energy efficiency and average throughput maximization. The proposed Deep-EERA method learns optimal policies to conserve energy and enhance throughput within the dynamic and complex UAV-empowered B5G environment. Through extensive simulations, we validate the performance of the proposed approach, demonstrating that it outperforms other baseline methods in energy efficiency and throughput maximization.","PeriodicalId":48690,"journal":{"name":"Tsinghua Science and Technology","volume":"30 1","pages":"418-432"},"PeriodicalIF":6.6,"publicationDate":"2024-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10676362","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142169652","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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}
Indoor localization has been challenging research due to the invalidity of the global navigation satellite system in indoor scenarios. Recent advances in ambient assistive living have shown great power in detecting and locating persons living in their homes, especially using the ON/OFF binary sensors. In this paper, we exploit the Bluetooth low-energy beacons as device-based binary anchors under the lowest transmission power to turn any indoor activity and facility interaction into a binary location indicator. The binary anchors are fused with an extended Kalman filter based pedestrian dead-reckoning using a factor graph optimization, with extra constraints including the normalized magnetic loop closure which is optimized using an attenuation factor, and a rapidly-exploring random tree-based map collision validation. The proposed system provides a cost-effective, scalable, and robust localization for common indoor scenarios. The experimental results show an effective sub-meter precision for the long-term trajectories, and a small amount of anchors is enough for significant calibration in large scenarios.
{"title":"BASIL: Binary Anchor-Based Smart Indoor Localization","authors":"Zhe Yang;Yanjun Li;Yufan Zhang;Yun Pan;Chung Shue Chen;Yi-hua Zhu","doi":"10.26599/TST.2024.9010008","DOIUrl":"https://doi.org/10.26599/TST.2024.9010008","url":null,"abstract":"Indoor localization has been challenging research due to the invalidity of the global navigation satellite system in indoor scenarios. Recent advances in ambient assistive living have shown great power in detecting and locating persons living in their homes, especially using the ON/OFF binary sensors. In this paper, we exploit the Bluetooth low-energy beacons as device-based binary anchors under the lowest transmission power to turn any indoor activity and facility interaction into a binary location indicator. The binary anchors are fused with an extended Kalman filter based pedestrian dead-reckoning using a factor graph optimization, with extra constraints including the normalized magnetic loop closure which is optimized using an attenuation factor, and a rapidly-exploring random tree-based map collision validation. The proposed system provides a cost-effective, scalable, and robust localization for common indoor scenarios. The experimental results show an effective sub-meter precision for the long-term trajectories, and a small amount of anchors is enough for significant calibration in large scenarios.","PeriodicalId":48690,"journal":{"name":"Tsinghua Science and Technology","volume":"30 1","pages":"1-17"},"PeriodicalIF":6.6,"publicationDate":"2024-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10676357","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142169637","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 : 2024-09-11DOI: 10.26599/TST.2024.9010001
Zixiang Xian;Rubing Huang;Dave Towey;Chuan Yue
Although Convolutional Neural Networks (CNNs) have achieved remarkable success in image classification, most CNNs use image datasets in the Red-Green-Blue (RGB) color space (one of the most commonly used color spaces). The existing literature regarding the influence of color space use on the performance of CNNs is limited. This paper explores the impact of different color spaces on image classification using CNNs. We compare the performance of five CNN models with different convolution operations and numbers of layers on four image datasets, each converted to nine color spaces. We find that color space selection can significantly affect classification accuracy, and that some classes are more sensitive to color space changes than others. Different color spaces may have different expression abilities for different image features, such as brightness, saturation, hue, etc. To leverage the complementary information from different color spaces, we propose a pseudo-Siamese network that fuses two color spaces without modifying the network architecture. Our experiments show that our proposed model can outperform the single-color-space models on most datasets. We also find that our method is simple, flexible, and compatible with any CNN and image dataset.
{"title":"Convolutional Neural Network Image Classification Based on Different Color Spaces","authors":"Zixiang Xian;Rubing Huang;Dave Towey;Chuan Yue","doi":"10.26599/TST.2024.9010001","DOIUrl":"https://doi.org/10.26599/TST.2024.9010001","url":null,"abstract":"Although Convolutional Neural Networks (CNNs) have achieved remarkable success in image classification, most CNNs use image datasets in the Red-Green-Blue (RGB) color space (one of the most commonly used color spaces). The existing literature regarding the influence of color space use on the performance of CNNs is limited. This paper explores the impact of different color spaces on image classification using CNNs. We compare the performance of five CNN models with different convolution operations and numbers of layers on four image datasets, each converted to nine color spaces. We find that color space selection can significantly affect classification accuracy, and that some classes are more sensitive to color space changes than others. Different color spaces may have different expression abilities for different image features, such as brightness, saturation, hue, etc. To leverage the complementary information from different color spaces, we propose a pseudo-Siamese network that fuses two color spaces without modifying the network architecture. Our experiments show that our proposed model can outperform the single-color-space models on most datasets. We also find that our method is simple, flexible, and compatible with any CNN and image dataset.","PeriodicalId":48690,"journal":{"name":"Tsinghua Science and Technology","volume":"30 1","pages":"402-417"},"PeriodicalIF":6.6,"publicationDate":"2024-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10676405","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142169647","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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}
Recent years have seen growing demand for the use of edge computing to achieve the full potential of the Internet of Things (IoTs), given that various IoT systems have been generating big data to facilitate modern latency-sensitive applications. Network Dismantling (ND), which is a basic problem, attempts to find an optimal set of nodes that will maximize the connectivity degradation in a network. However, current approaches mainly focus on simple networks that model only pairwise interactions between two nodes, whereas higher-order groupwise interactions among an arbitrary number of nodes are ubiquitous in the real world, which can be better modeled as hypernetwork. The structural difference between a simple and a hypernetwork restricts the direct application of simple ND methods to a hypernetwork. Although some hypernetwork centrality measures (e.g., betweenness) can be used for hypernetwork dismantling, they face the problem of balancing effectiveness and efficiency. Therefore, we propose a betweenness approximation-based hypernetwork dismantling method with a Hypergraph Neural Network (HNN). The proposed approach, called “HND”, trains a transferable HNN-based regression model on plenty of generated small-scale synthetic hypernetworks in a supervised way, utilizing the well-trained model to approximate the betweenness of the nodes. Extensive experiments on five actual hypernetworks demonstrate the effectiveness and efficiency of HND compared with various baselines.
{"title":"Betweenness Approximation for Edge Computing with Hypergraph Neural Networks","authors":"Yaguang Guo;Wenxin Xie;Qingren Wang;Dengcheng Yan;Yiwen Zhang","doi":"10.26599/TST.2023.9010106","DOIUrl":"https://doi.org/10.26599/TST.2023.9010106","url":null,"abstract":"Recent years have seen growing demand for the use of edge computing to achieve the full potential of the Internet of Things (IoTs), given that various IoT systems have been generating big data to facilitate modern latency-sensitive applications. Network Dismantling (ND), which is a basic problem, attempts to find an optimal set of nodes that will maximize the connectivity degradation in a network. However, current approaches mainly focus on simple networks that model only pairwise interactions between two nodes, whereas higher-order groupwise interactions among an arbitrary number of nodes are ubiquitous in the real world, which can be better modeled as hypernetwork. The structural difference between a simple and a hypernetwork restricts the direct application of simple ND methods to a hypernetwork. Although some hypernetwork centrality measures (e.g., betweenness) can be used for hypernetwork dismantling, they face the problem of balancing effectiveness and efficiency. Therefore, we propose a betweenness approximation-based hypernetwork dismantling method with a Hypergraph Neural Network (HNN). The proposed approach, called “HND”, trains a transferable HNN-based regression model on plenty of generated small-scale synthetic hypernetworks in a supervised way, utilizing the well-trained model to approximate the betweenness of the nodes. Extensive experiments on five actual hypernetworks demonstrate the effectiveness and efficiency of HND compared with various baselines.","PeriodicalId":48690,"journal":{"name":"Tsinghua Science and Technology","volume":"30 1","pages":"331-344"},"PeriodicalIF":6.6,"publicationDate":"2024-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10676406","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142169668","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 : 2024-09-11DOI: 10.26599/TST.2024.9010038
Hongyi Wang;Yang Li;Jing Yang;Daqiang Hu;Zhi Liao
Recent years have witnessed the widespread adoption of mobile applications (apps for short). For quality-of-service and commercial competitiveness, sufficient Graphical User Interface (GUI) testing is required to verify the robustness of the apps. Given that testing with manual efforts is time-consuming and error-prone, automated GUI testing has been widely studied. However, existing approaches mostly focus on GUI exploration while lacking attention to complex interactions with apps, especially generating appropriate text inputs like real users. In this paper, we introduce CamDroid, a lightweight context-aware automated GUI testing tool, which can efficiently explore app activities through (1) a model-based UI-guided testing strategy informed by the context of previous event-activity transitions and (2) a data-driven text input generation approach regarding the GUI context. We evaluate CamDroid on 20 widely-used apps. The results show that CamDroid outperforms non-trivial baselines in activity coverage, crash detection, and test efficiency.
{"title":"CamDroid: Context-Aware Model-Based Automated GUI Testing for Android Apps","authors":"Hongyi Wang;Yang Li;Jing Yang;Daqiang Hu;Zhi Liao","doi":"10.26599/TST.2024.9010038","DOIUrl":"https://doi.org/10.26599/TST.2024.9010038","url":null,"abstract":"Recent years have witnessed the widespread adoption of mobile applications (apps for short). For quality-of-service and commercial competitiveness, sufficient Graphical User Interface (GUI) testing is required to verify the robustness of the apps. Given that testing with manual efforts is time-consuming and error-prone, automated GUI testing has been widely studied. However, existing approaches mostly focus on GUI exploration while lacking attention to complex interactions with apps, especially generating appropriate text inputs like real users. In this paper, we introduce CamDroid, a lightweight context-aware automated GUI testing tool, which can efficiently explore app activities through (1) a model-based UI-guided testing strategy informed by the context of previous event-activity transitions and (2) a data-driven text input generation approach regarding the GUI context. We evaluate CamDroid on 20 widely-used apps. The results show that CamDroid outperforms non-trivial baselines in activity coverage, crash detection, and test efficiency.","PeriodicalId":48690,"journal":{"name":"Tsinghua Science and Technology","volume":"30 1","pages":"55-67"},"PeriodicalIF":6.6,"publicationDate":"2024-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10676359","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142169735","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 : 2024-09-11DOI: 10.26599/TST.2023.9010089
Zhixiong Wu;Yong Cui
Gait recognition has a wide range of application scenarios in the fields of intelligent security and transportation. Gait recognition currently faces challenges: inadequate feature methods for environmental interferences and insufficient local-global information correlation. To address these issues, we propose a gait recognition model based on feature fusion and dual attention. Our model utilizes the ResNet architecture as the backbone network for fundamental gait features extraction. Subsequently, the features from different network layers are passed through the feature pyramid for feature fusion, so that multi-scale local information can be fused into global information, providing a more complete feature representation. The dual attention module enhances the fused features in multiple dimensions, enabling the model to capture information from different semantics and scale information. Our model proves effective and competitive results on CASIA-B (NM: 95.6%, BG: 90.9%, CL: 73.7%) and OU-MVLP (88.1%). The results of related ablation experiments show that the model design is effective and has strong competitiveness.
{"title":"GaitFFDA: Feature Fusion and Dual Attention Gait Recognition Model","authors":"Zhixiong Wu;Yong Cui","doi":"10.26599/TST.2023.9010089","DOIUrl":"https://doi.org/10.26599/TST.2023.9010089","url":null,"abstract":"Gait recognition has a wide range of application scenarios in the fields of intelligent security and transportation. Gait recognition currently faces challenges: inadequate feature methods for environmental interferences and insufficient local-global information correlation. To address these issues, we propose a gait recognition model based on feature fusion and dual attention. Our model utilizes the ResNet architecture as the backbone network for fundamental gait features extraction. Subsequently, the features from different network layers are passed through the feature pyramid for feature fusion, so that multi-scale local information can be fused into global information, providing a more complete feature representation. The dual attention module enhances the fused features in multiple dimensions, enabling the model to capture information from different semantics and scale information. Our model proves effective and competitive results on CASIA-B (NM: 95.6%, BG: 90.9%, CL: 73.7%) and OU-MVLP (88.1%). The results of related ablation experiments show that the model design is effective and has strong competitiveness.","PeriodicalId":48690,"journal":{"name":"Tsinghua Science and Technology","volume":"30 1","pages":"345-356"},"PeriodicalIF":6.6,"publicationDate":"2024-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10676356","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142169640","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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}
The diversified development of the service ecosystem, particularly the rapid growth of services like cloud and edge computing, has propelled the flourishing expansion of the service trading market. However, in the absence of appropriate pricing guidance, service providers often devise pricing strategies solely based on their own interests, potentially hindering the maximization of overall market profits. This challenge is even more severe in edge computing scenarios, as different edge service providers are dispersed across various regions and influenced by multiple factors, making it challenging to establish a unified pricing model. This paper introduces a multi-participant stochastic game model to formalize the pricing problem of multiple edge services. Subsequently, an incentive mechanism based on Pareto improvement is proposed to drive the game towards Pareto optimal direction, achieving optimal profits. Finally, an enhanced PSO algorithm was proposed by adaptively optimizing inertia factor across three stages. This optimization significantly improved the efficiency of solving the game model and analyzed equilibrium states under various evolutionary mechanisms. Experimental results demonstrate that the proposed pricing incentive mechanism promotes more effective and rational pricing allocations, while also demonstrating the effectiveness of our algorithm in resolving game problems.
{"title":"Maximizing Overall Service Profit: Multi-Edge Service Pricing as a Stochastic Game Model","authors":"Shengye Pang;Xinkui Zhao;Jiayin Luo;Jintao Chen;Fan Wang;Jianwei Yin","doi":"10.26599/TST.2024.9010050","DOIUrl":"https://doi.org/10.26599/TST.2024.9010050","url":null,"abstract":"The diversified development of the service ecosystem, particularly the rapid growth of services like cloud and edge computing, has propelled the flourishing expansion of the service trading market. However, in the absence of appropriate pricing guidance, service providers often devise pricing strategies solely based on their own interests, potentially hindering the maximization of overall market profits. This challenge is even more severe in edge computing scenarios, as different edge service providers are dispersed across various regions and influenced by multiple factors, making it challenging to establish a unified pricing model. This paper introduces a multi-participant stochastic game model to formalize the pricing problem of multiple edge services. Subsequently, an incentive mechanism based on Pareto improvement is proposed to drive the game towards Pareto optimal direction, achieving optimal profits. Finally, an enhanced PSO algorithm was proposed by adaptively optimizing inertia factor across three stages. This optimization significantly improved the efficiency of solving the game model and analyzed equilibrium states under various evolutionary mechanisms. Experimental results demonstrate that the proposed pricing incentive mechanism promotes more effective and rational pricing allocations, while also demonstrating the effectiveness of our algorithm in resolving game problems.","PeriodicalId":48690,"journal":{"name":"Tsinghua Science and Technology","volume":"29 6","pages":"1872-1889"},"PeriodicalIF":6.6,"publicationDate":"2024-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10566008","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141435459","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 : 2024-06-20DOI: 10.26599/TST.2023.9010136
Mengqiu Zhang;Qiang Qu;Li Ning;Jianping Fan
With the widespread adoption of blockchain applications, the imperative for seamless data migration among decentralized applications has intensified. This necessity arises from various factors, including the depletion of blockchain disk space, transitions between blockchain systems, and specific requirements such as temporal data analysis. To meet these challenges and ensure the sustained functionality of applications, it is imperative to conduct time-aware cross-blockchain data migration. This process is designed to facilitate the smooth iteration of decentralized applications and the construction of a temporal index for historical data, all while preserving the integrity of the original data. In various application scenarios, this migration task may encompass the transfer of data between multiple blockchains, involving movements from one chain to another, from one chain to several chains, or from multiple chains to a single chain. However, the success of data migration hinges on the careful consideration of factors such as the reliability of the data source, data consistency, and migration efficiency. This paper introduces a time-aware cross-blockchain data migration approach tailored to accommodate diverse application scenarios, including migration between multiple chains. The proposed solution integrates a collective mechanism for controlling, executing, and storing procedures to address the complexities of data migration, incorporating elements such as transaction classification and matching. Extensive experiments have been conducted to validate the efficacy of the proposed approach.
{"title":"On Time-Aware Cross-Blockchain Data Migration","authors":"Mengqiu Zhang;Qiang Qu;Li Ning;Jianping Fan","doi":"10.26599/TST.2023.9010136","DOIUrl":"https://doi.org/10.26599/TST.2023.9010136","url":null,"abstract":"With the widespread adoption of blockchain applications, the imperative for seamless data migration among decentralized applications has intensified. This necessity arises from various factors, including the depletion of blockchain disk space, transitions between blockchain systems, and specific requirements such as temporal data analysis. To meet these challenges and ensure the sustained functionality of applications, it is imperative to conduct time-aware cross-blockchain data migration. This process is designed to facilitate the smooth iteration of decentralized applications and the construction of a temporal index for historical data, all while preserving the integrity of the original data. In various application scenarios, this migration task may encompass the transfer of data between multiple blockchains, involving movements from one chain to another, from one chain to several chains, or from multiple chains to a single chain. However, the success of data migration hinges on the careful consideration of factors such as the reliability of the data source, data consistency, and migration efficiency. This paper introduces a time-aware cross-blockchain data migration approach tailored to accommodate diverse application scenarios, including migration between multiple chains. The proposed solution integrates a collective mechanism for controlling, executing, and storing procedures to address the complexities of data migration, incorporating elements such as transaction classification and matching. Extensive experiments have been conducted to validate the efficacy of the proposed approach.","PeriodicalId":48690,"journal":{"name":"Tsinghua Science and Technology","volume":"29 6","pages":"1810-1820"},"PeriodicalIF":6.6,"publicationDate":"2024-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10566005","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141435220","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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}
In the critical field of electrical grid maintenance, ensuring the integrity of power line insulators is a primary concern. This study introduces an innovative approach for monitoring the condition of insulators using aerial surveillance via drone-mounted cameras. The proposed method is a composite deep learning framework that integrates the “You Only Look Once” version 3 (YOLO3) model with deep convolutional generative adversarial networks (DCGAN) and super-resolution generative adversarial networks (SRGAN). The YOLO3 model excels in rapidly and accurately detecting insulators, a vital step in assessing their health. Its effectiveness in distinguishing insulators against complex backgrounds enables prompt detection of defects, essential for proactive maintenance. This rapid detection is enhanced by DCGAN's precise classification and SRGAN's image quality improvement, addressing challenges posed by low-resolution drone imagery. The framework's performance was evaluated using metrics such as sensitivity, specificity, accuracy, localization accuracy, damage sensitivity, and false alarm rate. Results show that the SRGAN+DCGAN+YOLO3 model significantly outperforms existing methods, with a sensitivity of 98%, specificity of 94%, an overall accuracy of 95.6%, localization accuracy of 90%, damage sensitivity of 92%, and a reduced false alarm rate of 8%. This advanced hybrid approach not only improves the detection and classification of insulator conditions but also contributes substantially to the maintenance and health of power line insulators, thus ensuring the reliability of the electrical power grid.
{"title":"Enhancing Power Line Insulator Health Monitoring with a Hybrid Generative Adversarial Network and YOLO3 Solution","authors":"Ramakrishna Akella;Sravan Kumar Gunturi;Dipu Sarkar","doi":"10.26599/TST.2023.9010137","DOIUrl":"https://doi.org/10.26599/TST.2023.9010137","url":null,"abstract":"In the critical field of electrical grid maintenance, ensuring the integrity of power line insulators is a primary concern. This study introduces an innovative approach for monitoring the condition of insulators using aerial surveillance via drone-mounted cameras. The proposed method is a composite deep learning framework that integrates the “You Only Look Once” version 3 (YOLO3) model with deep convolutional generative adversarial networks (DCGAN) and super-resolution generative adversarial networks (SRGAN). The YOLO3 model excels in rapidly and accurately detecting insulators, a vital step in assessing their health. Its effectiveness in distinguishing insulators against complex backgrounds enables prompt detection of defects, essential for proactive maintenance. This rapid detection is enhanced by DCGAN's precise classification and SRGAN's image quality improvement, addressing challenges posed by low-resolution drone imagery. The framework's performance was evaluated using metrics such as sensitivity, specificity, accuracy, localization accuracy, damage sensitivity, and false alarm rate. Results show that the SRGAN+DCGAN+YOLO3 model significantly outperforms existing methods, with a sensitivity of 98%, specificity of 94%, an overall accuracy of 95.6%, localization accuracy of 90%, damage sensitivity of 92%, and a reduced false alarm rate of 8%. This advanced hybrid approach not only improves the detection and classification of insulator conditions but also contributes substantially to the maintenance and health of power line insulators, thus ensuring the reliability of the electrical power grid.","PeriodicalId":48690,"journal":{"name":"Tsinghua Science and Technology","volume":"29 6","pages":"1796-1809"},"PeriodicalIF":6.6,"publicationDate":"2024-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10566001","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141435221","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 : 2024-06-20DOI: 10.26599/TST.2023.9010083
Xiujun Wang;Lei Mo;Xiao Zheng;Zhe Dang
Continuously publishing histograms in data streams is crucial to many real-time applications, as it provides not only critical statistical information, but also reduces privacy leaking risk. As the importance of elements usually decreases over time in data streams, in this paper we model a data stream by a sequence of weighted sliding windows, and then study how to publish histograms over these windows continuously. The existing literature can hardly solve this problem in a real-time way, because they need to buffer all elements in each sliding window, resulting in high computational overhead and prohibitive storage burden. In this paper, we overcome this drawback by proposing an online algorithm denoted by Efficient Streaming Histogram Publishing (ESHP) to continuously publish histograms over weighted sliding windows. Specifically, our method first creates a novel sketching structure, called Approximate-Estimate Sketch (AESketch), to maintain the counting information of each histogram interval at every time instance; then, it creates histograms that satisfy the differential privacy requirement by smartly adding appropriate noise values into the sketching structure. Extensive experimental results and rigorous theoretical analysis demonstrate that the ESHP method can offer equivalent data utility with significantly lower computational overhead and storage costs when compared to other existing methods.
{"title":"Streaming Histogram Publication Over Weighted Sliding Windows Under Differential Privacy","authors":"Xiujun Wang;Lei Mo;Xiao Zheng;Zhe Dang","doi":"10.26599/TST.2023.9010083","DOIUrl":"https://doi.org/10.26599/TST.2023.9010083","url":null,"abstract":"Continuously publishing histograms in data streams is crucial to many real-time applications, as it provides not only critical statistical information, but also reduces privacy leaking risk. As the importance of elements usually decreases over time in data streams, in this paper we model a data stream by a sequence of weighted sliding windows, and then study how to publish histograms over these windows continuously. The existing literature can hardly solve this problem in a real-time way, because they need to buffer all elements in each sliding window, resulting in high computational overhead and prohibitive storage burden. In this paper, we overcome this drawback by proposing an online algorithm denoted by Efficient Streaming Histogram Publishing (ESHP) to continuously publish histograms over weighted sliding windows. Specifically, our method first creates a novel sketching structure, called Approximate-Estimate Sketch (AESketch), to maintain the counting information of each histogram interval at every time instance; then, it creates histograms that satisfy the differential privacy requirement by smartly adding appropriate noise values into the sketching structure. Extensive experimental results and rigorous theoretical analysis demonstrate that the ESHP method can offer equivalent data utility with significantly lower computational overhead and storage costs when compared to other existing methods.","PeriodicalId":48690,"journal":{"name":"Tsinghua Science and Technology","volume":"29 6","pages":"1674-1693"},"PeriodicalIF":6.6,"publicationDate":"2024-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10566026","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141435222","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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：