Journal of Systems Architecture最新文献_第2页

Improving UI responsiveness in Android by restructured rendering 通过重构渲染改善Android的UI响应能力

IF 4.1 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-09-16 DOI: 10.1016/j.sysarc.2025.103580

Mingsong Lv , Tao Hu , Menglong Cui , Tao Yang , Yiyang Zhou , Qingxu Deng , Nan Guan

{"title":"Improving UI responsiveness in Android by restructured rendering","authors":"Mingsong Lv , Tao Hu , Menglong Cui , Tao Yang , Yiyang Zhou , Qingxu Deng , Nan Guan","doi":"10.1016/j.sysarc.2025.103580","DOIUrl":"10.1016/j.sysarc.2025.103580","url":null,"abstract":"<div><div>Mobile operating systems, such as Android, are increasingly used across diverse applications, where ensuring high responsiveness to user interactions is critical, particularly in mission-critical and real-time scenarios. Mobile operating systems typically process user interaction events and UI rendering on the same thread, commonly referred to as the main thread of a mobile application. As a result, user interaction handling can face significant delays when blocked by overloaded UI rendering tasks, compromising responsiveness. Existing mobile operating systems lack effective mechanisms to mitigate this issue. This paper addresses the problem by restructuring the UI rendering workflow to improve responsiveness in the presence of heavy rendering workloads. Specifically, two techniques are proposed that are tailored to whether the event handling results require screen display. Experimental results demonstrate improvements in both average-case and worst-case response times of event handling, enhancing the UI responsiveness. Although the implementation focuses on Android, the proposed approaches are adaptable to other mobile operating systems with similar rendering architectures, such as iOS and HarmonyOS.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"168 ","pages":"Article 103580"},"PeriodicalIF":4.1,"publicationDate":"2025-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145117618","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}

引用次数: 0

Trajectory design for data collection under insufficient UAV energy: A staged actor–critic reinforcement learning approach 无人机能量不足情况下数据收集的轨迹设计：一种分阶段行为批判强化学习方法

IF 4.1 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-09-15 DOI: 10.1016/j.sysarc.2025.103566

Jing Mei , Yuejia Zhang , Zhao Tong , Keqin Li

{"title":"Trajectory design for data collection under insufficient UAV energy: A staged actor–critic reinforcement learning approach","authors":"Jing Mei , Yuejia Zhang , Zhao Tong , Keqin Li","doi":"10.1016/j.sysarc.2025.103566","DOIUrl":"10.1016/j.sysarc.2025.103566","url":null,"abstract":"<div><div>Fixed-wing unmanned aerial vehicles (UAVs) offer distinct advantages for large-scale environmental sensor data collection. In forest and marine scenarios, UAVs typically depart from a fixed location, collecting data along a route, and return. Unlike existing work aiming to minimizing energy consumption on data collection task, this study focus on the scenario where a UAV’s initial energy may not be sufficient to visit all sensor nodes. We aim to maximize data collection under insufficient battery energy while make a safety return. To solve this, we adopt the twin delayed deep deterministic policy gradient (TD3) algorithm with three designed reward functions, and introduce a stage-based safe action algorithm, termed Staged Safe-Action TD3 (SS-TD3). An energy consumption model incorporating acceleration and a segmented time model are used to enhance exploration efficiency. To tackle sparse binary rewards and the suboptimal convergence of complex reward function in reinforcement learning, a staged training approach, Staged Actor–Critic based reinforcement Learning (S-ACL) is proposed, as the one of the fundamental component of SS-TD3. Experimental results show that SS-TD3 achieves the best energy efficiency compared to baselines, while S-ACL significantly improves policy performance in complex reward environments.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"168 ","pages":"Article 103566"},"PeriodicalIF":4.1,"publicationDate":"2025-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145104799","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}

引用次数: 0

Log2Evt: Constructing high-level events for IoT Systems through log-code execution path correlation Log2Evt：通过日志代码执行路径关联为物联网系统构建高级事件

IF 4.1 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-09-13 DOI: 10.1016/j.sysarc.2025.103578

Teng Li , Baichuan Zheng , Yebo Feng , Xiaowen Quan , Jiahua Xu , Yang Liu , Jianfeng Ma

{"title":"Log2Evt: Constructing high-level events for IoT Systems through log-code execution path correlation","authors":"Teng Li , Baichuan Zheng , Yebo Feng , Xiaowen Quan , Jiahua Xu , Yang Liu , Jianfeng Ma","doi":"10.1016/j.sysarc.2025.103578","DOIUrl":"10.1016/j.sysarc.2025.103578","url":null,"abstract":"<div><div>The detection of cyberattacks in IoT ecosystems requires comprehensive log auditing across distributed devices, yet the volume and heterogeneity of IoT logs exceed traditional analysis capabilities. Therefore, it is essential to narrow down the scope of forensics precisely and efficiently to target attack-related events. Existing schemes have the disadvantage of low accuracy and flexibility. We propose a novel approach that synthesizes high-level security events from low-level IoT logs by correlating firmware execution traces with runtime call stack contexts. Our approach implements lightweight monitoring probes at critical IoT workflow points and employs an IoT-optimized Common Ancestor algorithm for log sequence analysis. The experiments demonstrate a 15% improvement in accuracy compared to the rule-based matching scheme. Additionally, the results highlight the influence of the threshold parameter and show that the approach has minimal impact on program operation. The approach effectively addresses the challenges of protocol fragmentation and resource constraints in IoT environments, providing a foundation for robust security monitoring in smart city deployments.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"168 ","pages":"Article 103578"},"PeriodicalIF":4.1,"publicationDate":"2025-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145157979","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}

引用次数: 0

Balancing privacy and fairness: Client selection in differential privacy-based federated learning 平衡隐私和公平：基于差异隐私的联邦学习中的客户端选择

IF 4.1 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-09-12 DOI: 10.1016/j.sysarc.2025.103576

Xu Zhao , Gang Li , Yuan Yao , Bo Cui

{"title":"Balancing privacy and fairness: Client selection in differential privacy-based federated learning","authors":"Xu Zhao , Gang Li , Yuan Yao , Bo Cui","doi":"10.1016/j.sysarc.2025.103576","DOIUrl":"10.1016/j.sysarc.2025.103576","url":null,"abstract":"<div><div>Federated Learning (FL) tackles the data island problem by enabling collaborative model training across clients while safeguarding privacy. However, most existing work only cared about the parameter privacy protection, while ignoring the model training efficay. To this end, in this paper, we propose AdaDPCS-FL, an Adaptive Budget Allocation and Client Selection method tailored for Federated Learning. This approach consists of two steps, where in the step, an adaptive privacy budget allocation strategy based on model similarity and a reversion mechanism are designed to speed up training convergence while still keeping privacy preservation in FL. In the second step, addressing the fairness of client selection in the FL process, it proposes a contribution-based online client selection mechanism is further proposed with the consideration of the fairness of client selection, in which, a multi-armed bandit scheme is tailored to optimize the client selection. Theoretically, the proposed method satisfies the properties of differential privacy, convergence guarantee, and a constant upper bound on cumulative regret <span><math><mrow><mi>O</mi><mrow><mo>(</mo><mi>K</mi><msqrt><mrow><mo>ln</mo><mi>R</mi><mo>ln</mo><mi>R</mi></mrow></msqrt><mo>)</mo></mrow></mrow></math></span>. Experiments on real datasets demonstrate superior performance over baselines like FedProx and FedAvg. Moreover, with the privacy guaranteed, the test accuracy by our proposed method can be improved approximately 4% compared to DP-FedAvg and FedBDP in heterogeneous settings.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"168 ","pages":"Article 103576"},"PeriodicalIF":4.1,"publicationDate":"2025-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145060426","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}

引用次数: 0

Efficient Dropout-resilient and Verifiable Federated learning scheme for AIoT healthcare system 面向AIoT医疗系统的高效抗辍学可验证联邦学习方案

IF 4.1 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-09-12 DOI: 10.1016/j.sysarc.2025.103575

Yazhe Kang , Yong Xie , Libing Wu , Pingchao Zhou , Yue Du

{"title":"Efficient Dropout-resilient and Verifiable Federated learning scheme for AIoT healthcare system","authors":"Yazhe Kang , Yong Xie , Libing Wu , Pingchao Zhou , Yue Du","doi":"10.1016/j.sysarc.2025.103575","DOIUrl":"10.1016/j.sysarc.2025.103575","url":null,"abstract":"<div><div>Federated learning (FL) with healthcare AIoT data from multiple healthcare institutions can enhance device intelligence and further protect people’s health. The preservation of confidentiality in healthcare data has emerged as a central concern in healthcare FL systems. Under the premise of emphasizing privacy protection, it is particularly inefficient to maintain system robustness after a participant drops out. Furthermore, it proves more complex to confirm whether the aggregation server faithfully executes the aggregation task, especially given that the aggregation server may collude with corrupt users to intentionally return carefully crafted aggregation results. At present, few works focus on the two issues concurrently. To address this challenge, an efficient dropout-resilient and verifiable FL scheme (EDV-FL for short) is proposed in this paper. Our scheme addresses the issue of dropped users rejoining in the future, while reducing both communication and computation overhead. Moreover, we ensure that even if the server colludes with corrupt users to forge the aggregation result, users can still detect the correctness of the aggregation result. We theoretically demonstrate the effectiveness of EDV-FL and reproduce the scheme using Convolutional Neural Network (CNN) models on the MNIST, CIFAR-10, and Fashion-MNIST datasets. Theoretical proofs and experimental analyses show that our EDV-FL is an efficient, dropout-resistant, and collusion-resistant, verifiable FL scheme.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"168 ","pages":"Article 103575"},"PeriodicalIF":4.1,"publicationDate":"2025-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145104800","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}

引用次数: 0

SMCCA: A sharded multi-task collaborative consensus algorithm for unmanned vehicle networks SMCCA：一种面向无人驾驶车辆网络的分片多任务协同共识算法

IF 4.1 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-09-09 DOI: 10.1016/j.sysarc.2025.103572

Song Peng , Yongming Fu , Yingwen Chen , Mengyuan Zhu , Huan Zhou , Jiachao Wang , Jinshu Su

{"title":"SMCCA: A sharded multi-task collaborative consensus algorithm for unmanned vehicle networks","authors":"Song Peng , Yongming Fu , Yingwen Chen , Mengyuan Zhu , Huan Zhou , Jiachao Wang , Jinshu Su","doi":"10.1016/j.sysarc.2025.103572","DOIUrl":"10.1016/j.sysarc.2025.103572","url":null,"abstract":"<div><div>Consensus mechanisms are a critical technology in modern networked collaborative systems and play a pivotal role in emerging scenarios such as Unmanned Vehicle Networks (UVNs). However, the escalating complexity and scale of unmanned systems have exposed two principal bottlenecks in existing consensus algorithms: (1) high communication overhead resulting from global consensus mechanisms, and (2) inadequate handling of intricate inter-task constraints, which severely limits the practical deployment of UVNs. To address these challenges, we propose a Sharded Multi-task Collaborative Consensus Algorithm (SMCCA), which employs a divide-and-conquer approach to decouple traditional global consensus problems into parallelizable local consensus sub-problems, thereby optimizing communication costs and enhancing task coordination. Specifically, SMCCA ensures efficient collaboration within UVNs through three core mechanisms. First, the algorithm partitions unmanned devices into multiple autonomous yet interoperable clusters using a sharding architecture, where leader shards act as global coordinators to enable fine-grained cross-shard task cooperation. Second, a Task Relationship Graph (TRG) is constructed to precisely quantify the dependencies and conflicts among tasks. Based on this, a hierarchical topological sorting algorithm is applied to generate an optimal execution sequence without constrained interactions. Third, a hierarchical global task view based on a Directed Acyclic Graph (DAG) is established to support efficient fault recovery and consistent state maintenance. Experimental results demonstrate that in a network with 256 nodes distributed across four shards, SMCCA achieves a throughput of 11,835 tasks per second and an average latency of approximately 0.5 s, significantly outperforming existing consensus algorithms.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"168 ","pages":"Article 103572"},"PeriodicalIF":4.1,"publicationDate":"2025-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145048344","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}

引用次数: 0

Lightweight multi-layered de-identification architecture: Secure client selection in federated learning 轻量级多层去标识架构：联邦学习中的安全客户端选择

IF 4.1 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-09-09 DOI: 10.1016/j.sysarc.2025.103569

Jiheon Choi, Sangyoon Oh

{"title":"Lightweight multi-layered de-identification architecture: Secure client selection in federated learning","authors":"Jiheon Choi, Sangyoon Oh","doi":"10.1016/j.sysarc.2025.103569","DOIUrl":"10.1016/j.sysarc.2025.103569","url":null,"abstract":"<div><div>Despite federated learning (FL) let us avoid raw data-sharing, recent studies shows that existing client selection schemes poses privacy vulnerabilities that might be led to organizational identity leaks. To address this problem, we present a lightweight, multi-layered de-identification architecture that enables privacy-preserving client selection while maintaining selection efficiency in FL environments. The architecture comprises lightweight dynamic hash-based identifier, a secure salt distribution protocol, and an enhanced, resizable Bloom filter verifier. Together, these components provide three critical security properties, formal k-anonymity, <span><math><mi>ϵ</mi></math></span>-connection resistance, and <span><math><mi>α</mi></math></span>-membership confusion, protecting FL training against dictionary attacks, cross-round linkability attacks, and membership-inference attacks. We also offer a theoretical framework that allows fine-grained control of the privacy-efficiency trade-off through adjustable parameters. Empirical experiment on MNIST, Fashion-MNIST, CIFAR-10) with Non-IID datasets and heterogeneous client environments show that our method keeps model accuracy within 0.02 to 15.3 pp of FedAvg and Power-of-Choice while adding only 3.6% to 6.3% of computation overhead. These results demonstrate the effectiveness of our approach and show that balanced client selection between performance and privacy is achievable.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"168 ","pages":"Article 103569"},"PeriodicalIF":4.1,"publicationDate":"2025-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145048346","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}

引用次数: 0

Bit-level compiler optimization for ultra low-power embedded systems 超低功耗嵌入式系统的位级编译器优化

IF 4.1 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-09-08 DOI: 10.1016/j.sysarc.2025.103546

Seonyeong Heo , Jiho Kim , Woohyeop Im , Jiyun Moon , Daehee Jang

{"title":"Bit-level compiler optimization for ultra low-power embedded systems","authors":"Seonyeong Heo , Jiho Kim , Woohyeop Im , Jiyun Moon , Daehee Jang","doi":"10.1016/j.sysarc.2025.103546","DOIUrl":"10.1016/j.sysarc.2025.103546","url":null,"abstract":"<div><div>Achieving ultra low-power consumption is essential for embedded systems deployed in harsh environments, such as space and deep sea locations, where energy resources are scarce and physical accessibility is limited. Typically, these systems employ ultra low-power microcontrollers that operate on narrow data widths of 8 or 16 bits at the microarchitecture level. If software developers do not carefully consider the data widths during programming, the resulting programs may be suboptimally optimized for these ultra low-power systems. To address this issue and enable more efficient low-power computing, this work proposes a novel optimizing compiler that supports bit-level analyses and transformations. The proposed compiler analyzes how each individual bit of a data item is utilized within a program to determine its optimal width. Consequently, the proposed compiler reduces unnecessary data movements and computational overhead on ultra low-power processors. This work implements the prototype compiler on top of the LLVM compiler framework and evaluates the performance impact of the optimized embedded applications with a processor simulator.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"168 ","pages":"Article 103546"},"PeriodicalIF":4.1,"publicationDate":"2025-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145104798","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}

引用次数: 0

LODAP: On-device incremental learning via lightweight operations and data pruning LODAP：通过轻量级操作和数据修剪在设备上进行增量学习

IF 4.1 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-09-08 DOI: 10.1016/j.sysarc.2025.103571

Biqing Duan , Qing Wang , Di Liu , Wei Zhou , Zhenli He , Shengfa Miao

{"title":"LODAP: On-device incremental learning via lightweight operations and data pruning","authors":"Biqing Duan , Qing Wang , Di Liu , Wei Zhou , Zhenli He , Shengfa Miao","doi":"10.1016/j.sysarc.2025.103571","DOIUrl":"10.1016/j.sysarc.2025.103571","url":null,"abstract":"<div><div>Incremental learning, which enables a deployed model to continually learn new classes over time, is becoming increasingly crucial for industrial edge systems, where communicating with remote servers for computation-intensive training is often difficult. As edge devices are expected to continue learning more classes after deployment, designing efficient on-device learning frameworks is essential. In this paper, we propose <span>LODAP</span>, a lightweight on-device incremental learning framework for edge systems. The key part of <span>LODAP</span> is a new module, namely Efficient Incremental Module (EIM), which combines normal convolutions with lightweight operations. During incremental learning, <span>EIM</span> employs <em>adapters</em> to effectively and efficiently learn features for new classes to improve the accuracy of incremental learning while reducing model complexity and training overhead. To further improve efficiency, <span>LODAP</span> integrates a data pruning strategy that removes redundant training samples, significantly lowering the training cost. We conducted extensive experiments on the CIFAR-100, Tiny-ImageNet, and CUB-200-2011 datasets. Experimental results show that <span>LODAP</span> improves the accuracy by up to 4.32% over existing methods while reducing around 50% of model complexity. In addition, evaluations on real edge systems demonstrate its applicability for on-device machine learning. The code is available at <span><span>https://github.com/duanbiqing/LODAP</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"168 ","pages":"Article 103571"},"PeriodicalIF":4.1,"publicationDate":"2025-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145048345","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}

引用次数: 0

LUFT-CAN: A lightweight unsupervised learning based intrusion detection system with frequency-time analysis for vehicular CAN bus LUFT-CAN：基于轻型无监督学习的车载CAN总线频率时间分析入侵检测系统

IF 4.1 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-09-07 DOI: 10.1016/j.sysarc.2025.103567

Yu Xin , Xiaohang Wang , Li Lu , Shuguo Zhuo , Yingtao Jiang , Amit Kumar Singh , Kui Ren , Mei Yang , Kaiwei Wu

{"title":"LUFT-CAN: A lightweight unsupervised learning based intrusion detection system with frequency-time analysis for vehicular CAN bus","authors":"Yu Xin , Xiaohang Wang , Li Lu , Shuguo Zhuo , Yingtao Jiang , Amit Kumar Singh , Kui Ren , Mei Yang , Kaiwei Wu","doi":"10.1016/j.sysarc.2025.103567","DOIUrl":"10.1016/j.sysarc.2025.103567","url":null,"abstract":"<div><div>The Controller Area Network (CAN) bus is critical for data transmission among electronic control units (ECUs) in modern vehicles, necessitating robust intrusion detection systems (IDS) for security. However, existing IDS approaches have several limitations. For example, rule based IDS methods depend on proprietary protocol knowledge, while most machine learning approaches rely on supervised training using outdated or limited datasets, hindering their ability to detect emerging threats. Furthermore, deep learning based IDS models often have high computational complexity, making them unsuitable for resource-constrained vehicular environments. To overcome these challenges, we propose LUFT-CAN, a novel, lightweight, unsupervised IDS that integrates frequency and time domain analysis of CAN traffic. By leveraging spectral characteristics of CAN ID sequences, LUFT-CAN effectively distinguishes between normal and anomalous traffic patterns. A tailored neural network architecture extracts these features, and the system is optimized via quantization-aware training for real-time inference on embedded systems. Experiments performed on datasets collected from modern vehicles, Tesla Model 3 2022 and LeapMotor C10 2024 as well as a public benchmark dataset demonstrate that LUFT-CAN achieves promising F1-scores of 97.1% and 96.7%, significantly outperforming previous approaches. We implemented the proposed IDS on a 2024 LeapMotor C10 test vehicle equipped with a Qualcomm 8295 microcontroller unit(MCU). The model’s inference time is 14.27 s per 100,000 frames, demonstrating its effectiveness and efficiency for in-vehicle deployment.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"168 ","pages":"Article 103567"},"PeriodicalIF":4.1,"publicationDate":"2025-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145026994","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}

引用次数: 0