{"title":"RPS-DFN: Residual perception self-attention deep fusion network for multimodal IIoT device state identification","authors":"Anying Chai, Zhaobo Fang, Ping Huang, Chenyang Guo, Lei Wang, Wanda Yin","doi":"10.1016/j.iot.2025.101790","DOIUrl":"10.1016/j.iot.2025.101790","url":null,"abstract":"<div><div>The Industrial Internet of Things (IIoT) integrates advanced technologies such as Internet of Things (loT) technology and Artificial Intelligence (Al) into various aspects of industrial production and achieves accurate identification of equipment status through the deployment of a large number of sensors. However, due to the heterogeneity of data and the limitations of traditional data fusion methods, which often overlook cross-modal interactions and feature contributions, leading to poor fusion performance. In this paper, we propose an end-to-end Residual Perceptive Self-attention Deep Fusion Network (RPS-DFN) to effectively fuse time-series signals such as force, vibration, and acoustic emission with device images captured at the same time. We propose a multi-modal data unification method based on Mel-spectrogram transformation to align the dimensions of signals and images. Then, we improve the ResNet18 pre-trained on ImageNet by designing a shared dimensionality reduction layer and a cross-modal attention module. The general visual representations learned by its pre-trained weights can be transferred to the small-sample equipment status detection task, enhancing the differences between features of different statuses. Finally, we design a two-layer Transformer encoder to learn the contributions and interactions of different features for downstream tasks, modeling and analyzing different features to achieve self-attentive deep fusion of features. The experimental results show that our method achieves an accuracy of 95.24% on PHM2010 and 75.71% on the cross-tool status detection task on the small-sample Mudestera dataset, verifying the practical applicability of the proposed method.</div></div>","PeriodicalId":29968,"journal":{"name":"Internet of Things","volume":"34 ","pages":"Article 101790"},"PeriodicalIF":7.6,"publicationDate":"2025-10-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145266676","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}
{"title":"High-level petri nets based approach for designing and analyzing IoT systems","authors":"Abdelouahab Fortas , Elhillali Kerkouche , Allaoua Chaoui","doi":"10.1016/j.iot.2025.101783","DOIUrl":"10.1016/j.iot.2025.101783","url":null,"abstract":"<div><div>In the last decade, the Internet of Things (IoT) has emerged as a concept for intelligent systems that touch various areas of life, such as health, industry, smart cities, and many other fields. Failure of these systems can lead to catastrophic losses that affect people’s lives, the environment, health, and the economy. Therefore, it is necessary to analyze these IoT systems. However, the complexity of IoT systems makes modeling and analyzing these systems challenging. Formal verification methods offer a rigorous approach to analyzing and verifying complex systems by mathematically proving their correctness. This paper proposes a new approach based on high-level Petri nets (G-Nets) for designing and analyzing IoT systems. G-Nets support a modular or compositional methodology. Using a modular method in IoT development is particularly appropriate due to the inherently modular nature of IoT systems themselves. This approach addresses IoT challenges by offering essential benefits such as flexibility, scalability, ease of maintenance, reusability, and improved complexity management. However, the G-Nets formalism requires more specialized and sophisticated tools to verify and analyze its specifications. To bridge this gap, we propose and implement formal semantics for the G-Nets formalism using the Maude language. The Maude tools provide powerful analysis techniques that enable rigorous analysis and verification of G-Nets specifications. We illustrated our approach with a practical example of an IoT-based home camera system.</div></div>","PeriodicalId":29968,"journal":{"name":"Internet of Things","volume":"34 ","pages":"Article 101783"},"PeriodicalIF":7.6,"publicationDate":"2025-10-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145266678","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}
{"title":"An innovative intrusion detection framework using GAN-augmented Deep Ensemble Neural Network for cross-domain IoT–cloud security","authors":"Sadia Nazim , Syed Shujaa Hussain , Bilal Yousuf , Saima Sultana , Eraj Tanweer","doi":"10.1016/j.iot.2025.101773","DOIUrl":"10.1016/j.iot.2025.101773","url":null,"abstract":"<div><div>The rising popularity of smart cities and their impact across multiple sectors, including healthcare, transportation, and industry, is due to the rapid expansion of the Internet of Things (IoT). The growing popularity of IoT environments has made them susceptible to a wide array of cybersecurity hazards, such as denial-of-service (DoS), brute-force, and malicious access assaults. Robust intrusion detection and forensic investigation approaches must be developed to counter the aforementioned hazards. These frameworks primarily benefit from authentic and well-organized datasets for successful training and validation.</div><div>This study introduces an innovative Generative Adversarial Networks GAN-enhanced Deep Ensemble Neural Network (DENNW) framework designed specifically for cross-domain intrusion detection in cloud and IoT environments. This method significantly improves intrusion detection across various datasets by combining a multi-source learning architecture with GAN-based oversampling to address class imbalance. The Bot-IoT and CSE-CIC-IDS-2018 datasets are used in this research, containing both real and synthetic network traffic, covering a broad range of IoT and cloud-related incidents. The proposed GAN-based DENNW framework outperforms existing cloud-based approaches that use similar measures, providing comprehensive class-wise metric evaluation with 97.22% overall accuracy, surpassing many current studies. Although the DENNW framework achieves 93% accuracy with detailed class-wise analysis, the suggested approach enhances operational efficiency in the IoT sector. The results highlight that the proposed framework for protecting emerging IoT–cloud systems is adaptable and practical.</div></div>","PeriodicalId":29968,"journal":{"name":"Internet of Things","volume":"34 ","pages":"Article 101773"},"PeriodicalIF":7.6,"publicationDate":"2025-10-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145324438","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}
{"title":"A lightweight UAV secure communication scheme integrating cross-domain group authentication and reputation awareness","authors":"Zigang Chen , Chenfeng Zhu , Hongwei Zhang , Fuhai Zhang , Haihua Zhu","doi":"10.1016/j.iot.2025.101781","DOIUrl":"10.1016/j.iot.2025.101781","url":null,"abstract":"<div><div>With the widespread adoption of unmanned aerial vehicles (UAVs) in multi-domain collaborative tasks, traditional centralized authentication mechanisms face significant challenges in handling cross-domain migration, cooperative communication, and resistance to physical attacks. To address these issues, this paper proposes a lightweight multi-domain UAV authentication protocol based on Physically Unclonable Functions (PUFs) and a pseudo-identity mechanism. The scheme integrates cross-domain mutual authentication and group-based collaborative verification, while supporting dynamic path construction and reputation-aware authentication strategies. By leveraging challenge index mapping and aggregated authentication messages, the protocol enables path-level secure coordination among UAV. Additionally, a dynamic reputation mechanism is introduced to regulate authentication privileges of abnormal devices. In terms of security, the protocol is formally verified under the Dolev-Yao model using the ProVerif tool to ensure identity authentication and key confidentiality. For performance evaluation, we measure the average execution time of cryptographic operations using standard cryptographic libraries on a PC platform, and estimate the corresponding computational, communication, and storage overhead. Experimental results demonstrate that the proposed scheme achieves a strong balance between security and efficiency, making it particularly suitable for resource-constrained and highly dynamic UAV collaboration scenarios across multiple domains.</div></div>","PeriodicalId":29968,"journal":{"name":"Internet of Things","volume":"34 ","pages":"Article 101781"},"PeriodicalIF":7.6,"publicationDate":"2025-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145266677","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}
Enver Hamiti , Bujar Krasniqi , Luis M. Correia , Mimoza Ibrani , Bernardo Galego
{"title":"Experimental evaluation of in-vehicle RF-EMF induced by IoT devices of smart parking systems and other wireless technologies","authors":"Enver Hamiti , Bujar Krasniqi , Luis M. Correia , Mimoza Ibrani , Bernardo Galego","doi":"10.1016/j.iot.2025.101789","DOIUrl":"10.1016/j.iot.2025.101789","url":null,"abstract":"<div><div>This study empirically investigates the impact of Internet of Things (IoT) devices in urban settings on human exposure to radiofrequency electromagnetic fields (RF-EMF). It addresses a notable research gap by presenting a systematic evaluation of RF-EMF exposure from a specific IoT implementation. The novelty of this work lies in its dedicated focus on LoRaWAN and its provision of a comparative analysis with other wireless technologies. Measurements were performed in a vehicle with a LoRaWAN device at 868 MHz, moving in several urban environments (university and street parking sites), for three cases: inside and outside a regular (petrol engine) vehicle, and inside an electric (engine). Power density was measured with a broadband probe, covering all current bands of wireless technologies (from 0.79 GHz to 5.85 GHz). The main goal was to determine the potential change in EMF level during the stay of a person in a car that uses IoT devices for monitoring the availability of parking lots and other possible services. As expected, values were very low compared with ICNIRP’s Guidelines (the strictest reference value in the whole measured band is 4 W/m<sup>2</sup>). In the LoRaWAN band, the highest temporal variation of the average power density fluctuated between 5.16 µW/m<sup>2</sup> and 35.37 µW/m<sup>2</sup>, this system contributing only with around 0.02 % of the total power. The measured values fit well a Log-Normal Distribution. The observed low exposure levels and compliance with ICNIRP’s Guidelines (the maximum value is 14.63 dB below ICNIRP’s threshold) support the safe deployment of LoRaWAN based parking systems.</div></div>","PeriodicalId":29968,"journal":{"name":"Internet of Things","volume":"34 ","pages":"Article 101789"},"PeriodicalIF":7.6,"publicationDate":"2025-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145266681","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}
{"title":"FGLIoT: IoT device identification via federated graph learning and spatio-temporal feature fusion","authors":"Xuhui Wang, Guanglu Sun, Xin Liu","doi":"10.1016/j.iot.2025.101785","DOIUrl":"10.1016/j.iot.2025.101785","url":null,"abstract":"<div><div>The device silo problem poses a significant challenge to the management and security of the Internet of Things (IoT). The key to solving this issue is to accurately identify IoT devices connected to the network while protecting data privacy. However, existing solutions overlook inter-packet semantic correlations, a fact which renders them unable to fully explore the potential behavior patterns in device communication traffic. Therefore, we propose FGLIoT, a federated graph learning-based method for IoT device identification. FGLIoT first represents the communication traffic data generated by IoT devices as packet sequence graphs, preserving the semantic information of packets. It then employs a graph learning module to capture inter-packet semantic correlations and learn representations of device communication behaviors. Subsequently, the representations are processed by spatial and temporal feature extractors to capture their spatial correlations and temporal dependences, respectively. Finally, residual connections are used to fuse the behavior representations with their spatial and temporal features, generating behavioral fingerprints for IoT device identification. Experimental results on three public IoT device datasets demonstrate the effectiveness of FGLIoT in solving the device silo problem.</div></div>","PeriodicalId":29968,"journal":{"name":"Internet of Things","volume":"34 ","pages":"Article 101785"},"PeriodicalIF":7.6,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145220826","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}
{"title":"A systematic literature review on AI in IoT systems: Tasks, applications, and deployment","authors":"Umair Khadam , Paul Davidsson , Romina Spalazzese","doi":"10.1016/j.iot.2025.101779","DOIUrl":"10.1016/j.iot.2025.101779","url":null,"abstract":"<div><div>The integration of Artificial Intelligence (AI) into Internet of Things (IoT) systems has garnered considerable attention for its ability to enhance efficiency, functionality, and decision making. To drive further research and practical applications, it is essential to gain a deeper understanding of the different roles of AI in IoT systems. In this systematic literature review, we analyze 103 articles describing Artificial Intelligence of Things (AIoT) systems found in three databases, i.e. Scopus, IEEE Xplore, and Web of Science. For each article, we examined the tasks for which AI was used, the input and output data, the application domain, the maturity level of the system, the AI methods used, and where the AI components were deployed. As a result, we identified six general tasks of AI in IoT systems, and thirteen subtasks, the most frequent being prediction, object and event recognition, and operational decision-making. Moreover, we conclude that most AI components in IoT systems process numeric data as input and that healthcare is the most common application domain followed by farming and transportation. Our analysis further revealed that most AIoT systems are in early development stages not validated in real environments. We also identified that Convolutional Neural Networks is the most frequently employed AI method, with supervised learning being the dominant approach. Additionally, we found that both AI deployment, either in the cloud or at the edge, are frequent, but that hybrid deployment is not that common. Finally, we identified key gaps in current AIoT research and based on this, we suggest directions for future research.</div></div>","PeriodicalId":29968,"journal":{"name":"Internet of Things","volume":"34 ","pages":"Article 101779"},"PeriodicalIF":7.6,"publicationDate":"2025-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145220821","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}
Azad Shokrollahi , Fredrik Karlsson , Reza Malekian , Jan A. Persson , Arezoo Sarkheyli-Hägele
{"title":"Non-invasive occupancy estimation and space utilization in smart buildings: Leveraging machine learning with PIR sensors and booking data","authors":"Azad Shokrollahi , Fredrik Karlsson , Reza Malekian , Jan A. Persson , Arezoo Sarkheyli-Hägele","doi":"10.1016/j.iot.2025.101777","DOIUrl":"10.1016/j.iot.2025.101777","url":null,"abstract":"<div><div>Occupancy estimation in smart buildings is essential for optimizing resource usage and enhancing operational efficiency. Existing estimation methods predominantly rely on cameras or advanced sensor fusion techniques, which, while accurate, are often expensive, invasive, and raise privacy concerns. Additionally, these approaches frequently require extra hardware, increasing installation complexity and operational costs. A significant gap in the literature lies in the limited use of existing smart building infrastructure, such as detection systems and booking data, for people counting. This study addresses these limitations by exclusively utilizing two binary PIR sensors (in-door and in-room) and booking data. Since PIR sensors and booking systems are already integrated into most smart building infrastructures, leveraging these existing resources helps reduce costs and simplifies implementation. The primary goal is to estimate the number of people between each in-door sensor trigger using machine learning models by incorporating people counting levels and time thresholds. Among the evaluated machine learning algorithms, the Extra Trees Classifier delivered strong performance, achieving 68.5% accuracy when the estimated occupancy differed from the actual count by at most one person, and 81.56% with a tolerance of two. These results are based on periods when the room was occupied. When both occupied and unoccupied periods were included, the accuracy was 96.10% for ±1 tolerance. Moreover, incorporating booking data enhanced people counting accuracy by 4%. The study also explores the method’s ability to identify underutilization and overutilization by comparing estimated occupancy with booking records and seating capacity, thereby supporting enhanced space management in smart buildings.</div></div>","PeriodicalId":29968,"journal":{"name":"Internet of Things","volume":"34 ","pages":"Article 101777"},"PeriodicalIF":7.6,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145220718","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}
Andrea De Simone, Giovanna Turvani, Fabrizio Riente
{"title":"Incremental firmware update over-the-air for low-power IoT devices over LoRaWAN","authors":"Andrea De Simone, Giovanna Turvani, Fabrizio Riente","doi":"10.1016/j.iot.2025.101772","DOIUrl":"10.1016/j.iot.2025.101772","url":null,"abstract":"<div><div>Remote firmware updates in Internet of Things (IoT) devices remain a major challenge due to the constraints of many IoT communication protocols. In particular, transmitting full firmware images over low-bandwidth links such as Long Range Wide Area Network (LoRaWAN) is often impractical. Existing techniques, such as firmware partitioning, can alleviate the problem but are often insufficient, especially for battery-powered devices where time and energy are critical constraints. Consequently, physical maintenance is still frequently required, which is costly and impractical in large-scale deployments. In this work, we introduce <em>bpatch</em>, a lightweight method for generating highly compact delta patches that enable on-device firmware reconstruction. The algorithm is explicitly designed for low-power devices, minimizing memory requirements and computational overhead during the update process. We evaluate <em>bpatch</em> on 173 firmware images across three architectures. Results show that it reduces update payloads by up to 39,000×for near-identical updates and by 9–18×for typical minor revisions, eliminating the need to transmit full firmware images. Experimental results further demonstrate significant time and energy savings, with performance comparable to more complex alternatives. <em>bpatch</em> is released as open-source and, although demonstrated on LoRaWAN, the approach is flexible and can be adapted to other IoT communication technologies.</div></div>","PeriodicalId":29968,"journal":{"name":"Internet of Things","volume":"34 ","pages":"Article 101772"},"PeriodicalIF":7.6,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145220827","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}
{"title":"A Secured Swarm Intelligence-based Path Selection framework for Malicious Low-power and Lossy Networks under RPL protocol","authors":"Hanin Almutairi , Salem AlJanah , Ning Zhang","doi":"10.1016/j.iot.2025.101776","DOIUrl":"10.1016/j.iot.2025.101776","url":null,"abstract":"<div><div>Low-power and Lossy Networks (LLNs) face persistent challenges, including dynamic topologies, unreliable links, limited energy, and constrained computational resources. These issues are exacerbated under malicious conditions such as Packet Dropping Attacks (PDAs), where conventional routing and security mechanisms fall short due to their high computational overhead. To address these challenges, this paper proposes the Secured Swarm Intelligence-based Path Selection (S-SIPaS) framework, designed to enhance reliability and security in Malicious LLNs (MLLNs). S-SIPaS builds on our previous SIPaS framework by integrating a lightweight trust model and a novel Secured Ant Colony Objective Function (S-ACOF) into the RPL protocol. S-ACOF applies Ant Colony Optimisation (ACO) principles to compute globally optimal, trustworthy paths while reducing energy consumption and control overhead. A key feature of S-SIPaS is its three-phase trust model: monitoring, trust measurement, and trust determination, which detects and isolates malicious nodes based on packet-forwarding behaviour, without relying on cryptographic techniques.</div><div>The framework combines multiple routing metrics, including physical distance, energy level, link quality, and trust score, enabling adaptive and efficient path selection in dynamic LLNs. Simulation results show that S-SIPaS improves Packet Delivery Ratio (PDR) by up to 51% over existing methods, especially in high-density and high-attack scenarios.</div><div>Despite strong performance, the framework has limitations: (i) it requires C1-class nodes (e.g., Z1); (ii) evaluation is limited to simulations; and (iii) it currently addresses only PDA threats and static topologies. Overall, S-SIPaS offers an effective, scalable, and secure routing solution for enhancing MLLNs and IoT systems.</div></div>","PeriodicalId":29968,"journal":{"name":"Internet of Things","volume":"34 ","pages":"Article 101776"},"PeriodicalIF":7.6,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145220824","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}