Artificial intelligence (AI) plays a critical role in improving safety across modern transportation systems, enabling collision prediction, cooperative perception, autonomous navigation, and real-time traffic control. However, many AI-enabled safety functions rely on closed-box models that create a significant responsibility gap in which automated decisions cannot be explained or justified. This gap is especially concerning in the safety applications of autonomous vehicles (AVs) and smart public transit systems, where incorrect or unjustified AI decisions may result in severe accidents or loss of life. Traditional interpretability approaches provide limited post hoc explanations that are insufficient for transportation safety certification or incident analysis. This article proposes a causal AI-driven framework for responsible, auditable, and safety-aligned intelligence in transportation systems. Using structural causal models (SCMs), counterfactual reasoning, and causal intervention analysis, the proposed framework provides root-cause attribution, fairness assurance, and audit-ready explanations across land, maritime, and air transportation systems. Two representative use cases are presented: the causal evaluation of vehicle-to-everything (V2X)-assisted collision avoidance and the causal justification of autonomous driving policies. Experimental validation confirms that the proposed causal framework reduces collision rates by 80.6% and maintains safety even under severe communication latencies and sensor noise. Furthermore, compared to existing probabilistic causal approaches, our proposed method achieves an optimal balance between safety and travel efficiency while operating at submillisecond latency. These results demonstrate that causally grounded intelligence is essential for public safety, system reliability, and regulatory compliance in next-generation transportation systems.
Low-Earth-orbit (LEO) satellite constellations are poised to become a cornerstone of the 6G Internet of Things (IoT), providing truly global coverage and ubiquitous connectivity. This article presents a holistic 2D system architecture for 6G LEO satellite IoT that incorporates composition and functional perspectives to facilitate the seamless integration of LEO satellites and terrestrial networks. Building upon this architecture, we evaluate three pivotal enabling technologies that target uplink, downlink, and inter-satellite links (ISLs). Specifically, we analyze massive grant-free random access (GF-RA) for efficient uplink connectivity, investigate deep-learning (DL)-based multibeam precoding for robust downlink transmission, and examine distributed cooperative routing for resilient ISL data delivery. Furthermore, we present an on-orbit verification platform that validates the real-world feasibility and performance of the proposed solutions. Finally, we outline key open challenges and future research directions to guide the realization of future LEO satellite IoT.
Direct handset-to-satellite (DHTS) systems represent a key enabler of nonterrestrial networks (NTNs), allowing mass-market handheld devices to directly access low-Earth-orbit (LEO) satellites. This article provides a tutorial perspective on the architecture-aware roles and system-level challenges of signal detection in DHTS systems. We show that regenerative payload configuration, multisatellite cooperation, and control plane (CP)/user plane (UP) separation fundamentally reshape where and how detection is performed. Meanwhile, severe Doppler shifts, asynchronous multilink superposition and onboard processing constraints invalidate many terrestrial assumptions, motivating waveform-aware, structure-exploiting, and cooperative detection strategies. These challenges also drive the adoption of soft-output detection schemes to suppress error propagation in iterative decoding. Such emerging detection requirements motivate tighter integration with access control mechanisms and the development of onboard intelligence. Overall, the evolving role of detection highlights the need for scalable, reliable, and resource-efficient solutions in future DHTS deployments.
High-altitude platform stations (HAPSs) as international mobile telecommunications (IMT) base stations (HIBSs) have the advantages of wide coverage, flexible deployment, low latency, and low construction cost and can be used as a supplement and extension of the existing network to promote the realization of full coverage of the space–air–ground integrated network. However, in the context of limited low-frequency resources, promoting compatibility and coexistence between HIBSs and existing ground IMT networks to enhance system capacity and quality of service is challenging, and some issues should be addressed. Thus, this article focuses on studying and analyzing the compatibility of HIBSs and ground IMT systems in the 694–960-MHz ultrahigh frequency band and then showing coexistence conditions based on simulation results.
Driven by the rapid development of low-Earth orbit (LEO) satellite networks, machine learning (ML) model training and data processing face significant challenges owing to constrained communication bandwidth (BW) and privacy concerns. Federated learning (FL) has emerged as a promising solution for in situ data processing; however, centralized FL techniques are vulnerable to single-point failures and malicious attacks. To address these issues, we propose SpaceChain, a blockchain-enhanced FL framework tailored to LEO satellite networks. The SpaceChain framework includes a decentralized validation system for assessing local model updates. The system implements a proof-of-stake (PoS)-inspired consortium consensus mechanism in a permissioned blockchain system, thereby giving precedence to satellites based on cumulative learning contributions for validation, leading to a reduction of the energy expenditure when compared with a proof-of-work (PoW) protocol. FL maintains the privacy of data, with raw data stored on satellites, and the blockchain maintains tamper resistance in coordinating model updates. Satellites are assigned specific roles during each communication round, including ML nodes for local updates, consensus verifiers for validation, and blockchain stake validators for the final model integration. This role-based architecture ensures dynamic node participation and enhances the reliability and efficiency of the system. The simulation results demonstrate that the SpaceChain framework enhances the security, reliability, and performance of ML training in LEO networks. This article outlines the principles, architecture, and potential applications of SpaceChain, providing a foundation for secure and efficient ML training in next-generation satellite communication systems.
Point-to-point optical wireless communication (OWC) systems and derived point-to-multipoint optical Internet of Things (O-IoT) broadcasting networks can serve as a secure and reliable means for information protection, emergency communication, and confidentiality management within localized environments. Meanwhile, they may also act as potential tools for covert eavesdropping, weak signal detection, or latent surveillance. We summarize recent advances and potential progress in physical-layer security (PLS) of OWC. Furthermore, several case studies are presented, covering a broad spectral range from ultraviolet to near-infrared (NIR), to illustrate how system-level OWC architectures can enable or defend against optical attacks, interception, and confidential transmission. These cases involve both hardware- and software-defined implementations, including photon counting, digital signal processing, and protocol design, highlighting practical approaches for optical information security management, covert transmission, eavesdropping, and weak signal detection. The discussed methods provide complementary solutions for the design and deployment of secure LANs in the emerging information security era
Fiber–visible light communication (VLC) technology has emerged as a promising solution for high-speed, low-latency wired–wireless communications by combining the advantages of both fiber and VLC systems. However, existing fiber–VLC architectures rely on optical–electrical (O–E) conversions, which disrupt the continuity of optical transmission and introduce additional latency. In this work, we propose an all-optical fiber–VLC system architecture that eliminates conventional O–E conversions, which leverages degenerate four-wave mixing (dFWM) for seamless optical signal processing. This novel approach enables direct wavelength translation and amplification through dFWM, thereby maintaining the signal entirely in the optical domain. Specifically, the proposed all-optical approach integrates fiber transmission, nonlinear spectral conversion, and VLC emission, forming a seamless optical path from source to receiver. By eliminating the electrical process, the system achieves lower latency, higher energy efficiency, and enhanced physical-layer security. Moreover, key implementation challenges and future integration opportunities are also discussed.
Risk assessment is intuitive for human drivers. Often, a look in the mirror or noticing a minor anomaly in the traffic pattern is enough to become aware that all is not well. Autonomous vehicles (AVs), however, require measurable cues and structured signals to detect risk. Current models for risk assessment in AVs have not been able to capture the early intuitive awareness that humans get from subtle interaction cues. This means that stronger risk assessment methods are required when dealing with real traffic and its unpredictable behaviors. Human drivers achieve this awareness through experience and intuition. This article proposes a dual-view context-aware framework that fuses the strengths of a data-driven supervised approach to emulate experience and an unsupervised approach to replicate human instinct. The dual view cancels out the peaks and valleys of the two models and therefore does not need any extra calibration. Traditional systems use features that look at how soon two vehicles might meet or how hard a driver would need to brake to stay safe. These help with timing and braking for an individual vehicle but do not account for how vehicles influence one another while driving. To fill this gap, we introduce two new context-based features: a risk buffer distance (RBD) ratio, which measures how much safe space is left before braking becomes necessary, and relative interaction (RI), which measures how much one vehicle’s movement affects another. A reasoning layer is added such that for every high-risk prediction, it gives reasons as to why the situation is dangerous and how it can be mitigated. While most existing models respond to noise due to false high-risk alerts, the dual view applies smoothing to its risk predictions to reduce spurious alerts. To sum up, the proposed model has proved to be highly reliable, self-calibrated, and fast in response, making it suitable for real-world deployment.
This article advocates for the transformative exploitation of connected and intelligent vehicular systems (CIVS) in smart cities. It highlights the opportunity presented by underutilized CIVS resources, including processing and communication hardware, roadside installations, mobile spaces, large batteries, and sensors, to improve urban performance, sustainability, and the overall human experience. The article proposes a shift toward optimal resource utilization, enabling on-demand services that move beyond traditional CIVS roles and support adaptive human-centric smart cities. When effectively integrated, these resources can power intelligent, context-aware services that address real urban challenges, such as congestion, energy demand, public safety, and infrastructure monitoring. The article reviews recent advancements in CIVS technologies and proposes a quality of experience (QoE)-driven approach that incorporates system performance, contextual awareness, and human factors into the design and optimization of service delivery. A service-oriented architecture (SOA) is introduced to automate the integration of vehicular resources using intelligent multiagent systems and to extend cloud-based services. The architecture is discussed in the context of several smart city applications, including real-time traffic management, cooperative driving, environmental monitoring through sensor data, and autonomous vehicle navigation, demonstrating its potential to turn untapped vehicular capabilities into meaningful urban solutions. Simulation results and case study analysis confirm that the proposed framework is both feasible and effective in enabling scalable, QoE-aware, and human-centered smart city transformation.
Artificial intelligence (AI) has revolutionized multiple sectors, ranging from health care to entertainment. Remarkably, despite this progress, today’s AI tools, including deep learning and generative AI (e.g., large language models), still fail when embedded into physical systems, such as robots or vehicles, that operate under the physical laws of the real world. This limitation stems from the inability of physical AI systems to maintain reliable world models for long-horizon planning under uncertainty and to generalize to unseen scenarios. In this context, wireless networks, through their pervasive sensing and communication infrastructure, can act as orchestrators of physical intelligence. However, current wireless architectures optimize for throughput, latency, and reliability and cannot support real-time physical AI coordination where agents must maintain a shared spatiotemporal context. To overcome such challenges, in this article, a network of holonic digital twins (HDT-Nets) is proposed to deliver the real-time physical AI inference requirements through holonic agents that actively reason about their environment, rather than passively mirror physical assets. Each HDT is realized as a hierarchical structure spanning the physical agent and the network edge, reasoning autonomously at the local level while cooperating with neighboring HDTs to form larger collectively intelligent units at the network level. In HDT-Nets, causal Markov blankets (MBs) spanning sensing, communication, and control domains determine which agents must coordinate and enable counterfactual reasoning over multidomain interventions. Active inference, operating within these boundaries, unifies perception, action, and learning by minimizing expected free energy, simultaneously deciding which beliefs to transmit based on their cognitive value to the receiver. Category theory guarantees that these transmitted beliefs preserve semantic structure when crossing between heterogeneous agents with incompatible representations. Finally, integrated information theory (IIT) provides a quantifiable metric to capture when collective intelligence exceeds independent operation and how network intelligence evolves through coordinated learning and information exchange.
Marine networks are more complex than terrestrial networks because of the challenging sea movements and sparse marine nodes. Infrequent link connections are caused by this sparsity, leading to high network latency and overhead. These instabilities in communication links among marine nodes can lead to unstable paths, resulting in data messages not reaching their destination. This can risk the safety of marine transportation when messages involve ship navigation, weather alert messages, or information-related messages for the helmsman. Integration of autonomous vehicles from different environments can result in safe and stable communication links while decreasing node sparsity. However, one of the main problems with these autonomous nodes is battery consumption. Incorporating software-defined networking (SDN) technology into the integrated marine-air network can result in increased network intelligence and flexibility. In this article, we consider a software-defined reconfigurable intelligent surface (RIS)-assisted integrated network to address the problems of unstable network paths and energy consumption.
Connected autonomous vehicles (CAVs) are envisioned to transform future mobility by enhancing safety, efficiency, and passenger experience through communication. However, the integration of networking into autonomous driving also introduces new vulnerabilities and complex tradeoffs across multiple performance dimensions. This article positions trustworthiness as the cornerstone of CAV design and explores its interdependence with efficiency and comfort—the three core performance metrics (CPMs) for evaluating system performance and public acceptance. We review the elements of trustworthiness, including safety, security, privacy, reliability, resilience, fairness, and accountability, and analyze their inherent synergies and conflicts with efficiency and comfort. Furthermore, we present frameworks for building trustworthy CAVs under both centralized decision-making (CDM) and distributed decision-making (DDM) paradigms, supported by simulation-based evaluations of reliability and latency across different network topologies. The findings demonstrate that centralized architectures generally favor efficiency and comfort, whereas distributed approaches improve resilience and accountability at the cost of higher latency. By highlighting standards, tradeoffs, and architectural choices, this article offers practical insights into designing CAV systems that balance trust, performance, and user experience in real-world deployments.
Traffic safety in the 6G era will depend on vehicles exchanging information that is not only timely but also accurate and contextually relevant for cooperative awareness, collective perception, and automated driving. Classic performance metrics such as latency and reliability describe communication behavior, but do not capture how shared data supports situational awareness. Timeliness- and semantic-oriented metrics including Age of Information (AoI) and Value of Information (VoI) provide deeper insight into freshness and usefulness, yet open questions remain on how to assess performance across increasingly complex vehicular functions. This paper reviews the state of the art in timeliness metrics for vehicular networks, surveys the corresponding standardization landscape, and examines the applicability and limitations of AoI-based and related metrics. It highlights persistent gaps in evaluating cooperative and automated driving services and outlines directions toward goal-oriented performance assessment suited to safety-critical 6G transportation systems.
Ensuring public safety in intelligent transportation systems (ITS) requires reliable and energy-efficient execution of vehicular applications under highly dynamic conditions. However, the inherent mobility of vehicles, fluctuating network quality, and constrained energy budgets often lead to task failures in 5G-enabled vehicular edge computing (VEC) environments. This article proposes an energy-aware task orchestration framework designed to minimize failure rates and optimize execution paths across mobile, edge, and cloud tiers. The framework integrates two adaptive mechanisms, transmission power control (TPC) and dynamic power scaling (DPS), to respond to real-time variations in signal strength, task complexity, and energy availability. A classification model predicts task feasibility across tiers, while a regression model estimates service responsiveness. The system selects the optimal execution path based on predicted success likelihood, energy profiles, and network conditions. Evaluated in a SUMO-integrated EdgeCloudSim environment with more than 1,000 vehicles, the proposed approach demonstrates more than 40% reduction in task failures under high-load scenarios. The framework supports diverse ITS applications, including danger assessment, traffic management, and infotainment, and offers a scalable solution for enhancing the reliability and safety of 5G vehicular networks.
This article investigates satellite navigation and communication systems in both low-Earth orbit (LEO) and medium-Earth orbit (MEO) satellites and systematically outlines the fundamental principles of satellite navigation systems (SNSs), satellite communication systems (SCSs), and integrated navigation and communication (INAC) systems. By exploring the enhanced capabilities of satellite systems, the article emphasizes how INAC systems improve overall functionality by enabling efficient signal multiplexing and multiple access, positioning multifunctional satellites as promising alternatives to traditional architectures. Moreover, the article introduces emerging frontiers for LEO-based SNSs and MEO-based SCSs through the integration of advanced 6G wireless technologies, which cannot be realized through mere extensions of existing communication or navigation techniques. Motivated by these insights, the article further discusses various conceptual transitions required to unlock the full potential of INAC systems, with a particular focus on channel capacity, positioning accuracy, and artificial intelligence (AI)-enabled waveform design.
In emergency response vehicular networks (ERVNs), where each second might dictate a life-or-death outcome, the performance of underlying communication systems is critical. Blockchain technology is a promising backbone for authenticating emergency response vehicles (ERVs) at smart traffic signal intersections. However, the decentralized nature of blockchain networks makes them vulnerable to distributed denial of service (DDoS) attacks. A DDoS attack can compromise safetycritical functions by overwhelming the network with fake authentication requests. Therefore, this article presents a comparative performance analysis of two well-known blockchain technologies, namely Ethereum (permissionless) and Hyperledger Fabric (HLF, permissioned), under DDoS-based network congestion. We clarify that this article does not propose any DDoS detection or mitigation mechanisms; rather, we evaluate the ability of the abovedescribed blockchains to maintain service for legitimate transactions under DDoS attacks. We design three DDoS attack scenarios that simulate realistic threats to ERVNs and benchmark throughput and latency using Hyperledger Caliper. Our results show that HLF, because of its architecture, achieves up to 2.7 times higher throughput and lower latency compared to Ethereum. This finding provides an important insight: A permissioned blockchain offers better performance under DDoS attacks, making it a more reliable communication system for safety-critical transportation systems.
As the crown jewel of intelligent transportation, connected and automated vehicles (CAVs) have sparked a phenomenal burst of interest to ultimately achieve the accident-free transportation system with high control capacity. However, current CAVs rely mainly on vehicle onboard equipment and edge devices for sensing and communication, and their accuracy and reliability are severely restricted by complex road environments and arbitrary human behavior. Additionally, the testing costs for intelligent vehicle-road collaboration systems and automated vehicles are exorbitant, coupled with extremely high deployment and operational complexities. Moreover, improper decision making can easily precipitate traffic accidents, thereby reducing the safety and efficiency of transportation systems. In lieu of this, we propose a digital twin (DT)-enabled vehicle-road cooperation system (VRCS) with the large-scale deployment of roadside Internet of Things (IoT) devices, which enables the ubiquitous perception and virtualization of CAVs. It also provides real-time digital mapping and intelligent control for the coordination between CAVs and smart road infrastructure. Specifically, we first propose an integrated "device-edge-DT" vehicle-road cooperation architecture to achieve the optimum utilization and digitalization of both vehicle onboard and roadside resources. Furthermore, a DT-driven sensor scheduling framework is employed to satisfy the information requirements for the VRCS. Finally, simulations show that the proposed DT-enabled vehicle-road cooperation architecture outperforms the baseline with more than 15.6% greater robustness, 53.6% higher traffic efficiency, and 8.3% better real-time responsiveness, advancing the vision of safe, efficient, and intelligent transportation.
Free-space optical (FSO) communication is emerging as a key backhaul technology for next-generation vertical heterogeneous networks (VHetNets), whose architecture spans satellites, high-altitude platform stations (HAPSs), autonomous aerial vehicles (AAVs), and terrestrial nodes. Along these vertical and slant paths, optical beams traverse successive atmospheric layers that may contain clouds, fog, rain, and aerosols, conditions that conventional single-coefficient Beer-Lambert (BL) models typically handle only in isolation. Instead of such simplified formulas, we present a unified attenuation model that incorporates aerosols, fog, rain, cloud layers, and drizzle, accounts for the zenith angle, and provides a holistic estimate of the cumulative power loss across atmospheric layers. Numerical results show several-decibel attenuation variations across representative weather scenarios, while the difference between the proposed model predictions and the layer-resolved Moderate Resolution Atmospheric Transmission (MODTRAN) simulations remains within 1 dB, thereby validating the accuracy of the proposed model and its practical relevance for VHetNet link budget studies.
The sixth generation of mobile networks is expected to revolutionize the way people, devices, and services connect by combining ultralow latency, massive scalability, and pervasive intelligence. To achieve these ambitious goals, a complete redefinition of the network architecture and the integration of artificial intelligence (AI) as a core enabler are equally important. A flexible and adaptive architecture is essential for supporting diverse requirements, ranging from immersive applications and autonomous systems to real-time digital twins and mission-critical communications. This article explores the pivotal role of AI in shaping 6G, with particular emphasis on distributed AI as a foundation for managing applications and services across heterogeneous infrastructures. By extending intelligence from traditional cloud environments to the edge and far-edge/device layer, distributed AI empowers networks to dynamically allocate resources, efficiently orchestrate services, and adapt in real time to user demands. We present the vision of an AI-driven fully distributed network architecture that integrates orchestration frameworks, closed-loop automation, and collaborative AI/machine learning (ML) pipelines. Ultimately, distributed AI emerges as the cornerstone of a service-oriented 6G infrastructure, paving the way for intelligent connectivity that can meet the diverse needs of future digital societies.
Autonomous mobile robots (AMRs) operating in industrial and urban environments must function efficiently under dynamic and energy-constrained conditions. Conventionally, motion planning and energy management are designed separately: planners generate feasible trajectories, while the energy management system (EMS) reactively supplies power without anticipating future demand. In hybrid robots powered by batteries and fuel cells (FCs), this separation reduces long-term efficiency and accelerates component degradation. This work introduces a health-aware planning and predictive EMS architecture in which the motion planner explicitly incorporates battery state-of-charge (SoC) and FC state-of-health (SoH) predictions into trajectory optimization, while the EMS performs anticipative power allocation. By shaping motion to avoid stress-inducing load patterns, the framework promotes stable operation and durability. Long-duration evaluations across warehouse, urban delivery, and industrial forklift scenarios demonstrate reduced hydrogen consumption, smoother FC operation, and controlled battery cycling.
