Pub Date : 2025-12-27DOI: 10.1016/j.pmcj.2025.102158
Marcello Pietri , Natalia Selini Hadjidimitriou , Marco Mamei , Marco Picone , Enrico Rossini , Edoardo Maria Sanna , Jovanka Adzic , Andrea Buldorini
In this research, we propose automating network management through data-driven intelligence, with a particular focus on anomalies and network traffic during specific events or periods. We analyze a large dataset collected by Orange mobile network operator in France with the goal of forecasting mobile demand for different classes of services. To model the underlying network infrastructure, we introduce a model for the underlying network based on a hierarchy of virtualization layers and slices. Building on this model, we propose algorithms to optimize the resources allocated to network slices and traffic distribution within the operator’s network. Network performance is evaluated as the fraction of time the mobile traffic is within the capacity of the network. Our results demonstrate that dynamic reallocation of resources among slices, and dynamic load balancing (traffic shaping) between nodes notably improves network performance. These results provide insights into critical aspects related to future 5G network management.
{"title":"Traffic analysis and resource adaptation in large-scale 5G multi-layer edge networks","authors":"Marcello Pietri , Natalia Selini Hadjidimitriou , Marco Mamei , Marco Picone , Enrico Rossini , Edoardo Maria Sanna , Jovanka Adzic , Andrea Buldorini","doi":"10.1016/j.pmcj.2025.102158","DOIUrl":"10.1016/j.pmcj.2025.102158","url":null,"abstract":"<div><div>In this research, we propose automating network management through data-driven intelligence, with a particular focus on anomalies and network traffic during specific events or periods. We analyze a large dataset collected by Orange mobile network operator in France with the goal of forecasting mobile demand for different classes of services. To model the underlying network infrastructure, we introduce a model for the underlying network based on a hierarchy of virtualization layers and slices. Building on this model, we propose algorithms to optimize the resources allocated to network slices and traffic distribution within the operator’s network. Network performance is evaluated as the fraction of time the mobile traffic is within the capacity of the network. Our results demonstrate that dynamic reallocation of resources among slices, and dynamic load balancing (traffic shaping) between nodes notably improves network performance. These results provide insights into critical aspects related to future 5G network management.</div></div>","PeriodicalId":49005,"journal":{"name":"Pervasive and Mobile Computing","volume":"116 ","pages":"Article 102158"},"PeriodicalIF":3.5,"publicationDate":"2025-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884604","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Federated Learning (FL) enables distributed learning of sensitive and multi-location data to enhance data privacy and avoid data leakages. In FL, participating clients only share the trained local model updates to the external server for global aggregation. However, FL is vulnerable to security threats like adaptive data poisoning attacks in non-IID data scenarios, especially coordinated attacks by colluding sybils. Existing defences struggle against sybils as their focus has been on either analysing model behaviour or strengthening aggregation mechanisms. Sybils can collude by sharing their model updates with each other in the external environment. They can then strategically manipulate and share poisoned model updates for global aggregation, specifically employing label-flipping attacks. This paper introduces a novel defence framework that shifts the focus from model analysis to sybil behaviour analysis utilizing historical pairwise cosine similarity of client updates during the training process. By establishing a dynamic threshold and analysing patterns of similarity change among participating clients, the proposed defence framework detects sybils exhibiting coordinated data poisoning and excludes them from the subsequent global aggregation process. This approach is adaptable to various underlying aggregation methods, providing a robust defence against collusive data poisoning attacks and improving model resilience and convergence even under challenging non-IID settings. Comprehensive evaluations demonstrate a significant reduction in the Attack Success Ratio, i.e. from over 40% to below 1%, showcasing its superior efficacy compared to state-of-the-art defences against targeted data poisoning attacks.
{"title":"Sybil-aware adaptive defence framework for robust federated learning","authors":"Dnyanesh Khedekar , Tanmaya Mahapatra , Amitesh Singh Rajput","doi":"10.1016/j.pmcj.2025.102157","DOIUrl":"10.1016/j.pmcj.2025.102157","url":null,"abstract":"<div><div>Federated Learning (FL) enables distributed learning of sensitive and multi-location data to enhance data privacy and avoid data leakages. In FL, participating clients only share the trained local model updates to the external server for global aggregation. However, FL is vulnerable to security threats like adaptive data poisoning attacks in non-IID data scenarios, especially coordinated attacks by colluding sybils. Existing defences struggle against sybils as their focus has been on either analysing model behaviour or strengthening aggregation mechanisms. Sybils can collude by sharing their model updates with each other in the external environment. They can then strategically manipulate and share poisoned model updates for global aggregation, specifically employing label-flipping attacks. This paper introduces a novel defence framework that shifts the focus from model analysis to sybil behaviour analysis utilizing historical pairwise cosine similarity of client updates during the training process. By establishing a dynamic threshold and analysing patterns of similarity change among participating clients, the proposed defence framework detects sybils exhibiting coordinated data poisoning and excludes them from the subsequent global aggregation process. This approach is adaptable to various underlying aggregation methods, providing a robust defence against collusive data poisoning attacks and improving model resilience and convergence even under challenging non-IID settings. Comprehensive evaluations demonstrate a significant reduction in the Attack Success Ratio, i.e. from over 40% to below 1%, showcasing its superior efficacy compared to state-of-the-art defences against targeted data poisoning attacks.</div></div>","PeriodicalId":49005,"journal":{"name":"Pervasive and Mobile Computing","volume":"116 ","pages":"Article 102157"},"PeriodicalIF":3.5,"publicationDate":"2025-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145790984","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2025-12-14DOI: 10.1016/j.pmcj.2025.102156
Zhengmiao Jin , Renxiang Chen , Ke Wu , Dong Liang , Li Yan
The flexibility of unmanned aerial vehicle (UAV) provides a promising solution for large-scale urban services. However, their limited energy remains one of the primary constraints affecting task efficiency. Trajectory optimization is required for energy management, as incorrect path decisions can result in lower task performance and potentially cause damage to the UAV or the urban environment. This paper investigates the path decision-making problem of UAV in dense, high-rise urban environments, characterized by optimal decisions for flying over or around obstacles to minimize energy consumption. Firstly, this study establishes a UAV energy consumption model based on the differences in energy usage across various flight states, and frames the UAV trajectory optimization problem as a Markov Decision Process (MDP), solved using the Deep Deterministic Policy Gradient (DDPG) framework. Secondly, within the Deep Reinforcement Learning (DRL) environment, when the UAV faces a choice between flying over or around obstacles, the exploration-exploitation dilemma arises due to the target-proximity-based dense reward function setup. This research proposes a three-stage learning framework, with a notable feature in the second stage, where exploration is driven by the gradient features of obstacle height to counteract excessive exploitation induced by the reward function. Finally, to address the issue of the algorithm’s experience sampling strategy neglecting the mismatch between the current state and past experiences, which arises due to the progression of the training process, this paper proposes a two-stage experience replay strategy. One notable feature of this strategy is the pattern-matching filtering method in the second stage, which selects experiences that closely match the current state for sampling, thereby accelerating the training process. Extensive simulation experiments demonstrate the necessity and effectiveness of the proposed exploration strategy and experience replay strategy.
{"title":"Gradient-driven exploration and pattern matching experience replay for efficient UAV path planning: Flying over or around?","authors":"Zhengmiao Jin , Renxiang Chen , Ke Wu , Dong Liang , Li Yan","doi":"10.1016/j.pmcj.2025.102156","DOIUrl":"10.1016/j.pmcj.2025.102156","url":null,"abstract":"<div><div>The flexibility of unmanned aerial vehicle (UAV) provides a promising solution for large-scale urban services. However, their limited energy remains one of the primary constraints affecting task efficiency. Trajectory optimization is required for energy management, as incorrect path decisions can result in lower task performance and potentially cause damage to the UAV or the urban environment. This paper investigates the path decision-making problem of UAV in dense, high-rise urban environments, characterized by optimal decisions for flying over or around obstacles to minimize energy consumption. Firstly, this study establishes a UAV energy consumption model based on the differences in energy usage across various flight states, and frames the UAV trajectory optimization problem as a Markov Decision Process (MDP), solved using the Deep Deterministic Policy Gradient (DDPG) framework. Secondly, within the Deep Reinforcement Learning (DRL) environment, when the UAV faces a choice between flying over or around obstacles, the exploration-exploitation dilemma arises due to the target-proximity-based dense reward function setup. This research proposes a three-stage learning framework, with a notable feature in the second stage, where exploration is driven by the gradient features of obstacle height to counteract excessive exploitation induced by the reward function. Finally, to address the issue of the algorithm’s experience sampling strategy neglecting the mismatch between the current state and past experiences, which arises due to the progression of the training process, this paper proposes a two-stage experience replay strategy. One notable feature of this strategy is the pattern-matching filtering method in the second stage, which selects experiences that closely match the current state for sampling, thereby accelerating the training process. Extensive simulation experiments demonstrate the necessity and effectiveness of the proposed exploration strategy and experience replay strategy.</div></div>","PeriodicalId":49005,"journal":{"name":"Pervasive and Mobile Computing","volume":"117 ","pages":"Article 102156"},"PeriodicalIF":3.5,"publicationDate":"2025-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145980647","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2025-12-04DOI: 10.1016/j.pmcj.2025.102153
Euijong Lee , Jaemin Jeong , Gyuchan Jo , Taegyeom Lee , Gee-Myung Moon , Young-Duk Seo , Ji-Hoon Jeong
Childhood overweight and obesity have emerged as some of the most serious global public health challenges, as they can lead to a variety of health-related problems and the early development of chronic diseases. The healthcare domain has been transformed by the integration of the Internet of Things (IoT), leading to the development of digital healthcare solutions. This integration has led to an increase in the health data collected from a variety of IoT sources. Consequently, advanced technologies are required to analyze health data, and Artificial Intelligence (AI) has been employed to extract meaningful insights from the data. Moreover, these technologies can be effectively applied in healthcare to address childhood overweight and obesity. A self-adaptive framework is proposed to manage childhood weight using lifelog data from IoT environments. An ensemble-based learning model is applied to predict weight using the lifelog data. A smartphone application providing real-world services was developed, and lifelog data were collected from a cohort of 362 children aged 104 to 152 months. Reasonable results were obtained, indicating that the proposed ensemble model can predict childhood weight change using lifelog data with an accuracy of 0.9711 and an F1-score 0.9725. Also, the results indicated that appropriate rewards can encourage human involvement and positively influence data quality.The results demonstrated the efficiency of the proposed framework with human involvement in weight prediction. The experimental results demonstrate the efficiency of the proposed framework and its potential application in healthcare services to enhance children’s health.
{"title":"A self-adaptive framework for child healthcare in IoT environment using AI-based prediction","authors":"Euijong Lee , Jaemin Jeong , Gyuchan Jo , Taegyeom Lee , Gee-Myung Moon , Young-Duk Seo , Ji-Hoon Jeong","doi":"10.1016/j.pmcj.2025.102153","DOIUrl":"10.1016/j.pmcj.2025.102153","url":null,"abstract":"<div><div>Childhood overweight and obesity have emerged as some of the most serious global public health challenges, as they can lead to a variety of health-related problems and the early development of chronic diseases. The healthcare domain has been transformed by the integration of the Internet of Things (IoT), leading to the development of digital healthcare solutions. This integration has led to an increase in the health data collected from a variety of IoT sources. Consequently, advanced technologies are required to analyze health data, and Artificial Intelligence (AI) has been employed to extract meaningful insights from the data. Moreover, these technologies can be effectively applied in healthcare to address childhood overweight and obesity. A self-adaptive framework is proposed to manage childhood weight using lifelog data from IoT environments. An ensemble-based learning model is applied to predict weight using the lifelog data. A smartphone application providing real-world services was developed, and lifelog data were collected from a cohort of 362 children aged 104 to 152 months. Reasonable results were obtained, indicating that the proposed ensemble model can predict childhood weight change using lifelog data with an accuracy of 0.9711 and an F1-score 0.9725. Also, the results indicated that appropriate rewards can encourage human involvement and positively influence data quality.The results demonstrated the efficiency of the proposed framework with human involvement in weight prediction. The experimental results demonstrate the efficiency of the proposed framework and its potential application in healthcare services to enhance children’s health.</div></div>","PeriodicalId":49005,"journal":{"name":"Pervasive and Mobile Computing","volume":"116 ","pages":"Article 102153"},"PeriodicalIF":3.5,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145738157","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Rapid advances in electronic textiles have enabled the development of smart socks with temperature-sensing capabilities for real-time foot temperature monitoring, where stretchability, user comfort, durability, bending, washability, sensing location error during wear, and higher manufacturing cost are drawbacks. This study introduces a smart sock prototype that integrates highly stretchable hybrid electronic temperature-sensing (SHETS) yarns in key physiological foot regions—hallux, metatarsal, midfoot, and heel—to detect temperature variations and address the challenges of existing and commercial products. Structural design of SHETS yarns involved wrapping, novel interconnection, encapsulation, and braiding techniques to integrate miniature thermistor within the yarn structure, showing high stretchability, durability, washability, and sensing accuracy. Low-power microcontroller transmits analog data to a digital format, and the web-based interface enables users to monitor data through mobile phone applications in real-time. The prototype demonstrates high accuracy, durability, and reliability under various conditions, with average temperature error ranging from ±0.23 °C to ±0.27 °C. The prototype maintains durability and stability in hot, cold, and sweaty conditions. Physical activities like walking, running, and cycling demonstrate the durability and stability of foot temperature changes, while extended wear shows low power consumption and stability. The device withstands over 30 washing cycles with minimal accuracy loss (maximum ±0.28 °C error) and retains the mechanical and electrical properties of SHETS yarn under repeated stretching and bending. Additionally, the android-based intelligent foot alert system enhances usability by providing real-time monitoring and alerts on smartphones, offering a cost-effective, energy-efficient, and user-friendly solution for proactive foot health management. Available code: https://github.com/Sumonta-e-textile/Smart-Materials-and-Electronic-Textiles-Lab.
{"title":"The structure design of the smart sock prototype integrated with stretchable hybrid electronic temperature sensing yarn for real-time temperature monitoring","authors":"Sumonta Ghosh , Fenye Meng , Rony Shaha , Jiyong Hu","doi":"10.1016/j.pmcj.2025.102136","DOIUrl":"10.1016/j.pmcj.2025.102136","url":null,"abstract":"<div><div>Rapid advances in electronic textiles have enabled the development of smart socks with temperature-sensing capabilities for real-time foot temperature monitoring, where stretchability, user comfort, durability, bending, washability, sensing location error during wear, and higher manufacturing cost are drawbacks. This study introduces a smart sock prototype that integrates highly stretchable hybrid electronic temperature-sensing (SHETS) yarns in key physiological foot regions—hallux, metatarsal, midfoot, and heel—to detect temperature variations and address the challenges of existing and commercial products. Structural design of SHETS yarns involved wrapping, novel interconnection, encapsulation, and braiding techniques to integrate miniature thermistor within the yarn structure, showing high stretchability, durability, washability, and sensing accuracy. Low-power microcontroller transmits analog data to a digital format, and the web-based interface enables users to monitor data through mobile phone applications in real-time. The prototype demonstrates high accuracy, durability, and reliability under various conditions, with average temperature error ranging from ±0.23 °C to ±0.27 °C. The prototype maintains durability and stability in hot, cold, and sweaty conditions. Physical activities like walking, running, and cycling demonstrate the durability and stability of foot temperature changes, while extended wear shows low power consumption and stability. The device withstands over 30 washing cycles with minimal accuracy loss (maximum ±0.28 °C error) and retains the mechanical and electrical properties of SHETS yarn under repeated stretching and bending. Additionally, the android-based intelligent foot alert system enhances usability by providing real-time monitoring and alerts on smartphones, offering a cost-effective, energy-efficient, and user-friendly solution for proactive foot health management. Available code: <span><span>https://github.com/Sumonta-e-textile/Smart-Materials-and-Electronic-Textiles-Lab</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":49005,"journal":{"name":"Pervasive and Mobile Computing","volume":"116 ","pages":"Article 102136"},"PeriodicalIF":3.5,"publicationDate":"2025-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145659149","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2025-11-24DOI: 10.1016/j.pmcj.2025.102134
João Gameiro , Rodrigo Rosmaninho , Gonçalo Perna , Pedro Rito , Susana Sargento , Carlos Marques , Filipe Pinto
The growing scale of smart city sensing devices and infrastructure entails a wide variety of available sensing information that can provide valuable insights into user mobility and traffic congestion. This information can be used to optimize service delivery through the development of mobility-aware services. 5G systems and their associated technologies provide an ideal environment with capabilities to efficiently support edge computing and bring the processing and storage resources closer to the end users, which results in a latency and backhaul usage reduction.
This article proposes the integration of edge computing in 5G operator network and a mobility/road-side infrastructure with edge orchestration to provide mobility-aware services to the end-users on demand. With this approach, a service instantiation can be translated into resource allocation both on the 5G platform through multi-slicing and the edge infrastructure. Resource management is then optimized for the users on the move by continuously allocating the necessary virtual network slices, processing, and storage resources in the appropriate locations for the user to consume its services while maintaining the appropriate QoS levels and optimized resource distribution in the edge platform. This approach is evaluated in a real mobile 5G network with emulated Radio Access Network (RAN) resources through two use cases based on infotainment and emergency services. The results show that the approach is efficient in using mobility, service requirements, and platform’s resources information to enable a proactive resource reservation both in the 5G base stations and edge computing nodes throughout the path traversed by the users.
{"title":"Edge computing and 5G network integration for mobility-aware service deployments","authors":"João Gameiro , Rodrigo Rosmaninho , Gonçalo Perna , Pedro Rito , Susana Sargento , Carlos Marques , Filipe Pinto","doi":"10.1016/j.pmcj.2025.102134","DOIUrl":"10.1016/j.pmcj.2025.102134","url":null,"abstract":"<div><div>The growing scale of smart city sensing devices and infrastructure entails a wide variety of available sensing information that can provide valuable insights into user mobility and traffic congestion. This information can be used to optimize service delivery through the development of mobility-aware services. 5G systems and their associated technologies provide an ideal environment with capabilities to efficiently support edge computing and bring the processing and storage resources closer to the end users, which results in a latency and backhaul usage reduction.</div><div>This article proposes the integration of edge computing in 5G operator network and a mobility/road-side infrastructure with edge orchestration to provide mobility-aware services to the end-users on demand. With this approach, a service instantiation can be translated into resource allocation both on the 5G platform through multi-slicing and the edge infrastructure. Resource management is then optimized for the users on the move by continuously allocating the necessary virtual network slices, processing, and storage resources in the appropriate locations for the user to consume its services while maintaining the appropriate QoS levels and optimized resource distribution in the edge platform. This approach is evaluated in a real mobile 5G network with emulated Radio Access Network (RAN) resources through two use cases based on infotainment and emergency services. The results show that the approach is efficient in using mobility, service requirements, and platform’s resources information to enable a proactive resource reservation both in the 5G base stations and edge computing nodes throughout the path traversed by the users.</div></div>","PeriodicalId":49005,"journal":{"name":"Pervasive and Mobile Computing","volume":"115 ","pages":"Article 102134"},"PeriodicalIF":3.5,"publicationDate":"2025-11-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145615395","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper considers the localization problem for Intelligent Transport Systems (ITS) where micromobility vehicles are localized by determining their own coordinates according to the surrounding neighbors. In this work, we develop two approaches for the self-localization service by adopting efficient methods, exploiting V2X traffic exchange to offer an accurate positioning technique and to fulfill energy efficiency. Micromobility entities in this case are not supposed to be equipped with specific modules or particular sensors. First, we propose a Local Localization Approach (LLA), based on advanced KNN machine learning called combined KNN, where the algorithm adjusts the K parameter in order to estimate neighbors distances by considering all possible combinations. To update the K-value, we take into account a predefined error threshold in order to enhance the accuracy of the localization service. The proposed LLA demonstrates significant improvements compared to existing approaches. The estimated error depends on the K value. In fact, it is equal to 0.0163 m (for K = 3), 0.0151 m (for K = 5) and 0.0143 m (for K = 7). Secondly, we develop the Energy Saving version (ES-LLA) to enhance the energy efficiency. The numerical results demonstrate the performance of the two algorithms. The ES-LLA offers a gain of 0.2 W, compared to the LLA in terms of power consumption. Both algorithms present advantageous approaches because of their low cost, scalability, and fast adaptation in dynamic environments. We offer a sustainable solution for total transmit power reduction that optimizes the usage of resources on the vehicular network.
本文研究了智能交通系统中微机动车辆的定位问题，微机动车辆的定位方法是根据周边车辆确定自身坐标。在这项工作中，我们通过采用有效的方法开发了两种自定位服务方法，利用V2X流量交换提供准确的定位技术并实现能源效率。在这种情况下，微移动实体不应该配备特定的模块或特定的传感器。首先，我们提出了一种基于高级KNN机器学习的局部定位方法（LLA），称为组合KNN，其中算法通过考虑所有可能的组合来调整K参数以估计邻居距离。为了更新k值，我们考虑了预定义的错误阈值，以提高定位服务的准确性。与现有方法相比，所提出的LLA显示出显著的改进。估计误差取决于K值。实际上，它等于0.0163 m (K = 3), 0.0151 m （K = 5）和0.0143 m （K = 7）。其次，我们开发了节能版（ES-LLA），以提高能源效率。数值结果验证了两种算法的性能。在功耗方面，ES-LLA提供0.2 W的增益，与LLA相比。这两种算法都具有低成本、可扩展性和快速适应动态环境的优点。我们提供了一个可持续的解决方案，降低总传输功率，优化车辆网络上的资源使用。
{"title":"Energy-aware vehicle localization in dynamic environments via efficient machine learning techniques for positioning and power management","authors":"Hend Marouane , Mohamed Mosbah , Hassene Mnif , Amel Meddeb Makhlouf","doi":"10.1016/j.pmcj.2025.102135","DOIUrl":"10.1016/j.pmcj.2025.102135","url":null,"abstract":"<div><div>This paper considers the localization problem for Intelligent Transport Systems (ITS) where micromobility vehicles are localized by determining their own coordinates according to the surrounding neighbors. In this work, we develop two approaches for the self-localization service by adopting efficient methods, exploiting V2X traffic exchange to offer an accurate positioning technique and to fulfill energy efficiency. Micromobility entities in this case are not supposed to be equipped with specific modules or particular sensors. First, we propose a Local Localization Approach (LLA), based on advanced KNN machine learning called combined KNN, where the algorithm adjusts the K parameter in order to estimate neighbors distances by considering all possible combinations. To update the K-value, we take into account a predefined error threshold in order to enhance the accuracy of the localization service. The proposed LLA demonstrates significant improvements compared to existing approaches. The estimated error depends on the K value. In fact, it is equal to 0.0163 m (for K = 3), 0.0151 m (for K = 5) and 0.0143 m (for K = 7). Secondly, we develop the Energy Saving version (ES-LLA) to enhance the energy efficiency. The numerical results demonstrate the performance of the two algorithms. The ES-LLA offers a gain of 0.2 W, compared to the LLA in terms of power consumption. Both algorithms present advantageous approaches because of their low cost, scalability, and fast adaptation in dynamic environments. We offer a sustainable solution for total transmit power reduction that optimizes the usage of resources on the vehicular network.</div></div>","PeriodicalId":49005,"journal":{"name":"Pervasive and Mobile Computing","volume":"115 ","pages":"Article 102135"},"PeriodicalIF":3.5,"publicationDate":"2025-11-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145618285","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2025-11-10DOI: 10.1016/j.pmcj.2025.102131
Huimin Yao , Dengao Li , Jumin Zhao
Accurate human action recognition (HAR) using sparse millimeter-wave (mmWave) radar point clouds faces significant challenges. Existing approaches suffer from ineffective feature extraction in sparse point clouds, vulnerability to radar noise and multi-path interference, and significant intra-class variance induced by distance–angle variations. To overcome these limitations, we propose SMC-HAR, a novel Supervised Momentum Contrast framework for HAR. SMC-HAR leverages contrastive learning with a joint loss function that integrates supervised contrastive loss and cross-entropy loss. This design enhances feature discriminability, mitigates intra-class dispersion, and promotes feature aggregation within classes while improving separation between classes. Our momentum mechanism dynamically optimizes the feature distribution reference space and bolsters robustness against noise and multi-path interference. Furthermore, we design a domain-specific augmentation optimization strategy tailored for mmWave radar point clouds in HAR, which explores optimal synergistic combinations of augmentations to better adapt to point cloud sparsity and action pattern characteristics. Experimental results on the widely used MM-Fi dataset show that SMC-HAR achieves a classification accuracy of 88.40%, marking a substantial 8.40% improvement over the baseline cross-entropy model. This demonstrates the effectiveness of our framework in enhancing feature discriminability and robustness for mmWave point cloud-based HAR.
{"title":"Supervised momentum contrastive learning for mmWave-based human action recognition","authors":"Huimin Yao , Dengao Li , Jumin Zhao","doi":"10.1016/j.pmcj.2025.102131","DOIUrl":"10.1016/j.pmcj.2025.102131","url":null,"abstract":"<div><div>Accurate human action recognition (HAR) using sparse millimeter-wave (mmWave) radar point clouds faces significant challenges. Existing approaches suffer from ineffective feature extraction in sparse point clouds, vulnerability to radar noise and multi-path interference, and significant intra-class variance induced by distance–angle variations. To overcome these limitations, we propose SMC-HAR, a novel Supervised Momentum Contrast framework for HAR. SMC-HAR leverages contrastive learning with a joint loss function that integrates supervised contrastive loss and cross-entropy loss. This design enhances feature discriminability, mitigates intra-class dispersion, and promotes feature aggregation within classes while improving separation between classes. Our momentum mechanism dynamically optimizes the feature distribution reference space and bolsters robustness against noise and multi-path interference. Furthermore, we design a domain-specific augmentation optimization strategy tailored for mmWave radar point clouds in HAR, which explores optimal synergistic combinations of augmentations to better adapt to point cloud sparsity and action pattern characteristics. Experimental results on the widely used MM-Fi dataset show that SMC-HAR achieves a classification accuracy of 88.40%, marking a substantial 8.40% improvement over the baseline cross-entropy model. This demonstrates the effectiveness of our framework in enhancing feature discriminability and robustness for mmWave point cloud-based HAR.</div></div>","PeriodicalId":49005,"journal":{"name":"Pervasive and Mobile Computing","volume":"115 ","pages":"Article 102131"},"PeriodicalIF":3.5,"publicationDate":"2025-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145520651","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2025-11-07DOI: 10.1016/j.pmcj.2025.102133
Valentin Stangaciu , Cristina Stangaciu , Daniel-Ioan Curiac , Mihai V. Micea
The Internet of Things concept has expanded to a large area of applications evolving to the point of providing even real-time support. Critical applications become increasingly suitable at the Edge Layer where real-time operations need to be supported at both node and network level thus communication becomes crucial. This paper presents a real-time communication solution based on the highly popular XBee modules. We describe a predictable and modular driver for such modules along with a full communication platform ready to be integrated into an IoT design for real-time applications. The proposed communication module has been implemented at prototype level and successfully validated through an extensive set of simulations and experiments.
{"title":"RTXBEE: Real-time communication module for critical Internet of Things applications","authors":"Valentin Stangaciu , Cristina Stangaciu , Daniel-Ioan Curiac , Mihai V. Micea","doi":"10.1016/j.pmcj.2025.102133","DOIUrl":"10.1016/j.pmcj.2025.102133","url":null,"abstract":"<div><div>The Internet of Things concept has expanded to a large area of applications evolving to the point of providing even real-time support. Critical applications become increasingly suitable at the Edge Layer where real-time operations need to be supported at both node and network level thus communication becomes crucial. This paper presents a real-time communication solution based on the highly popular XBee modules. We describe a predictable and modular driver for such modules along with a full communication platform ready to be integrated into an IoT design for real-time applications. The proposed communication module has been implemented at prototype level and successfully validated through an extensive set of simulations and experiments.</div></div>","PeriodicalId":49005,"journal":{"name":"Pervasive and Mobile Computing","volume":"115 ","pages":"Article 102133"},"PeriodicalIF":3.5,"publicationDate":"2025-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145468054","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2025-11-06DOI: 10.1016/j.pmcj.2025.102132
Guowen Li , Zihang Huang , Teng Fei , Dunxin Jia , Meng Bian
Accurate real-time traffic flow monitoring is crucial for intelligent transportation systems (ITS), enabling optimized traffic management, urban planning, and policy-making. However, conventional methods face cost, deployment, weather, and privacy challenges. Addressing these shortcomings, this study investigates the potential of utilizing ubiquitous traffic noise, an inherently accessible, cost-efficient, non-intrusive, and privacy-preserving signal, as a viable data source. We propose the Lightweight Noise Spectrogram Transformer (LNST), a novel deep learning model for analyzing traffic noise spectrograms as a Proof of Concept. LNST leverages the Transformer architecture's self-attention mechanism to effectively capture long-range temporal and spectral dependencies crucial for interpreting complex traffic acoustics. Trained and evaluated on diverse urban traffic scenarios, LNST demonstrates significant advantages. Experimental results show it consistently outperforms baseline models, achieving superior prediction accuracy (MSE, MAE, R²). Furthermore, through transfer learning and model pruning, LNST achieves high computational efficiency with substantially fewer parameters and faster inference speeds. Its lighter design also ensures its feasibility for deployment on resource-constrained edge computing platforms. This work validates the practicality of acoustic sensing for traffic monitoring and presents an accurate, computationally efficient, and LNST as a cost-effective, easily deployable, and privacy-respecting solution, offering a valuable supplementary tool for advancing ITS.
{"title":"Listen to the road: acoustic traffic monitoring on edge platforms via Lightweight Noise Spectrogram Transformer (LNST)","authors":"Guowen Li , Zihang Huang , Teng Fei , Dunxin Jia , Meng Bian","doi":"10.1016/j.pmcj.2025.102132","DOIUrl":"10.1016/j.pmcj.2025.102132","url":null,"abstract":"<div><div>Accurate real-time traffic flow monitoring is crucial for intelligent transportation systems (ITS), enabling optimized traffic management, urban planning, and policy-making. However, conventional methods face cost, deployment, weather, and privacy challenges. Addressing these shortcomings, this study investigates the potential of utilizing ubiquitous traffic noise, an inherently accessible, cost-efficient, non-intrusive, and privacy-preserving signal, as a viable data source. We propose the Lightweight Noise Spectrogram Transformer (LNST), a novel deep learning model for analyzing traffic noise spectrograms as a Proof of Concept. LNST leverages the Transformer architecture's self-attention mechanism to effectively capture long-range temporal and spectral dependencies crucial for interpreting complex traffic acoustics. Trained and evaluated on diverse urban traffic scenarios, LNST demonstrates significant advantages. Experimental results show it consistently outperforms baseline models, achieving superior prediction accuracy (MSE, MAE, R²). Furthermore, through transfer learning and model pruning, LNST achieves high computational efficiency with substantially fewer parameters and faster inference speeds. Its lighter design also ensures its feasibility for deployment on resource-constrained edge computing platforms. This work validates the practicality of acoustic sensing for traffic monitoring and presents an accurate, computationally efficient, and LNST as a cost-effective, easily deployable, and privacy-respecting solution, offering a valuable supplementary tool for advancing ITS.</div></div>","PeriodicalId":49005,"journal":{"name":"Pervasive and Mobile Computing","volume":"115 ","pages":"Article 102132"},"PeriodicalIF":3.5,"publicationDate":"2025-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145569417","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号