Semantic communication improves task effectiveness by transmitting task-relevant information. However, most existing schemes remain organized as task-specific, end-to-end pipelines, which are difficult to reuse across models, applications, and deployment environments. Against this background, we propose the Semantic Internet of Everything (SIoE), a composable service architecture that represents heterogeneous communication and artificial intelligence (AI) functions as capability-profiled services and coordinates them according to application objectives. SIoE comprises three planes: a task and service plane, an agentic orchestration plane, and a semantic capability plane. In this framework, task requirements are captured via a semantic service-level agreement (SLA), while an agentic planner discovers and composes candidate capabilities under deterministic compatibility, resource, privacy, and policy validation. Feedback from the communication, semantic, and task levels enables continuous adaptation and replanning. A lightweight vehicle-to-everything case study illustrates profile-grounded capability planning under explicit service constraints. The results demonstrate the feasibility of decoupling service objectives from fixed communication implementations and also highlight key open challenges, including semantic SLA design, capability interoperability, scalable planning, and trustworthy execution.
The Internet of Things has evolved from connecting devices to integrating human behavior and machine intelligence into system design. Yet mainstream frameworks—from Cyber-Physical Systems (CPS) to Cyber-Physical-Social Systems (CPSS)—still lack an explicit dimension for cognition. The Cyber-Physical-Social-Thinking (CPST) space theory, introduced in 2015, filled this gap by adding “Thinking” as a first-class dimension encompassing both human cognition and artificial intelligence. This article provides a practitioner-oriented retrospective on the CPST framework’s tenth anniversary. We explain the four-dimensional model, trace its evolution through General Cyberspace (2018) and Cyberology (2022), and compare it against six related paradigms. Through three deployed case studies—Hangzhou’s Urban Brain, AGI-enabled IoX ecosystems, and cognitive sovereignty in national governance—we demonstrate how CPST guides system architects toward designs that treat cognitive processes as first-class engineering concerns. We conclude with a three-horizon practitioner’s roadmap for adopting CPST principles in IoT system design over the next decade.
The emergence of 6G networks and the growing adoption of the medical Internet of Things (IoT) are transforming healthcare into increasingly real-time, distributed, and autonomous e-health systems. In such settings, secure zero-touch operation is becoming a fundamental requirement, as healthcare environments must protect not only confidential medical data, but also safety-critical control messages, continuous monitoring streams, and the trustworthiness of clinical decisions. Meanwhile, the rapid progress of quantum computing is beginning to challenge the long-standing security assumptions of public-key cryptography, creating an urgent need to rethink security for next-generation connected healthcare. In this article, we introduce three representative quantum cryptographic approaches, namely Quantum Key Distribution (QKD), Quantum Secret Sharing (QSS), and Quantum Secure Direct Communication (QSDC), and discuss their potential roles in strengthening security for medical IoT. QKD offers a promising direction for trusted key establishment over high-value medical communication links, QSS supports collaborative authorization and distributed trust for highly sensitive medical data and control functions, and QSDC is well suited to the direct protection of time-critical medical information and control messages. The article further discusses how these approaches offer a path toward unconditional security under ideal conditions and examines their potential roles in future healthcare applications. Collectively, these three approaches suggest a broader security framework in which key establishment, authorization, and secure communication are more tightly integrated.
Emotional wellbeing plays a critical role in learner engagement, cognitive performance, and academic success in smart educational environments. However, existing intelligent learning systems mainly focus on learning outcomes while providing limited integration of physiological wellbeing information and real-time adaptive healthcare support. To address these limitations, this paper proposes a Context-Aware Digital Service Loop Orchestration Framework for emotional wellbeing assessment in IoT-enabled smart English education healthcare systems. The framework adopts a multi-layer architecture comprising IoT sensing devices, edge computing, cloud intelligence, and application-layer intervention services to enable continuous wellbeing monitoring and adaptive support. The proposed system integrates educational affective information and physiological health signals within a unified ecosystem. Educational video data from the DAiSEE dataset and physiological signals from the WESAD dataset are processed through distributed sensing and data management modules. At the edge layer, MediaPipe Face Mesh is utilized for facial landmark extraction, while NeuroKit2 performs preprocessing, noise removal, artifact correction, and signal standardization for physiological data. Deep emotional representations are extracted using ConvNeXt, and wellbeing indicators are derived from Heart Rate Variability (HRV) and Electrodermal Activity (EDA) features. A Context-Aware Joint Multimodal Transformer (CA-JMT) with a Context-Aware Adaptive Cross Attention (CAACA) mechanism models dynamic interactions between educational behavior and physiological responses. Furthermore, an Enhanced Narwhale Optimization Algorithm (ENWOA) optimizes model parameters through adaptive exploration–exploitation balancing and multi-objective fitness evaluation. The assessed wellbeing state is integrated into a Digital Service Loop for continuous sensing, assessment, personalized intervention orchestration, outcome verification, and feedback-driven adaptation. Experimental results achieve 98.71% accuracy, 98.84% precision, 98.62% recall, 98.73% F1-score, 98.89% specificity, 0.0852 loss, and 0.982 AUC, demonstrating the framework’s effectiveness, scalability, and suitability for healthcare-oriented smart education environments.
The increasing frequency of Distributed Denial of Service (DDoS) attacks poses a major threat to large-scale Internet of Things (IoT) systems, particularly those operating over emerging 6G-enabled edge infrastructure. These systems support critical applications such as e-health, which are driven by Internet of Medical Things (IoMT) networks. IoMT networks are dynamic and resource-constrained. This makes it difficult to deploy computationally intensive security solutions directly on medical edge devices. In such environments, centralized solutions are often impractical. Federated learning offers a suitable approach for distributed IoT security. It provides collaborative model training while preserving data privacy. However, existing federated learning-based intrusion detection systems suffer from high communication overhead and inefficient client participation, which limits their scalability and responsiveness. To address these challenges, this paper proposes FedECHO, a communication-efficient federated ensemble framework for adaptive DDoS detection and mitigation in IoMT networks, with applicability to 6G-enabled e-health environments. The framework integrates a CNN–LSTM ensemble model with entropy-based update filtering and Lagrange multiplier-based client selection to reduce communication cost. FedECHO also supports autonomous operation. It triggers response actions based on detection outcomes, including rate limiting, traffic redirection, and honeypot-based isolation. These actions use a game-theoretic strategy and create a closed-loop system where detection and response run without operator input. Experimental results on the HL-IoT, CICIoT-2023 and CICIoMT 2024 datasets show that FedECHO achieves over 94% detection accuracy while reducing communication overhead by up to 1.23 times compared to existing federated learning approaches.
Integrated sensing and communication (ISAC) can enable sixth-generation (6G) unmanned aerial vehicle-assisted Internet of Things (UAV-IoT) networks to provide reliable Localization-as-a-Service (LaaS), but activating all aerial/terrestrial anchors and beams increases pilot overhead, energy use, and beam-training delay. This article proposes artificial intelligence (AI)-assisted ISAC resource selection for LaaS (AIRS-LaaS), an edge-intelligent framework that ranks candidate anchor–beam pairs using line-of-sight (LoS) likelihood, signal-to-interference-plus-noise ratio (SINR), sensing confidence, geometry, mobility risk, and resource cost. A lightweight selector then activates only a compact subset before localization. Simulations compare AIRS-LaaS with all-anchor, Fisher information matrix/Cramér–Rao lower bound (FIM/CRLB)-greedy, strongest-SINR, nearest-anchor, and random schemes under LoS/non-line-of-sight (NLoS) conditions and UAV mobility. Results show a balanced localization–communication–overhead tradeoff, while the discussion highlights standard-driven key performance indicators (KPIs), ISAC reporting, localization confidence, fallback operation, AI model management, and privacy-aware data exchange.
While facilitating communication services, cellular mobile networks also provide real-time and large-scale observations of individual mobility, offering significant potential far beyond their traditional connectivity role. This paper proposes a mobility-aware epidemiological framework that integrates the Susceptible-Exposed-Infectious-Recovered (SEIR) model with mobility data derived from cellular mobile networks to improve predictive accuracy and applicability. By incorporating fine-grained, cell-level mobility data, our approach captures localized movement patterns and their impact on epidemic dynamics in urban environments. Simulation results show that areas with high mobility and density reach infection peaks 66–70% earlier and over 2.5 times higher than predictions obtained using a standard SEIR model, while low-mobility regions exhibit minimal deviation from baseline behavior. These findings illustrate that network-level mobility indicators are effective tools for identifying localized epidemic risks, supporting targeted and real-time public health interventions.
Digital twin (DT) is emerging as a key enabler for cyber-physical integration in the Industrial Internet of Things (IIoT). By deploying virtual twins (VTs) at edge nodes, DT can provide low-latency and high-fidelity services for mobile physical twins (PTs), such as inspection robots, automated guided vehicles, and intelligent production equipment. However, sustainable DT deployment in industrial edge networks is challenged by two coupled factors: PT mobility and edge node lifetime degradation. The former may trigger frequent VT placement or data rerouting for PT-VT synchronization, while the latter makes edge service availability maintenance-dependent under harsh industrial environments. This paper investigates adaptive edge maintenance for fidelity-aware DT deployment in industrial IoT. We discuss how edge lifetime, maintenance status, PT mobility, and fidelity requirements jointly affect VT hosting, updating, and synchronization. Several key design challenges are highlighted, including lifetime evaluation, maintenance scheduling, VT placement, data rerouting, sensing-data sampling, and resource orchestration. As a potential solution, we present a two-timescale collaborative embodied learning framework to coordinate lifetime-aware deployment with adaptive maintenance and service decisions. Simulation results show that the proposed design reduces average energy consumption by up to 18% while maintaining stable VT accuracy above 85%.
Electric vehicle (EV) charging infrastructures are rapidly evolving into large-scale cyber-physical ecosystems that integrate cloud services, grid coordination platforms, and user-facing applications. While this connectivity improves operational efficiency, it also expands the cybersecurity attack surface of charging networks. Traditional intrusion detection systems often struggle to balance scalability, detection accuracy, and resource constraints in such distributed environments. This article presents the Clone-Swarm Sentinel Model (CSSM), an edge-native intrusion detection architecture designed for EV charging infrastructures. CSSM separates continuous anomaly detection from contextual validation through event-driven agent cloning. When suspicious behavior is detected, a temporary Clone Agent inherits the runtime state of the monitoring process and generates a counterfactual behavioral baseline to confirm whether the deviation represents a persistent anomaly. This design reduces false positives while maintaining bounded computational overhead at edge devices. The architecture combines lightweight edge detection with collaborative learning through federated model adaptation, enabling the system to adapt to evolving charging behaviors while preserving data locality. Evaluation using real-world EV charging datasets demonstrates improved detection accuracy, resilience under network degradation, and scalability across large deployments involving thousands of charging stations. Beyond EV charging systems, CSSM illustrates a broader architectural approach for building scalable and resilient security frameworks for distributed IoT cyber-physical infrastructures.
Industrial Internet of Things deployments increasingly rely on swarms of autonomous UAVs for real-time inspection, logistics monitoring, and predictive maintenance. These UAVs must communicate while moving rapidly in 3D space, where industrial machinery generates strong electromagnetic interference and localized radio dead zones, making reliable multihop communication difficult. Geographic forwarding selects the next hop closest to the destination on a map but ignores actual radio conditions, while most learning-based protocols update too slowly to track link changes that occur within seconds. This article presents the Fast Adaptive Interference-aware Routing (FAIR) scheme, a fully distributed protocol in which each UAV scores candidate next hops by combining link reliability estimated from radio signal quality, predicted link lifetime, and geographic progress toward the destination. This enables proactive avoidance of interfered links without any extra probing traffic. FAIR further adjusts how aggressively it updates routing preferences according to current link stability. Simulation results show that FAIR consistently achieves higher packet delivery ratio, lower end-to-end latency, and lower energy consumption than two benchmark protocols across a wide range of conditions, demonstrating its potential as a practical communication backbone for collaborative UAV systems in industrial environments.
Telemedicine is becoming a key service domain for future 6G systems, spanning remote consultation, continuous patient monitoring, emergency response, and distributed diagnosis. As these workflows become more mobile, data-intensive, and software-driven, manual service management is increasingly difficult to sustain. Zero-touch management offers a promising path to autonomous operation, but healthcare environments require more than generic automation because adaptation decisions must also reflect clinical priority, trustworthiness, and security risk. This article presents a trust-aware secure zero-touch Internet of Things (IoT) architecture for 6G-enabled telemedicine systems, where Internet of Medical Things (IoMT) devices form the medical sensing and device layer, and telemedicine is treated as a clinically demanding service subset of broader 6G-enabled e-health infrastructures. The proposed architecture combines multi-source observation, continuous trust reassessment, policy-bounded orchestration, and programmable enforcement across device, edge, network, and cloud domains. Structurally, it implements a MAPE-K-aligned closed-loop control cycle integrating a Trust Engine, a Policy Engine with tiered conflict resolution, and a Zero-Touch Orchestrator informed by lightweight AI/ML components. It supports trusted admission and onboarding, anomaly-aware adaptation, selective service degradation, trusted fallback, and auditable self-healing. A remote emergency ambulance scenario demonstrates the step-by-step closed-loop operation, showing how critical medical flows can be preserved while less essential services are adaptively downgraded under network degradation or cyber risk. The article concludes with comparative insights and key open challenges for deploying trustworthy autonomous telemedicine at scale.
Parkinson’s disease is characterized by a variety of motor symptoms, with rest tremor and dyskinesia among the most common and disruptive, significantly affecting patients’ quality of life. However, their clinical evaluation remains subjective and limited to in-clinic assessments and patient self-reports. Intelligent Internet of Medical Things (IIoMT) solutions based on wearable sensors can provide continuous and objective monitoring of these symptoms and support the extraction of clinically interpretable digital biomarkers. In this work, we present a deep learning-based IIoMT framework for wrist-worn accelerometer data that jointly supports rest tremor detection, rest tremor amplitude classification and dyskinesia detection. The proposed framework leverages symptom-specific deep learning pipelines based on spatiotemporal representation learning to capture the complex dynamics of Parkinson’s motor symptoms, while supporting scalable IIoMT deployment. We evaluate the proposed pipelines using data from two widely used datasets, namely the Michael J. Fox Foundation Levodopa Response Study dataset and a subset of the Verily Study Watch dataset. The results demonstrate strong performance of the rest tremor pipeline, including external validation, and show promising dyskinesia detection performance on MJFF-LRS. Overall, the proposed framework highlights the potential of wearable IIoMT systems for objective assessment of Parkinson’s motor symptoms and medication-related fluctuations.
Wireless capsule endoscopy (WCE) location estimation is crucial for the growing Internet of Medical Things (IoMT). It supports continuous monitoring and applications like targeted biopsy, location-based drug delivery, and minimally invasive surgery. While using received-signal-strength and time-of-arrival is common for localization, this work focuses on time-difference-of-arrival (TDoA) parameters within an IoMT framework. The transmitted signal is an ultra-wideband (UWB) pulse emitted by a wireless capsule antenna at three positions within the abdominal cavity. A network of nine on-body receivers, including a reference receiver, captures the signals. This represents a body-centric IoMT sensing layer. Propagation is modeled using a human voxel model in CST Studio Suite. Four TDoA extraction methods are applied to the received signals, and the in-voxel TDoA is compared to the actual TDoA to calculate the TDoA error. The strongest correlation peak yields the highest TDoA error, resulting in average localization errors of 46 mm at two WCE-positions and 62 mm at one WCE-position. Even though the WCE localization error in the voxel model remains higher than in open space, an additional experiment with more receivers reduces the average error. Moreover, the same number of receivers shows different localization accuracies across abdominal zones due to varying tissue composition. These findings encourage IoMT-driven optimization, including adjusting receiver density by zone and developing an IoMT-based intelligent receiver-selection strategy to improve localization accuracy and receiver utilization in real-world WCE IoMT systems.
Dyslexia screening remains a challenging clinical problem due to its multimodal cognitive manifestations, reliance on specialist assessment, and limited scalability in conventional healthcare settings. While the Intelligent Internet of Medical Things (IIoMT) enables data-driven dyslexic pattern assessment through distributed sensing, existing solutions often lack adaptive intelligence, contextual continuity, and integrated security. This article presents a secure multi-agentic hybrid IIoMT framework for multimodal dyslexia screening, combining agent-to-agent (A2A) intelligence, Model Context Protocol (MCP)-based orchestration, and longitudinal cognitive reasoning. The proposed framework integrates four complementary screening tasks—face drawing, clock drawing, handwriting analysis, and reading assessment with gaze tracking—captured via heterogeneous IIoMT endpoints. Specialized AI agents analyze each modality and collaborate through A2A coordination to enable multimodal feature fusion and unified dyslexia risk assessment. The use of persistent contextual intelligence is a key innovation, enabling the retention of clinically relevant interaction context across sessions for personalized and progressive screening. Security and privacy are embedded as system properties, addressing data protection, multi-agentic AI model trustworthiness, and secure, ethical, and authenticated inter-agent collaboration. Experimental evaluation on dyslexia screening demonstrates achieving 94.5% accuracy in face drawing analysis, 82.7% in clock drawing, 96.7% in handwriting assessment, and 83.3% accuracy in reading fluency detection (n = 120). The proposed framework advances dyslexia screening toward scalable, intelligent, and longitudinal digital assessment.
The vision of 6G networks is to realize a fully connected and intelligent world by integrating large-scale distributed intelligence with Federated Learning (FL) and the Internet of Things (IoT). However, conventional FL faces significant challenges in resource-constrained IoT environments, including slow convergence and high communication overhead during training operations. These challenges, combined with the stringent service requirements of IoT subsystems, highlight the need for more efficient FL mechanisms. Low Earth Orbit (LEO) satellites, with their global coverage and onboard edge computing capabilities, provide a promising platform for supporting large-scale distributed intelligence solutions. In this work, we propose a multi-layered Non-Terrestrial Network (NTN)-based FL framework that enables intelligent IoT applications. The framework leverages LEO satellite-assisted model search, transfer, and similarity-aware FL initialization to accelerate learning and reduce communication costs. We further analyze the trade-off between the overhead of model search and transfer over the LEO satellite network and the resulting improvements in FL convergence performance. Performance evaluations reveal that even a limited satellite-assisted search can cut convergence latency by nearly half compared to conventional FL, demonstrating the importance of LEO satellite-assisted model search, transfer, and similarity-aware FL initialization. These results highlight the potential of satellite-enabled FL as a key enabler of scalable, intelligent, and globally connected 6G IoT ecosystems.
Smart factories increasingly rely on embodied AI agents such as autonomous mobile robots and automated guided vehicles to perform real-time manufacturing, logistics, and inspection tasks. These systems require accurate and low-latency localization to enable safe navigation and coordinated multi-agent operations within dynamic industrial environments. However, conventional localization approaches often depend on centralized processing or raw multi-modal data transmission, which leads to excessive communication overhead, increased latency, and limited scalability in Industrial Internet of Things (IIoT) networks. This article presents an edge-enabled collaborative localization framework that integrates multi-agent sensing, task-oriented semantic communication, and edge-based collaborative fusion for smart factory environments. Instead of transmitting raw sensor streams, the proposed framework extracts task-relevant semantic features and transmits compact feature vectors to an edge server, significantly reducing bandwidth consumption while maintaining accurate localization. The edge layer performs collaborative pose fusion and global map refinement across multiple robots, enabling scalable and real-time localization under IIoT communication constraints. A case study-based evaluation demonstrates that the proposed approach improves localization accuracy, reduces communication latency, and enhances scalability compared with conventional cloud-centric and non-collaborative localization methods. The results highlight the potential of combining edge intelligence and semantic communication to support reliable multi-agent coordination in next-generation smart factory systems.
Internet of Medical Things (IoMT) devices operate under strict resource constraints while handling highly sensitive health data, making security and privacy critical concerns. Federated learning (FL) further complicates this landscape, as model updates exchanged during training may unintentionally expose private medical information. Emerging quantum computing capabilities threaten the long-term viability of conventional lightweight cryptographic mechanisms, motivating the integration of Post-Quantum Cryptography (PQC) into IoMT systems. This article discusses key enabling technologies for quantum-resilient IoMT, including post-quantum key establishment, lightweight encryption, and edge-native orchestration. We propose a scalable Kubernetes-based framework that integrates PQC into FL-enabled IoMT environments and validate it on a Raspberry Pi testbed. Results demonstrate that distributed cryptographic processing significantly reduces latency compared to sequential designs while maintaining feasible resource overhead. The primary contribution of this work lies in the design and validation of a secure orchestration and communication framework for FL-enabled IoMT systems. We conclude by outlining future directions toward energy-aware architectures, intelligent security optimization, and resilient next-generation Intelligent Internet of Medical Things (IIoMT) ecosystems.
Sixth-generation (6G) networks are expected to integrate intelligence as a native capability, enabling advanced verticals such as digital health (eHealth) supported by large-scale Internet of Medical Things (IoMT) deployments. In this context, Federated Learning (FL) is emerging as a promising paradigm for collaborative model training, allowing distributed medical devices to learn from data while preserving privacy. Nevertheless, conventional approaches such as Federated Averaging (FedAvg) often suffer from unstable convergence and performance degradation in non-independent and non-identically distributed (non-IID) environments, especially under partial and resource-constrained client participation. To address these challenges, we propose a Digital Twin (DT)-enabled FL framework in which the server evolves from a passive aggregator into an active orchestrator of the learning process. The DT consists of a virtual representation of participating devices and their learning dynamics, thereby enabling an informed selection of clients and coordinated updates based on both resource conditions and predictive uncertainty. Building on this orchestration layer, we introduce Twin2Twin (T2T), a hierarchical aggregation strategy that groups clients into similarity-based cohorts and combines their updates in a structured manner, improving training stability in dynamic environments. Specifically, the proposed framework integrates: (i) DT-driven client orchestration that accounts for both device heterogeneity and data informativeness, and (ii) a similarity-aware aggregation mechanism that reduces update variability across training rounds. Experimental results demonstrate that the proposed strategy achieves faster and more stable convergence compared to traditional FL, highlighting the potential of DT-driven orchestration to support reliable and scalable learning in IoMT systems deployed in 6G networks.
