Pub Date : 2024-08-26DOI: 10.1109/tce.2024.3449285
Detian Zhang, Yunjun Zhou, Jin Wang
{"title":"Game Theory Based Approach for Massive Route Planning in Dynamic Road Networks","authors":"Detian Zhang, Yunjun Zhou, Jin Wang","doi":"10.1109/tce.2024.3449285","DOIUrl":"https://doi.org/10.1109/tce.2024.3449285","url":null,"abstract":"","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"2011 1","pages":""},"PeriodicalIF":4.3,"publicationDate":"2024-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142189666","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The rise of the Internet of Things (IoT) and the Internet of Vehicles (IoV) has accelerated the realization of smart cities, where cameras as interconnected consumer electronics (CE) are deployed across cities to capture target images. The widespread deployment of monitoring equipment has prompted us to focus on the target re-identification (Re-ID) issue. One major challenge about this issue is that the identified targets are often obscured by different obstacles, which leads to bad performance. In practical applications, the occluded Re-ID task is very significant to complete. Previous approaches have focused on improving the occluded Re-ID performance but have neglected the lightweight problem, which makes the model difficult to deploy in the real world. Therefore, this paper proposes a lightweight framework that ensures occluded Re-ID performance and deploys at the edge to solve the problem of long transmission time and high latency caused by wireless and cloud technology in CE. This framework tackles occluded target Re-ID issues by integrating omni-scale features with human keypoint estimation and multi-head attention mechanism (OHMA). To solve the vehicle Re-ID problem, we use the cutout method to simulate an occlusion scene due to the lack of occluded vehicle data. Then, The multi-head attention mechanism combines with the omni-scale network (OSNet) to learn vehicles subtle features. To deal with occluded pedestrians, human keypoint estimation focuses on non-occluded areas of pedestrian images by paying attention to visible information about the human body. The generated heatmaps fuse omni-scale feature maps to explore better feature representations. In addition, the HUAWEI Atlas 200I DK A2 is used to simulate real edge devices and evaluate the experiments on both public and real-world private datasets. The results demonstrate that our framework improves the occluded Re-ID performance while ensuring lightweight. Compared with the previous methods, OHMA displays advantages in occlusion scenes.
{"title":"OHMA: An Edge-Based Lightweight Occluded Target Re-Identification Framework for Exploring Abundant Feature Expression","authors":"Xiaoyu Zhang;Yichao Wang;Xiting Peng;Mianxiong Dong;Kaoru Ota;Lexi Xu","doi":"10.1109/TCE.2024.3443336","DOIUrl":"10.1109/TCE.2024.3443336","url":null,"abstract":"The rise of the Internet of Things (IoT) and the Internet of Vehicles (IoV) has accelerated the realization of smart cities, where cameras as interconnected consumer electronics (CE) are deployed across cities to capture target images. The widespread deployment of monitoring equipment has prompted us to focus on the target re-identification (Re-ID) issue. One major challenge about this issue is that the identified targets are often obscured by different obstacles, which leads to bad performance. In practical applications, the occluded Re-ID task is very significant to complete. Previous approaches have focused on improving the occluded Re-ID performance but have neglected the lightweight problem, which makes the model difficult to deploy in the real world. Therefore, this paper proposes a lightweight framework that ensures occluded Re-ID performance and deploys at the edge to solve the problem of long transmission time and high latency caused by wireless and cloud technology in CE. This framework tackles occluded target Re-ID issues by integrating omni-scale features with human keypoint estimation and multi-head attention mechanism (OHMA). To solve the vehicle Re-ID problem, we use the cutout method to simulate an occlusion scene due to the lack of occluded vehicle data. Then, The multi-head attention mechanism combines with the omni-scale network (OSNet) to learn vehicles subtle features. To deal with occluded pedestrians, human keypoint estimation focuses on non-occluded areas of pedestrian images by paying attention to visible information about the human body. The generated heatmaps fuse omni-scale feature maps to explore better feature representations. In addition, the HUAWEI Atlas 200I DK A2 is used to simulate real edge devices and evaluate the experiments on both public and real-world private datasets. The results demonstrate that our framework improves the occluded Re-ID performance while ensuring lightweight. Compared with the previous methods, OHMA displays advantages in occlusion scenes.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"70 4","pages":"7424-7435"},"PeriodicalIF":4.3,"publicationDate":"2024-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142189683","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-26DOI: 10.1109/TCE.2024.3449451
Meng Xi;Zhijing Wang;Jingyi He;Yibo Wang;Jiabao Wen;Shuai Xiao;Jiachen Yang
With the rapid development of society, a wide variety of consumer applications are increasingly emerging. At the same time, the involvement of intelligent technologies such as deep learning, reinforcement learning, and quantum computing is empowering consumer applications by driving them to be smarter, more secure, and digitized. Among them, the underwater field is an important application direction, such as equipment overhaul, scientific research, resource exploration, and so on. This paper targets the detection, optimization, and inference tasks in underwater applications, aiming to design efficient and safe solution algorithms for them using new techniques. First, we establish an underwater mission scenario, using time-varying current data to create a 3D ocean environment model, which can satisfy the requirements of different underwater applications. Second, a safe and efficient underwater object detection algorithm is designed, which constructs a deep neural network to extract valid information from redundant environments. Finally, a path planning algorithm for underwater unmanned equipment clusters is developed to solve the optimization decision problem through deep reasoning computation. We carry out a series of comparative experiments, which adequately prove that the algorithm proposed in this paper has good superiority, can cope with the interference of different intensities of ocean currents, and ensures the operational effect of the cluster formation.
{"title":"High-Precision Underwater Perception and Path Planning of AUVs Based on Quantum-Enhanced","authors":"Meng Xi;Zhijing Wang;Jingyi He;Yibo Wang;Jiabao Wen;Shuai Xiao;Jiachen Yang","doi":"10.1109/TCE.2024.3449451","DOIUrl":"10.1109/TCE.2024.3449451","url":null,"abstract":"With the rapid development of society, a wide variety of consumer applications are increasingly emerging. At the same time, the involvement of intelligent technologies such as deep learning, reinforcement learning, and quantum computing is empowering consumer applications by driving them to be smarter, more secure, and digitized. Among them, the underwater field is an important application direction, such as equipment overhaul, scientific research, resource exploration, and so on. This paper targets the detection, optimization, and inference tasks in underwater applications, aiming to design efficient and safe solution algorithms for them using new techniques. First, we establish an underwater mission scenario, using time-varying current data to create a 3D ocean environment model, which can satisfy the requirements of different underwater applications. Second, a safe and efficient underwater object detection algorithm is designed, which constructs a deep neural network to extract valid information from redundant environments. Finally, a path planning algorithm for underwater unmanned equipment clusters is developed to solve the optimization decision problem through deep reasoning computation. We carry out a series of comparative experiments, which adequately prove that the algorithm proposed in this paper has good superiority, can cope with the interference of different intensities of ocean currents, and ensures the operational effect of the cluster formation.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"70 3","pages":"5607-5617"},"PeriodicalIF":4.3,"publicationDate":"2024-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142189664","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-26DOI: 10.1109/TCE.2024.3443328
Da Guo;Zhengjie Feng;Zhen Zhang;Fazlullah Khan;Chien-Ming Chen;Ruibin Bai;Marwan Omar;Saru Kumar
Adversarial examples are security risks in the implementation of artificial intelligence (AI) in 6G Consumer Electronics. Deep learning models are highly susceptible to adversarial attacks, and defense against such attacks is critical to the safety of 6G Consumer Electronics. However, there remains a lack of effective defensive mechanisms against adversarial attacks in the realm of deep learning. The primary issue lies in the fact that it is not yet understood how adversarial examples can deceive deep learning models. The potential operation mechanism of adversarial examples has not been fully explored, which constitutes a bottleneck in adversarial attack defense. This paper focuses on causality in adversarial examples such as combining the adversarial attack algorithms with the causal inference methods. Specifically, we will use a variety of adversarial attack algorithms to generate adversarial samples, and analyze the causal relationship between adversarial samples and original samples through causal inference. At the same time, we will compare and analyze the causal effect between them to reveal the mechanism and discover the reason of miscalculating. The expected contributions of this paper include: (1) Reveal the mechanism and influencing factors of counterattack, and provide theoretical support for the security of deep learning models; (2) Propose a defense strategy based on causal inference method to provide a practical method for the defense of deep learning models; (3) Provide new ideas and methods for adversarial attack defense in deep learning models.
{"title":"Causal Effects of Adversarial Attacks on AI Models in 6G Consumer Electronics","authors":"Da Guo;Zhengjie Feng;Zhen Zhang;Fazlullah Khan;Chien-Ming Chen;Ruibin Bai;Marwan Omar;Saru Kumar","doi":"10.1109/TCE.2024.3443328","DOIUrl":"10.1109/TCE.2024.3443328","url":null,"abstract":"Adversarial examples are security risks in the implementation of artificial intelligence (AI) in 6G Consumer Electronics. Deep learning models are highly susceptible to adversarial attacks, and defense against such attacks is critical to the safety of 6G Consumer Electronics. However, there remains a lack of effective defensive mechanisms against adversarial attacks in the realm of deep learning. The primary issue lies in the fact that it is not yet understood how adversarial examples can deceive deep learning models. The potential operation mechanism of adversarial examples has not been fully explored, which constitutes a bottleneck in adversarial attack defense. This paper focuses on causality in adversarial examples such as combining the adversarial attack algorithms with the causal inference methods. Specifically, we will use a variety of adversarial attack algorithms to generate adversarial samples, and analyze the causal relationship between adversarial samples and original samples through causal inference. At the same time, we will compare and analyze the causal effect between them to reveal the mechanism and discover the reason of miscalculating. The expected contributions of this paper include: (1) Reveal the mechanism and influencing factors of counterattack, and provide theoretical support for the security of deep learning models; (2) Propose a defense strategy based on causal inference method to provide a practical method for the defense of deep learning models; (3) Provide new ideas and methods for adversarial attack defense in deep learning models.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"70 3","pages":"5804-5813"},"PeriodicalIF":4.3,"publicationDate":"2024-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142224474","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Coronavirus disease (COVID-19) has been the most challenging public health issue during the past years. The current computer-aided methods of COVID19 detection face difficulty to distinguish between COVID19 and pneumonia since they share common symptoms. Traditional methods for solving binary classification problems with COVID-19 classes are limited in their calibre to balance efficiency and accuracy. On the other hand, medical devices like reverse transcription polymerase chain reaction (RT-PCR) take longer than an hour to produce test results, and Rapid Antigen Testing (RAT) is less effective at detecting COVID-19 because it can produce false positive or false negative results. The biggest challenges here are efficiency and accuracy. To address these issues, this study introduces a novel deep multi-layer COVID19 chest X-ray based lung contamination recognition network (XCR-Net) to detect COVID-19, pneumonia, and normal individuals. Our proposed XCR-Net has been tested with five different chest X-ray datasets, having normal, COVID19, and pneumonia case chest X-ray images, and the consistency of XCR-Net has been verified by a 10-fold cross validation scheme. This multi-class study reports the class-wise and overall performance of XCR-Net, and it outperforms all other multi-class COVID-19 endeavours. Future biomedical researchers and IT professionals will be able to advance chest X-ray research with the help of the envisioned XCR-Net.
{"title":"XCR-Net: A Computer Aided Framework to Detect COVID-19","authors":"Ashik Mostafa Alvi;Md. Jubaer Khan;Nishat Tasnim Manami;Zubair Azim Miazi;Kate Wang;Siuly Siuly;Hua Wang","doi":"10.1109/TCE.2024.3446793","DOIUrl":"10.1109/TCE.2024.3446793","url":null,"abstract":"Coronavirus disease (COVID-19) has been the most challenging public health issue during the past years. The current computer-aided methods of COVID19 detection face difficulty to distinguish between COVID19 and pneumonia since they share common symptoms. Traditional methods for solving binary classification problems with COVID-19 classes are limited in their calibre to balance efficiency and accuracy. On the other hand, medical devices like reverse transcription polymerase chain reaction (RT-PCR) take longer than an hour to produce test results, and Rapid Antigen Testing (RAT) is less effective at detecting COVID-19 because it can produce false positive or false negative results. The biggest challenges here are efficiency and accuracy. To address these issues, this study introduces a novel deep multi-layer COVID19 chest X-ray based lung contamination recognition network (XCR-Net) to detect COVID-19, pneumonia, and normal individuals. Our proposed XCR-Net has been tested with five different chest X-ray datasets, having normal, COVID19, and pneumonia case chest X-ray images, and the consistency of XCR-Net has been verified by a 10-fold cross validation scheme. This multi-class study reports the class-wise and overall performance of XCR-Net, and it outperforms all other multi-class COVID-19 endeavours. Future biomedical researchers and IT professionals will be able to advance chest X-ray research with the help of the envisioned XCR-Net.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"70 4","pages":"7551-7561"},"PeriodicalIF":4.3,"publicationDate":"2024-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142224476","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-23DOI: 10.1109/tce.2024.3448541
Junaid Akram, Ali Anaissi, Awais Akram, Rajkumar Singh Rathore, Rutvij H. Jhaveri
{"title":"Adversarial Label-Flipping Attack and Defense for Anomaly Detection in Spatial Crowdsourcing UAV Services","authors":"Junaid Akram, Ali Anaissi, Awais Akram, Rajkumar Singh Rathore, Rutvij H. Jhaveri","doi":"10.1109/tce.2024.3448541","DOIUrl":"https://doi.org/10.1109/tce.2024.3448541","url":null,"abstract":"","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"11 1","pages":""},"PeriodicalIF":4.3,"publicationDate":"2024-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142189669","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-22DOI: 10.1109/TCE.2024.3445458
Siyun Xu;Miao Zhang;Tong Wang
The global Blockchain networks are growing and demand for resources is also growing respectively. The systems are switching from traditional systems to advanced systems where there is a seamless connectivity with 6G communication channels and security of data due to decentralized nature of Blockchain environment. The resources play integral part in Blockchain networks such as computational resources, data storage resources, bandwidth, sensors and energy generation power resources. The forecasting of futuristic demand of resources is important for the smooth functioning of Blockchain networks. The advanced technologies like 6G networks and machine learning techniques, Internet of Things (IoT), Digital Twins, Cyber Physical systems and AI enabled tools are playing an important role in reshaping the Blockchain networks. This research work is utilizing deep learning and game theory to map the resource requirement and to evaluate the Blockchain systems to find the potential demand for resources for smooth functioning of Blockchain enabled systems. The sampling data has been collected from Blockchain nodes and parameter based migration methods are devised to improve the predictions of deep learning models. The resource needs of the software based Blockchain networks can be predicted where the future load can be predicted on Blockchain enabled networks. The trained model based on deep learning neural networks achieves multi-layer conversion combinations through nonlinear modules to make accurate predictions in Blockchain based systems for resource requirement. This article uses the migration theory, combined with the advantages of deep neural networks to produce accurate predictions. The forecasting prediction accuracy of the required futuristic resources on raw variables is attained at 85.87%. The proposed model helps to determine the futuristic need of the resources for smooth functioning of Blockchain systems as many applications nowadays are dependent upon the Blockchain environment due to decentralized and secured nature of Blockchain networks.
{"title":"Game Theory and Deep Learning for Predicting Demand for Future Resources Within Blockchain-Networks","authors":"Siyun Xu;Miao Zhang;Tong Wang","doi":"10.1109/TCE.2024.3445458","DOIUrl":"10.1109/TCE.2024.3445458","url":null,"abstract":"The global Blockchain networks are growing and demand for resources is also growing respectively. The systems are switching from traditional systems to advanced systems where there is a seamless connectivity with 6G communication channels and security of data due to decentralized nature of Blockchain environment. The resources play integral part in Blockchain networks such as computational resources, data storage resources, bandwidth, sensors and energy generation power resources. The forecasting of futuristic demand of resources is important for the smooth functioning of Blockchain networks. The advanced technologies like 6G networks and machine learning techniques, Internet of Things (IoT), Digital Twins, Cyber Physical systems and AI enabled tools are playing an important role in reshaping the Blockchain networks. This research work is utilizing deep learning and game theory to map the resource requirement and to evaluate the Blockchain systems to find the potential demand for resources for smooth functioning of Blockchain enabled systems. The sampling data has been collected from Blockchain nodes and parameter based migration methods are devised to improve the predictions of deep learning models. The resource needs of the software based Blockchain networks can be predicted where the future load can be predicted on Blockchain enabled networks. The trained model based on deep learning neural networks achieves multi-layer conversion combinations through nonlinear modules to make accurate predictions in Blockchain based systems for resource requirement. This article uses the migration theory, combined with the advantages of deep neural networks to produce accurate predictions. The forecasting prediction accuracy of the required futuristic resources on raw variables is attained at 85.87%. The proposed model helps to determine the futuristic need of the resources for smooth functioning of Blockchain systems as many applications nowadays are dependent upon the Blockchain environment due to decentralized and secured nature of Blockchain networks.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"70 4","pages":"6997-7006"},"PeriodicalIF":4.3,"publicationDate":"2024-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142189667","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-21DOI: 10.1109/TCE.2024.3445290
Sujit Bebortta;Subhranshu Sekhar Tripathy;Surbhi Bhatia Khan;Maryam M. Al Dabel;Ahlam Almusharraf;Ali Kashif Bashir
Recently, there has been a rise in the use of Unmanned Areal Vehicles (UAVs) in consumer electronics, particularly for the critical situations. Internet of Things (IoT) technology and the accessibility of inexpensive edge computing devices present novel prospects for enhanced functionality in various domains through the utilization of IoT-based UAVs. One major difficulty of this perspective is the challenges of computation offloading between resource-constrained edge devices, and UAVs. This paper proposes an innovative framework to solve the computation offloading problem using a multi-objective Deep reinforcement learning (DRL) technique. The proposed approach helps in finding a balance between delays and energy consumption by using the concept of Tiny Machine Learning (TinyML). It develops a low complexity frameworks that make it feasible for offloading tasks to edge devices. Catering to the dynamic nature of edge-based UAV networks, TinyDeepUAV suggests a vector reinforcement that can change weights dynamically based on various user preferences. It is further conjectured that the structure can be enhanced by Double Dueling Deep Q Network (D3QN) for optimal improvement of the optimization problem. The simulation results depicts a trade-off between delay and energy consumption, enabling more effective offloading decisions while outperforming benchmark approaches.
最近，在消费电子产品中，特别是在关键情况下，无人驾驶区域车辆（uav）的使用有所增加。物联网（IoT）技术和廉价边缘计算设备的可及性为利用基于物联网的无人机在各个领域增强功能提供了新的前景。这种观点的一个主要困难是在资源受限的边缘设备和无人机之间进行计算卸载的挑战。本文提出了一种利用多目标深度强化学习（DRL）技术解决计算卸载问题的创新框架。提出的方法有助于通过使用微型机器学习（TinyML）的概念在延迟和能耗之间找到平衡。它开发了一个低复杂性的框架，使得将任务卸载到边缘设备是可行的。为了迎合基于边缘的无人机网络的动态特性，TinyDeepUAV提出了一种向量强化，可以根据各种用户偏好动态改变权重。进一步推测双Dueling Deep Q Network （D3QN）可以对结构进行增强，从而对优化问题进行优化改进。仿真结果描述了延迟和能耗之间的权衡，在优于基准方法的同时实现了更有效的卸载决策。
{"title":"TinyDeepUAV: A Tiny Deep Reinforcement Learning Framework for UAV Task Offloading in Edge-Based Consumer Electronics","authors":"Sujit Bebortta;Subhranshu Sekhar Tripathy;Surbhi Bhatia Khan;Maryam M. Al Dabel;Ahlam Almusharraf;Ali Kashif Bashir","doi":"10.1109/TCE.2024.3445290","DOIUrl":"10.1109/TCE.2024.3445290","url":null,"abstract":"Recently, there has been a rise in the use of Unmanned Areal Vehicles (UAVs) in consumer electronics, particularly for the critical situations. Internet of Things (IoT) technology and the accessibility of inexpensive edge computing devices present novel prospects for enhanced functionality in various domains through the utilization of IoT-based UAVs. One major difficulty of this perspective is the challenges of computation offloading between resource-constrained edge devices, and UAVs. This paper proposes an innovative framework to solve the computation offloading problem using a multi-objective Deep reinforcement learning (DRL) technique. The proposed approach helps in finding a balance between delays and energy consumption by using the concept of Tiny Machine Learning (TinyML). It develops a low complexity frameworks that make it feasible for offloading tasks to edge devices. Catering to the dynamic nature of edge-based UAV networks, TinyDeepUAV suggests a vector reinforcement that can change weights dynamically based on various user preferences. It is further conjectured that the structure can be enhanced by Double Dueling Deep Q Network (D3QN) for optimal improvement of the optimization problem. The simulation results depicts a trade-off between delay and energy consumption, enabling more effective offloading decisions while outperforming benchmark approaches.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"70 4","pages":"7357-7364"},"PeriodicalIF":4.3,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142189672","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The data transfer volume is massive in next-generation Wireless Sensor Networks (6G-enabled WSNs) in smart city with consumer electronics-based high communication density, especially for multimedia data. Deploying multiple IoT nodes on such networks makes the process complex and challenging. In such cases, quality of Service (QoS) is critical as it ensures critical network performance and leverages improved end-user experience. There have been some existing heuristic/meta-heuristic works to address the QoS in next-generation WSNs; however, they are sensitive to their parametric values due to a lack of expert knowledge. Some are less robust and less adaptable in dynamic networks due to poorer balanced exploration of the solution space, exploitation of known semi-optimal/optimal solutions, and inefficient resource utilization in constrained environments such as edge devices. The suggested consumer electronics-based research presents an innovative solution, ‘RL-MODE,’ which incorporates Reinforcement Learning-Enhanced Multiobjective Optimisation Algorithms to address QoS management difficulties in edge-enabled WSN-IoT systems. The proposed methodology optimises competing objectives simultaneously, such as minimising energy use and latency while maximizing throughput and coverage, all while keeping the resource-constrained nature of edge devices in mind. The proposed RL-MODE Algorithm comprises Multiobjective Differential Evolution (MODE) Algorithm and a new Reinforcement Learning (RL) adaption technique to develop Pareto-optimal solutions by analysing the complicated linkages between input parameters, edge resources, and QoS parameters. Simulations and experiments with Next-Gen WSN-IoT applications show the effectiveness of the proposed method. This not only improves QoS in WSN-IoT applications, but it also increases resource utilisation and scalability in edge computing settings.
{"title":"Next-Gen WSN Enabled IoT for Consumer Electronics in Smart City: Elevating Quality of Service Through Reinforcement Learning-Enhanced Multi-Objective Strategies","authors":"Shailendra Pratap Singh;Naween Kumar;Norah Saleh Alghamdi;Gaurav Dhiman;Wattana Viriyasitavat;Assadaporn Sapsomboon","doi":"10.1109/TCE.2024.3446988","DOIUrl":"10.1109/TCE.2024.3446988","url":null,"abstract":"The data transfer volume is massive in next-generation Wireless Sensor Networks (6G-enabled WSNs) in smart city with consumer electronics-based high communication density, especially for multimedia data. Deploying multiple IoT nodes on such networks makes the process complex and challenging. In such cases, quality of Service (QoS) is critical as it ensures critical network performance and leverages improved end-user experience. There have been some existing heuristic/meta-heuristic works to address the QoS in next-generation WSNs; however, they are sensitive to their parametric values due to a lack of expert knowledge. Some are less robust and less adaptable in dynamic networks due to poorer balanced exploration of the solution space, exploitation of known semi-optimal/optimal solutions, and inefficient resource utilization in constrained environments such as edge devices. The suggested consumer electronics-based research presents an innovative solution, ‘RL-MODE,’ which incorporates Reinforcement Learning-Enhanced Multiobjective Optimisation Algorithms to address QoS management difficulties in edge-enabled WSN-IoT systems. The proposed methodology optimises competing objectives simultaneously, such as minimising energy use and latency while maximizing throughput and coverage, all while keeping the resource-constrained nature of edge devices in mind. The proposed RL-MODE Algorithm comprises Multiobjective Differential Evolution (MODE) Algorithm and a new Reinforcement Learning (RL) adaption technique to develop Pareto-optimal solutions by analysing the complicated linkages between input parameters, edge resources, and QoS parameters. Simulations and experiments with Next-Gen WSN-IoT applications show the effectiveness of the proposed method. This not only improves QoS in WSN-IoT applications, but it also increases resource utilisation and scalability in edge computing settings.","PeriodicalId":13208,"journal":{"name":"IEEE Transactions on Consumer Electronics","volume":"70 4","pages":"6507-6518"},"PeriodicalIF":4.3,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142189670","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","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：