The Internet of Vehicles (IoV) is emerging as a pivotal technology for enhancing traffic management and safety. Its rapid development demands solutions for enhanced communication efficiency and reduced latency. However, traditional centralized networks struggle to meet these demands, prompting the exploration of decentralized solutions such as blockchain. Addressing blockchain’s scalability challenges posed by the growing number of nodes and transactions calls for innovative solutions, among which sharding stands out as a pivotal approach to significantly enhance blockchain throughput. However, existing schemes still face challenges related to a) the impact of vehicle mobility on blockchain consensus, especially for cross-shard transaction; and b) the strict requirements of low latency consensus in a highly dynamic network. In this paper, we propose a DAG (Directed Acyclic Graph) consensus leveraging Robust Dynamic Sharding and Tree-broadcasting (DRDST) to address these challenges. Specifically, we first develop a standard for evaluating the network stability of nodes, combined with the nodes’ trust values, to propose a novel robust sharding model that is solved through the design of the Genetic Sharding Algorithm (GSA). Then, we optimize the broadcast latency of the whole sharded network by improving the tree-broadcasting to minimize the maximum broadcast latency within each shard. On this basis, we also design a DAG consensus scheme based on an improved hashgraph protocol, which can efficiently handle cross-shard transactions. Finally, the simulation proves the proposed scheme is superior to the comparison schemes in latency, throughput, consensus success rate, and node traffic load.
{"title":"DRDST: Low-Latency DAG Consensus Through Robust Dynamic Sharding and Tree-Broadcasting for IoV","authors":"Runhua Chen;Haoxiang Luo;Gang Sun;Hongfang Yu;Dusit Niyato;Schahram Dustdar","doi":"10.1109/TMC.2025.3599385","DOIUrl":"https://doi.org/10.1109/TMC.2025.3599385","url":null,"abstract":"The Internet of Vehicles (IoV) is emerging as a pivotal technology for enhancing traffic management and safety. Its rapid development demands solutions for enhanced communication efficiency and reduced latency. However, traditional centralized networks struggle to meet these demands, prompting the exploration of decentralized solutions such as blockchain. Addressing blockchain’s scalability challenges posed by the growing number of nodes and transactions calls for innovative solutions, among which sharding stands out as a pivotal approach to significantly enhance blockchain throughput. However, existing schemes still face challenges related to <italic>a</i>) the impact of vehicle mobility on blockchain consensus, especially for cross-shard transaction; and <italic>b</i>) the strict requirements of low latency consensus in a highly dynamic network. In this paper, we propose a DAG (Directed Acyclic Graph) consensus leveraging Robust Dynamic Sharding and Tree-broadcasting (DRDST) to address these challenges. Specifically, we first develop a standard for evaluating the network stability of nodes, combined with the nodes’ trust values, to propose a novel robust sharding model that is solved through the design of the Genetic Sharding Algorithm (GSA). Then, we optimize the broadcast latency of the whole sharded network by improving the tree-broadcasting to minimize the maximum broadcast latency within each shard. On this basis, we also design a DAG consensus scheme based on an improved hashgraph protocol, which can efficiently handle cross-shard transactions. Finally, the simulation proves the proposed scheme is superior to the comparison schemes in latency, throughput, consensus success rate, and node traffic load.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"25 1","pages":"1236-1253"},"PeriodicalIF":9.2,"publicationDate":"2025-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145659157","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 : 2025-08-15DOI: 10.1109/TMC.2025.3599519
Yinbin Miao;Jiaqi Yu;Jiliang Li;Xinghua Li;Jun Feng;Zhiquan Liu;Robert H. Deng
Privacy-preserving spatial range query allows users to obtain valid data based on specific spatial attributes or geographical location while ensuring privacy. However, many existing Privacy-Preserving Spatial Range Query (PSRQ) schemes generally face the problems of low query efficiency and insufficient security when dealing with large-scale mobile cloud data sets, and it is difficult to resist Indistinguishability under Chosen-Plaintext Attack (IND-CPA). To solve these challenges, we first propose an Efficient and Secure Spatial Range Query scheme (ESSRQ), which is based on a dual mobile cloud architecture by integrating Geohash algorithm, Circular Shift Coalesce Zero-Sum Garbled Bloom Filter (CSC-ZGBF) and Symmetric Homomorphic Encryption (SHE), achieving a constant search complexity. However, ESSRQ cannot protect the access patterns, where the cloud server still has the potential to infer attacks based on the index position and even obtain plaintext queries. On this basis, we further propose an extended scheme ESSRQ-PIR, which introduces Private Information Retrieval (PIR) into single mobile cloud-based architecture, effectively prevents the leakage of access patterns, enhances the security of ESSRQ and can also realize efficient query on large-scale cloud datasets. Formal security analysis proves that our proposed schemes are secure against IND-CPA, and extensive experiments demonstrate that our schemes improve the query efficiency by up to nearly 20 times when compared with previous solutions. These features make the proposed schemes particularly suitable for privacy-preserving spatial queries in mobile cloud computing environments.
保护隐私的空间范围查询允许用户在保证隐私的同时，根据特定的空间属性或地理位置获取有效的数据。然而，现有的许多保隐私空间范围查询（PSRQ）方案在处理大规模移动云数据集时普遍存在查询效率低、安全性不足的问题，且难以抵抗选择明文攻击（IND-CPA）下的不可分辨性。为了解决这些挑战，我们首先提出了一种高效安全的空间距离查询方案（ESSRQ），该方案基于双移动云架构，通过集成Geohash算法、圆移位合并零和乱码布隆滤波器（CSC-ZGBF）和对称同态加密（SHE），实现了恒定的搜索复杂度。然而，ESSRQ不能保护访问模式，云服务器仍然有可能根据索引位置推断攻击，甚至获得明文查询。在此基础上，我们进一步提出了一种扩展方案ESSRQ-PIR，该方案将私有信息检索（Private Information Retrieval， PIR）引入到单个基于移动云的架构中，有效地防止了访问模式的泄漏，增强了ESSRQ的安全性，也可以实现对大规模云数据集的高效查询。正式的安全性分析证明了我们提出的方案对IND-CPA是安全的，大量的实验表明，我们的方案与以前的方案相比，查询效率提高了近20倍。这些特征使得所提出的方案特别适合于移动云计算环境中保护隐私的空间查询。
{"title":"Security-Enhanced Spatial Range Query Over Large-Scale Encrypted Mobile Cloud Datasets","authors":"Yinbin Miao;Jiaqi Yu;Jiliang Li;Xinghua Li;Jun Feng;Zhiquan Liu;Robert H. Deng","doi":"10.1109/TMC.2025.3599519","DOIUrl":"https://doi.org/10.1109/TMC.2025.3599519","url":null,"abstract":"Privacy-preserving spatial range query allows users to obtain valid data based on specific spatial attributes or geographical location while ensuring privacy. However, many existing Privacy-Preserving Spatial Range Query (PSRQ) schemes generally face the problems of low query efficiency and insufficient security when dealing with large-scale mobile cloud data sets, and it is difficult to resist Indistinguishability under Chosen-Plaintext Attack (IND-CPA). To solve these challenges, we first propose an Efficient and Secure Spatial Range Query scheme (ESSRQ), which is based on a dual mobile cloud architecture by integrating Geohash algorithm, Circular Shift Coalesce Zero-Sum Garbled Bloom Filter (CSC-ZGBF) and Symmetric Homomorphic Encryption (SHE), achieving a constant search complexity. However, ESSRQ cannot protect the access patterns, where the cloud server still has the potential to infer attacks based on the index position and even obtain plaintext queries. On this basis, we further propose an extended scheme ESSRQ-PIR, which introduces Private Information Retrieval (PIR) into single mobile cloud-based architecture, effectively prevents the leakage of access patterns, enhances the security of ESSRQ and can also realize efficient query on large-scale cloud datasets. Formal security analysis proves that our proposed schemes are secure against IND-CPA, and extensive experiments demonstrate that our schemes improve the query efficiency by up to nearly 20 times when compared with previous solutions. These features make the proposed schemes particularly suitable for privacy-preserving spatial queries in mobile cloud computing environments.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"25 1","pages":"1184-1199"},"PeriodicalIF":9.2,"publicationDate":"2025-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145659182","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 : 2025-08-14DOI: 10.1109/TMC.2025.3599406
Aobo Liang;Yan Sun;Nadra Guizani
Various sensors of Internet of Things (IoT) generate massive amounts of mobile traffic data, forming multivariate time series (MTS). Accurate forecasting of MTS facilitates the enhancement of proactive autoscaling and resource allocation in edge networks. While recent Transformer-based models (Transformers) have achieved significant success in MTS forecasting (MTSF), they tend to rely solely on either time-domain or frequency-domain features, which captures inadequate trends and periodic characteristics. To this end, we propose a wavelet learning framework that seamlessly integrates wavelet transforms with Transformers to benefit from time and frequency characteristics. We design a mixing-splitting architecture to model multi-scale wavelet coefficients and utilizes the attention mechanism to capture inter-series dependencies in the wavelet domain. However, the vanilla softmax self-attention (SA) is high-computational-cost and its smoothing effect diminishes the contrast between strong and weak variable correlations. Therefore, we propose a novel attention mechanism: Rotary Route Attention (RoRA). RoRA incorporates rotary positional embeddings to enhance feature diversity and introduces a small number of routing tokens $r$ to aggregate information from the $KV$ matrices and redistribute it to the $Q$ matrix. Such design strengthens interactions among strongly correlated variables while mitigating the impact of weakly correlated noise. We further propose WaveRoRA, a unified model that leverages RoRA capturing inter-series dependencies in the wavelet domain. We conduct extensive experiments on eight real-world datasets. The results indicate that WaveRoRA outperforms existing state-of-the-art models while maintaining lower computational costs.
{"title":"WaveRoRA: Wavelet Rotary Route Attention for Multivariate Time Series Forecasting","authors":"Aobo Liang;Yan Sun;Nadra Guizani","doi":"10.1109/TMC.2025.3599406","DOIUrl":"https://doi.org/10.1109/TMC.2025.3599406","url":null,"abstract":"Various sensors of Internet of Things (IoT) generate massive amounts of mobile traffic data, forming multivariate time series (MTS). Accurate forecasting of MTS facilitates the enhancement of proactive autoscaling and resource allocation in edge networks. While recent Transformer-based models (Transformers) have achieved significant success in MTS forecasting (MTSF), they tend to rely solely on either time-domain or frequency-domain features, which captures inadequate trends and periodic characteristics. To this end, we propose a wavelet learning framework that seamlessly integrates wavelet transforms with Transformers to benefit from time and frequency characteristics. We design a mixing-splitting architecture to model multi-scale wavelet coefficients and utilizes the attention mechanism to capture inter-series dependencies in the wavelet domain. However, the vanilla softmax self-attention (SA) is high-computational-cost and its smoothing effect diminishes the contrast between strong and weak variable correlations. Therefore, we propose a novel attention mechanism: Rotary Route Attention (RoRA). RoRA incorporates rotary positional embeddings to enhance feature diversity and introduces a small number of routing tokens <inline-formula><tex-math>$r$</tex-math></inline-formula> to aggregate information from the <inline-formula><tex-math>$KV$</tex-math></inline-formula> matrices and redistribute it to the <inline-formula><tex-math>$Q$</tex-math></inline-formula> matrix. Such design strengthens interactions among strongly correlated variables while mitigating the impact of weakly correlated noise. We further propose WaveRoRA, a unified model that leverages RoRA capturing inter-series dependencies in the wavelet domain. We conduct extensive experiments on eight real-world datasets. The results indicate that WaveRoRA outperforms existing state-of-the-art models while maintaining lower computational costs.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"25 1","pages":"1287-1301"},"PeriodicalIF":9.2,"publicationDate":"2025-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145659191","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}
Sparse Crowdsensing has emerged as a crucial and flexible method for collecting spatio-temporal data in various applications, such as traffic management, environmental monitoring, and disaster response. By recruiting users and utilizing their diverse mobile devices, this approach often results in data that is both sparse and multi-scale, complicating the data completion process. Although numerous data completion algorithms have been developed to address data sparsity, most assume that the collected data is of the same or similar scale, rendering them ineffective for multi-scale data. To overcome this limitation, in this paper, we propose a spatio-temporal pyramid-based multi-scale data completion framework in Sparse Crowdsensing. The basic idea is to leverage a pyramid structure to efficiently capture the complex interrelations between different scales. We first develop a Spatial-Temporal Pyramid Construction Module (ST-PC) to handle multi-scale inputs, and then propose a Spatial-Temporal Pyramid Attention Mechanism (ST-PAM) to capture multi-scale correlations while reducing computational complexity. Furthermore, our method incorporates cross-scale constraints to optimize completion performance. Extensive experiments on four real-world spatio-temporal datasets demonstrate the effectiveness of our framework in multi-scale data completion.
{"title":"Spatio-Temporal Pyramid-Based Multi-Scale Data Completion in Sparse Crowdsensing","authors":"Wenbin Liu;Hao Du;En Wang;Jiajian Lv;Weiting Liu;Bo Yang;Jie Wu","doi":"10.1109/TMC.2025.3599322","DOIUrl":"https://doi.org/10.1109/TMC.2025.3599322","url":null,"abstract":"Sparse Crowdsensing has emerged as a crucial and flexible method for collecting spatio-temporal data in various applications, such as traffic management, environmental monitoring, and disaster response. By recruiting users and utilizing their diverse mobile devices, this approach often results in data that is both sparse and multi-scale, complicating the data completion process. Although numerous data completion algorithms have been developed to address data sparsity, most assume that the collected data is of the same or similar scale, rendering them ineffective for multi-scale data. To overcome this limitation, in this paper, we propose a spatio-temporal pyramid-based multi-scale data completion framework in Sparse Crowdsensing. The basic idea is to leverage a pyramid structure to efficiently capture the complex interrelations between different scales. We first develop a Spatial-Temporal Pyramid Construction Module (ST-PC) to handle multi-scale inputs, and then propose a Spatial-Temporal Pyramid Attention Mechanism (ST-PAM) to capture multi-scale correlations while reducing computational complexity. Furthermore, our method incorporates cross-scale constraints to optimize completion performance. Extensive experiments on four real-world spatio-temporal datasets demonstrate the effectiveness of our framework in multi-scale data completion.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"25 1","pages":"1270-1286"},"PeriodicalIF":9.2,"publicationDate":"2025-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145659218","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 : 2025-08-14DOI: 10.1109/TMC.2025.3599384
Xingyu Li;Wenzhe Zhang;Linfeng Liu;Ping Wang
The deployment of Unmanned Aerial Vehicle (UAV) swarm for Forest Fire Detection (FFD) missions presents unique challenges, e.g., the early forest fires are difficult to identify due to environment diversity and feature complexity, especially when some UAVs could be destroyed in harsh environments. To address these challenges, UAV swarm-based FFD missions can leverage advanced deep learning techniques, where online model updates, robustness, and communication overhead control become crucial for ensuring the effectiveness and adaptability of these missions. In this paper, we propose a Two-tier Submodel Partition Framework (TSPF) to enhance the robustness of UAV swarm conducting FFD missions. TSPF utilizes online model updates to adapt to diverse mission environments, thus strengthening the generalization capability of the model. In addition, a graph coloring method, an intragroup backup mechanism, and a Dynamic Server Selection (DSS) mechanism for the grouping are employed to enhance the robustness of FFD missions when some UAVs are destroyed, hence maintaining the high performance of FFD missions in harsh environments. Moreover, TSPF enables submodel updates by aggregating the parameters of selected layers within/between UAV groups, thereby effectively reducing the model parameter uploads (communication overhead) in model training. Experimental evaluations demonstrate that our proposed TSPF significantly improves the detection accuracy of forest fires, enhances the robustness of FFD missions against the destruction of some UAVs, and reduces the communication overhead in FFD missions.
{"title":"Two-Tier Submodel Partition Framework for Enhancing UAV Swarm Robustness in Forest Fire Detection","authors":"Xingyu Li;Wenzhe Zhang;Linfeng Liu;Ping Wang","doi":"10.1109/TMC.2025.3599384","DOIUrl":"https://doi.org/10.1109/TMC.2025.3599384","url":null,"abstract":"The deployment of Unmanned Aerial Vehicle (UAV) swarm for Forest Fire Detection (FFD) missions presents unique challenges, e.g., the early forest fires are difficult to identify due to environment diversity and feature complexity, especially when some UAVs could be destroyed in harsh environments. To address these challenges, UAV swarm-based FFD missions can leverage advanced deep learning techniques, where online model updates, robustness, and communication overhead control become crucial for ensuring the effectiveness and adaptability of these missions. In this paper, we propose a Two-tier Submodel Partition Framework (TSPF) to enhance the robustness of UAV swarm conducting FFD missions. TSPF utilizes online model updates to adapt to diverse mission environments, thus strengthening the generalization capability of the model. In addition, a graph coloring method, an intragroup backup mechanism, and a Dynamic Server Selection (DSS) mechanism for the grouping are employed to enhance the robustness of FFD missions when some UAVs are destroyed, hence maintaining the high performance of FFD missions in harsh environments. Moreover, TSPF enables submodel updates by aggregating the parameters of selected layers within/between UAV groups, thereby effectively reducing the model parameter uploads (communication overhead) in model training. Experimental evaluations demonstrate that our proposed TSPF significantly improves the detection accuracy of forest fires, enhances the robustness of FFD missions against the destruction of some UAVs, and reduces the communication overhead in FFD missions.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"25 1","pages":"1169-1183"},"PeriodicalIF":9.2,"publicationDate":"2025-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145659188","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 : 2025-07-31DOI: 10.1109/TMC.2025.3594488
Zhenning Wang;Yue Cao;Huan Zhou;Xiaokang Zhou;Jiawen Kang;Houbing Song
Recently, For-Hire Vehicles (FHVs) have emerged as major players in Vehicular CrowdSensing (VCS). However, the heterogeneity of tasks issued by Data Requesters (DRs) and the heterogeneity of sensors equipped on FHVs under different Vehicle Platforms (VPs) bring difficulties to task allocation and execution. It can be concluded that it is important to reasonably analyze the relationship among DRs, VPs, and FHVs, as well as to motivate VPs and FHVs to complete sensing tasks. Therefore, taking advantage of the real-time simulation and intelligent decision-making of Digital Twins (DT), this paper proposes a DT-driven Double-layer Reverse Auction Method (DRAM). In the first layer, the reverse auction is established between each DR and VPs, and in the second layer, the reverse auction is established between each VP and FHVs. Meanwhile, we also introduce a sensing fairness index to ensure the sensing balance of different sub-regions and consider it in the DRAM process. Here, the idea of backward induction is used to solve the above problems, with the goal of minimizing the overhead of winning VP and the average overhead of all DRs. Finally, the effectiveness of the DRAM proposed in this paper is verified based on the real data set. Compared with the baseline method, DRAM can reduce the average overhead of DR by about 4%-25%. Meanwhile, in terms of sensing fairness, it can be improved by up to 55%.
{"title":"DRAM: Digital Twin-Driven Double-Layer Reverse Auction Method for Multi-Platform Vehicular Crowdsensing","authors":"Zhenning Wang;Yue Cao;Huan Zhou;Xiaokang Zhou;Jiawen Kang;Houbing Song","doi":"10.1109/TMC.2025.3594488","DOIUrl":"https://doi.org/10.1109/TMC.2025.3594488","url":null,"abstract":"Recently, For-Hire Vehicles (FHVs) have emerged as major players in Vehicular CrowdSensing (VCS). However, the heterogeneity of tasks issued by Data Requesters (DRs) and the heterogeneity of sensors equipped on FHVs under different Vehicle Platforms (VPs) bring difficulties to task allocation and execution. It can be concluded that it is important to reasonably analyze the relationship among DRs, VPs, and FHVs, as well as to motivate VPs and FHVs to complete sensing tasks. Therefore, taking advantage of the real-time simulation and intelligent decision-making of Digital Twins (DT), this paper proposes a DT-driven <underline>D</u>ouble-layer <underline>R</u>everse <underline>A</u>uction <underline>M</u>ethod (DRAM). In the first layer, the reverse auction is established between each DR and VPs, and in the second layer, the reverse auction is established between each VP and FHVs. Meanwhile, we also introduce a sensing fairness index to ensure the sensing balance of different sub-regions and consider it in the DRAM process. Here, the idea of backward induction is used to solve the above problems, with the goal of minimizing the overhead of winning VP and the average overhead of all DRs. Finally, the effectiveness of the DRAM proposed in this paper is verified based on the real data set. Compared with the baseline method, DRAM can reduce the average overhead of DR by about 4%-25%. Meanwhile, in terms of sensing fairness, it can be improved by up to 55%.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 12","pages":"13725-13742"},"PeriodicalIF":9.2,"publicationDate":"2025-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145442694","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 : 2025-07-28DOI: 10.1109/TMC.2025.3591561
Rui Xing;Zhenzhe Zheng;Fan Wu;Guihai Chen
Large language models (LLMs) exhibit remarkable capabilities in natural language understanding and generation. However, the accuracy of the inference depends deeply on the contexts of queries, especially for personal services. Abundant mobile sensing data collected by sensors embedded in smart devices can proactively capture real-time user contexts. However, raw sensing data are low-quality (e.g., existing missing data and data redundancy) and are incapable of providing accurate contexts. In this work, we present ConGen, a user context generation framework for LLMs, aiming at prompting users’ contexts through their implicit mobile sensing information. ConGen integrates two components: refined data completion and multi-granularity context compression. Specifically, the refined data completion couples data-centric feature selection by leveraging the eXplainable AI (XAI) method into the data imputation model to generate fewer but more informative features for efficient and effective context generation. Additionally, we implement multi-granularity context compression, reducing timestep- and context-level data redundancy while further elevating context quality. Experiment results show that ConGen can generate more accurate context, surpassing competitive baselines by 1.3%-8.3% in context inference on all four datasets. Moreover, context compression significantly reduces redundancy to $1/70sim 1/40$ of the original data amount, and further improves the context accuracy. Finally, the enhanced performance of LLMs, as demonstrated by both quantitative and qualitative evaluations of prompting ConGen-generated user contexts, underscores the effectiveness of ConGen.
{"title":"User Context Generation for Large Language Models From Mobile Sensing Data","authors":"Rui Xing;Zhenzhe Zheng;Fan Wu;Guihai Chen","doi":"10.1109/TMC.2025.3591561","DOIUrl":"https://doi.org/10.1109/TMC.2025.3591561","url":null,"abstract":"Large language models (LLMs) exhibit remarkable capabilities in natural language understanding and generation. However, the accuracy of the inference depends deeply on the contexts of queries, especially for personal services. Abundant mobile sensing data collected by sensors embedded in smart devices can proactively capture real-time user contexts. However, raw sensing data are low-quality (e.g., existing missing data and data redundancy) and are incapable of providing accurate contexts. In this work, we present ConGen, a user context generation framework for LLMs, aiming at prompting users’ contexts through their implicit mobile sensing information. ConGen integrates two components: refined data completion and multi-granularity context compression. Specifically, the refined data completion couples data-centric feature selection by leveraging the eXplainable AI (XAI) method into the data imputation model to generate fewer but more informative features for efficient and effective context generation. Additionally, we implement multi-granularity context compression, reducing timestep- and context-level data redundancy while further elevating context quality. Experiment results show that ConGen can generate more accurate context, surpassing competitive baselines by 1.3%-8.3% in context inference on all four datasets. Moreover, context compression significantly reduces redundancy to <inline-formula><tex-math>$1/70sim 1/40$</tex-math></inline-formula> of the original data amount, and further improves the context accuracy. Finally, the enhanced performance of LLMs, as demonstrated by both quantitative and qualitative evaluations of prompting ConGen-generated user contexts, underscores the effectiveness of ConGen.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 12","pages":"13678-13695"},"PeriodicalIF":9.2,"publicationDate":"2025-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145442695","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 : 2025-07-25DOI: 10.1109/TMC.2025.3592929
Shaoran Li;Nan Jiang;Chengzhang Li;Shiva Acharya;Yubo Wu;Weijun Xie;Wenjing Lou;Y. Thomas Hou
MU-MIMO beamforming is a key technology for 5G networks, relying on Channel State Information (CSI). However, in practice, the estimated CSI in reality is prone to uncertainty. Further, a MU-MIMO beamforming solution must be derived within a millisecond to be useful for real-time 5G applications. We present ReDBeam—a real-time data-driven beamforming solution for MU-MIMO using limited CSI data samples. The main novelties of ReDBeam are a parallel algorithm and an optimized GPU implementation. ReDBeam delivers a MU-MIMO beamforming solution within 1 millisecond to meet the probabilistic data rate requirements from the users, and minimize a base station’s power consumption. Through extensive experiments, we show that ReDBeam consistently meets the stringent 1-millisecond real-time requirement and is orders of magnitude faster than other state-of-the-art algorithms. ReDBeam conclusively demonstrates that MU-MIMO beamforming with data rate requirements can be achieved in real-time using only limited CSI data samples.
{"title":"Real-Time MU-MIMO Beamforming With Limited Channel Samples in 5G Networks","authors":"Shaoran Li;Nan Jiang;Chengzhang Li;Shiva Acharya;Yubo Wu;Weijun Xie;Wenjing Lou;Y. Thomas Hou","doi":"10.1109/TMC.2025.3592929","DOIUrl":"https://doi.org/10.1109/TMC.2025.3592929","url":null,"abstract":"MU-MIMO beamforming is a key technology for 5G networks, relying on Channel State Information (CSI). However, in practice, the estimated CSI in reality is prone to uncertainty. Further, a MU-MIMO beamforming solution must be derived within a millisecond to be useful for real-time 5G applications. We present ReDBeam—a real-time data-driven beamforming solution for MU-MIMO using limited CSI data samples. The main novelties of ReDBeam are a parallel algorithm and an optimized GPU implementation. ReDBeam delivers a MU-MIMO beamforming solution within 1 millisecond to meet the probabilistic data rate requirements from the users, and minimize a base station’s power consumption. Through extensive experiments, we show that ReDBeam consistently meets the stringent 1-millisecond real-time requirement and is orders of magnitude faster than other state-of-the-art algorithms. ReDBeam conclusively demonstrates that MU-MIMO beamforming with data rate requirements can be achieved in real-time using only limited CSI data samples.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 12","pages":"13549-13566"},"PeriodicalIF":9.2,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145442707","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 : 2025-07-24DOI: 10.1109/TMC.2025.3592334
Borui Li;Tiange Xia;Weilong Wang;Jingyuan Zhang;Shuai Wang;Chenhong Cao;Zheng Dong;Shuai Wang
Mobile applications based on machine learning (ML) are increasingly relying on offloading to the edge devices for low-latency, resource-efficient computation. Applying serverless computing for these ML applications on the edge offers a promising solution for handling dynamic workloads while meeting user-specified latency service-level objectives (SLOs). However, existing serverless frameworks, with their coarse-grained data parallelism and rigid model partitioning, are inadequate for ML inference on widely adopted edge System-on-Chip (SoC) devices. This paper presents FluidEdge, an edge-native serverless inference framework. FluidEdge identifies bottleneck operators in ML models and addresses them through a novel fine-grained intra-function latency-sensitive auto-scaling approach that dynamically scales inference bottlenecks during online serving. Additionally, it employs inter-function scaling to further prevent latency SLO violations and leverages the unified memory of edge SoCs for efficient data sharing during inference. Experimental results demonstrate that FluidEdge achieves a 37.4% latency improvement and 67.3% -87.6% SLO violation reduction compared to best-performed state-of-the-art serverless inference frameworks.
{"title":"FluidEdge: Expediting Serverless Machine Learning Inference via Bottleneck-Aware Auto-Scaling on Edge SoCs","authors":"Borui Li;Tiange Xia;Weilong Wang;Jingyuan Zhang;Shuai Wang;Chenhong Cao;Zheng Dong;Shuai Wang","doi":"10.1109/TMC.2025.3592334","DOIUrl":"https://doi.org/10.1109/TMC.2025.3592334","url":null,"abstract":"Mobile applications based on machine learning (ML) are increasingly relying on offloading to the edge devices for low-latency, resource-efficient computation. Applying serverless computing for these ML applications on the edge offers a promising solution for handling dynamic workloads while meeting user-specified latency service-level objectives (SLOs). However, existing serverless frameworks, with their coarse-grained data parallelism and rigid model partitioning, are inadequate for ML inference on widely adopted edge System-on-Chip (SoC) devices. This paper presents FluidEdge, an edge-native serverless inference framework. FluidEdge identifies bottleneck operators in ML models and addresses them through a novel fine-grained intra-function latency-sensitive auto-scaling approach that dynamically scales inference bottlenecks during online serving. Additionally, it employs inter-function scaling to further prevent latency SLO violations and leverages the unified memory of edge SoCs for efficient data sharing during inference. Experimental results demonstrate that FluidEdge achieves a 37.4% latency improvement and 67.3% -87.6% SLO violation reduction compared to best-performed state-of-the-art serverless inference frameworks.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 12","pages":"13586-13599"},"PeriodicalIF":9.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145442687","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 : 2025-07-22DOI: 10.1109/TMC.2025.3590775
Xinjie Wang;Yifan Zhang;Xinpu Liu;Ke Xu;Jianwei Wan;Yulan Guo;Hanyun Wang
Efficiently compressing large-scale point cloud data under limited bandwidth and computing resource conditions has become a critical issue to be addressed in mobile computing platforms. Although the octree structure can efficiently represent large-scale and complex point clouds, existing octree-based Point Cloud Geometry Compression (PCGC) approaches typically focus on exploiting either spatial or channel features individually, neglecting the interaction across spatial-channel dimensions. In addition, current approaches are also limited to small-scale point clouds due to reliance on global Transformer or local convolutional neural network (CNN). To solve these issues, we introduce GCFI-Net, a global-local cross-spatial-channel feature interaction network for predicting the occupancy probability distribution of each octree node in this paper. In the GCFI-Net, we propose a Multiscale Convolutional Fusion-based Spatial Interaction (MCFSI) module to capture global context and model spatial interactions, and a Global-Local Cross-Channel Interaction (GLCCI) module with dual pathways to integrate global and local cross-channel information. Additionally, we propose a Multiscale-enhanced Spatial and Channel Interaction (MSCI) module to aggregate features from ancestor and sibling nodes, which further enhances the octree node representation ability. Extensive experiments on large-scale sparse LiDAR and dense human body point clouds demonstrate that the proposed GCFI-Net achieves superior compression performance with fewer parameters compared to state-of-the-art PCGC methods.
{"title":"GCFI-Net: Global-Local Cross-Spatial-Channel Feature Interaction Network for Point Cloud Geometry Compression","authors":"Xinjie Wang;Yifan Zhang;Xinpu Liu;Ke Xu;Jianwei Wan;Yulan Guo;Hanyun Wang","doi":"10.1109/TMC.2025.3590775","DOIUrl":"https://doi.org/10.1109/TMC.2025.3590775","url":null,"abstract":"Efficiently compressing large-scale point cloud data under limited bandwidth and computing resource conditions has become a critical issue to be addressed in mobile computing platforms. Although the octree structure can efficiently represent large-scale and complex point clouds, existing octree-based Point Cloud Geometry Compression (PCGC) approaches typically focus on exploiting either spatial or channel features individually, neglecting the interaction across spatial-channel dimensions. In addition, current approaches are also limited to small-scale point clouds due to reliance on global Transformer or local convolutional neural network (CNN). To solve these issues, we introduce GCFI-Net, a global-local cross-spatial-channel feature interaction network for predicting the occupancy probability distribution of each octree node in this paper. In the GCFI-Net, we propose a Multiscale Convolutional Fusion-based Spatial Interaction (MCFSI) module to capture global context and model spatial interactions, and a Global-Local Cross-Channel Interaction (GLCCI) module with dual pathways to integrate global and local cross-channel information. Additionally, we propose a Multiscale-enhanced Spatial and Channel Interaction (MSCI) module to aggregate features from ancestor and sibling nodes, which further enhances the octree node representation ability. Extensive experiments on large-scale sparse LiDAR and dense human body point clouds demonstrate that the proposed GCFI-Net achieves superior compression performance with fewer parameters compared to state-of-the-art PCGC methods.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 12","pages":"13663-13677"},"PeriodicalIF":9.2,"publicationDate":"2025-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145442746","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：