This paper investigates secrecy-oriented access point (AP) refinement for scalable cell-free massive multiple-input multiple-output (CF-mMIMO) systems after detection of an active pilot spoofing attack. A single-antenna eavesdropper transmits the same pilot sequence as the target user equipment (UE), contaminating local channel estimates and inducing coherent downlink leakage. For correlated Rayleigh fading and normalized maximum-ratio (MR) precoding, an expression is derived for the leakage contributed by each access point. The result shows that the leakage is governed by a coupling between the target-UE and eavesdropper covariance matrices weighted by the inverse covariance of the pilot observation. This relation motivates a score for ranking candidate AP–UE links. The score determines only the trial order; a link is removed only when the conservative secrecy spectral efficiency (SSE) metric increases. Because repeated SSE evaluations dominate the refinement cost, an evaluation budget limits the number of tested candidates. A residual-based procedure also estimates the effective attacker covariance from observations of the detected pilot while enforcing positive semidefiniteness. With a budget fraction of 0.2, the proposed ordering recovers approximately 90.4 percent and 87.9 percent of the reference gain obtained by testing all candidates under MR and local partial minimum mean-square error (LP-MMSE) precoding, respectively. Under a common genie-aided accept/reject benchmark, covariance estimates formed from finite samples retain most of the gain achieved with perfect covariance at the same budget. LP-MMSE precoding achieves higher SSE, whereas refinement yields a larger relative benefit under MR precoding.
Boundary node detection in wireless sensor networks (WSNs) and Internet of Things (IoT) networks is a fundamental problem with applications in perimeter surveillance, coverage verification, and network management. The Least Polar-angle Connected Node (LPCN) algorithm provides an exact solution by tracing the polygon hull of a connected Euclidean graph, but its distributed version (D-LPCN) requires sequential communication that is vulnerable to node failures and incompatible with dynamic networks. In this paper, we investigate whether a Graph Neural Network (GNN) can learn to approximate the LPCN boundary classification from data. We propose a complete methodology comprising: (1) automated generation of planar mesh networks, (2) extraction of 22 geometric and topological node features, (3) a message-passing GNN architecture with 3 propagation layers, and (4) a training data scaling study comparing models trained on 2K, 10K, and 100K networks. Comprehensive evaluation in 8 network configurations shows that the 100K-trained GNN achieves F1 = 0.992 in the training configuration and an average F1 = 0.963 in all settings, with perfect recall (zero missed boundary nodes) in 5 of 8 configurations. A real-world execution time analysis using ZigBee shows that D-LPCN requires ∼1.2s for 50 nodes due to sequential communication, while the distributed GNN completes in ∼141ms in parallel. We demonstrate that the GNN is fundamentally more robust to node failures and applicable to dynamic mobile networks where D-LPCN cannot operate. The 116 KB model runs entirely in-browser, enabling edge deployment without server infrastructure, as demonstrated by two companion web applications: an interactive step-by-step explainer of the complete pipeline and a real-time vehicle-tracking prototype operating on real OpenStreetMap road networks.
The rapid advances in wireless communication technologies have driven the adoption of distributed artificial intelligence in next-generation wireless networks. However, traditional centralized learning approaches become increasingly inefficient due to limited communication bandwidth, stringent latency constraints, and growing concerns regarding data privacy and security. Federated learning (FL) has emerged as a promising distributed paradigm that enables collaborative model training across edge devices without sharing raw data. Despite these advantages, its deployment in wireless networks remains challenging due to the scarcity of communication resources and the complexity of managing the heterogeneous devices. To date, however, resource allocation strategies in FL have not been extensively investigated from a systematic optimization perspective, while resource-aware system design largely underexplored. To bridge this gap, this survey provides a comprehensive and structured review of resource allocation in wireless FL systems. A unified analytical framework is introduced to characterize resource allocation strategies along three key dimensions: decision variable design, objective function formulation with associated performance metrics, and solution methodologies. Resource allocation strategies are further categorized to reflect increasing multi-resource demands. It highlights the importance of joint optimization and its underlying mathematical coupling in enhancing system efficiency and learning performance. Mathematical reformulation patterns are synthesized to explain common approaches for transforming complex and non-convex resource allocation problems into more tractable forms for optimization. A comparative and critical analysis of major solution paradigms is presented, encompassing classical optimization, metaheuristic algorithms, reinforcement learning, game-theoretic frameworks, and hybrid methods. Finally, this survey identifies key research gaps and promising future directions for integrating FL into next-generation wireless networks, including 5G/6G systems.
In cellular networks, mobility management and resource allocation are two critical tasks that directly impact the quality of service (QoS) for mobile users. Traditional approaches often address these tasks separately, which may not fully exploit the shared information between them and may be suboptimal in dynamic environments. In this paper, we propose a multi-task learning (MTL) framework that jointly learns handover and resource allocation decisions. Our model uses a shared encoder to extract a compact latent representation from multi-domain features, while two task-specific heads perform predictions. A reconstruction branch is incorporated to regularize the latent representation and improve model robustness. The model is trained and evaluated on a large-scale simulated dataset, generated under comprehensive network conditions and user mobility patterns. The proposed framework achieves competitive performance compared to strong baseline models for both tasks. Experimental results show that the proposed model achieves strong classification performance, reaching 94.15% accuracy for HO prediction and an F1-score of 0.925 for resource allocation strategy classification. More importantly, the joint learning framework provides clear system-level benefits. Compared with SVM, random forest, KNN, and single-task MLP baselines, the proposed method improves average user throughput by up to 15.6% and reduces delay by up to 19.5%, at speed 30 m/s. These results demonstrate that shared representation-based MTL can support joint handover (HO) and resource allocation (RA) strategy decision learning and can improve throughput and delay under the considered simulation setting, while maintaining efficient inference suitable for practical deployment in intelligent RAN systems.
The inverse-gamma (IG) distribution has recently emerged as a mathematically tractable and empirically supported model for large-scale shadowing in wireless communication systems. Although several IG-based fading and shadowing models have been investigated in terms of first-order statistics, their second-order temporal behavior has received limited attention. This paper develops an analytical framework for the level crossing rate (LCR) and average fade duration (AFD) of single- and double-IG shadowing processes. The joint probability density function (PDF) of the IG process and its time derivative is derived, leading to closed-form expressions for its LCR and AFD. The analysis is then extended to an independent double-IG product process, for which the joint PDF of the process and its time derivative is formulated, yielding an exact single-integral LCR expression and the corresponding AFD. The framework is further applied to a UAV-to-ground double-shadowing link, where the UAV-side and ground-station-side IG shadowing components may be statistically dependent. Under a conditional Gaussian derivative model, a closed-form PDF and expressions for the moments, together with exact integral expressions for the CDF, LCR, and AFD, are obtained for the received signal-to-noise ratio (SNR). Time-domain Monte Carlo simulations validate the single-IG second-order results, whereas conditional Monte Carlo evaluations of Rice’s formula confirm the numerical consistency of the correlated double-IG expressions under the adopted derivative model. The results quantify the effects of the IG parameters, shadowing dependence, and temporal variation rates on the frequency and duration of shadowing-induced outage events.
Diverse 5G slices (eMBB, URLLC, and mMTC) with distinct SLA/QoS requirements make bandwidth orchestration challenging in limited-spectrum environments. Static rule-based and model-based methods struggle with traffic, mobility, contention, and configuration changes, whereas DRL adapts by learning directly from interactions with the environment. We design and evaluate a model-free, DRL-based Multi-Objective Utility-based Intent-Driven Orchestration (MoU–IDO) framework in a 3GPP-compliant, multi-slice, multi-BS RAN. The framework prioritizes slice-intent fulfillment by enforcing SLA (latency and throughput) and QoS (service availability) requirements and subsequently optimizes spectral efficiency (SE) under limited bandwidth. Employing the proposed framework, five representative DRL algorithms (A2C, PPO, DQN, ACER, and SAC) are benchmarked against a heuristic baseline across RAN (MACRO-rural and DENSE-urban) scenarios using in-distribution and out-of-distribution train–test evaluations. Results show that MoU–IDO preserves latency feasibility, with eMBB and mMTC near the minimum operating region of ≤ 1 ms and URLLC within the subsequent sub-ms range. QoS satisfaction remains high for eMBB and mMTC, at approximately 95%, whereas URLLC is the primary bottleneck for generalization. Under MACRO cross-testing, DQN drops to nearly 46% URLLC QoS, whereas A2C remains around 91%, resulting in a 35–45 percentage-point reliability gap. During testing, eMBB throughput increases by approximately 30–50% in MACRO and 40–60% in DENSE, whereas URLLC throughput decreases by approximately 20%. RL agents also achieve an approximately eightfold improvement in utility over the heuristic, confirming stronger cross-slice adaptation.
Multi-hop wireless ad-hoc networks (WANETs) are expected to expand significantly over the next years. The limited resources characterizing many WANET deployments, make necessary the efficient use of resources. One common practice to improve efficiency is congestion control. The highly dynamic settings and the need for decentralized control found in WANETs mean that a congestion control protocol applied to such networks should be able to adapt to changes in the environment and allow each node to make its own decisions based on their own particular context. Reinforcement Learning (RL) allows each node (agent) to learn a congestion control model while operating in the network, adapting to changes in the environment. Although RL-based congestion control has been widely studied in TCP environments, where connection-based communications and packet acknowledgment messages make the feedback mechanism explicit, limited research is available for UDP-based, connectionless communications. In this paper, we propose a multi-agent RL congestion control mechanism for WANETs, introduce a traffic prioritization mechanism and, through simulations in diverse conditions, compare our solution with existing ones, both in terms of performance and energy awareness. The proposed method increases delivered throughput by 29–80% compared to the best baseline across varying congestion and physical-layer conditions, while maintaining low energy consumption.
Cell-free multiple-input multiple-output (CF-MIMO) systems assisted by reconfigurable intelligent surfaces (RISs) and rate-splitting multiple access (RSMA) are promising for interference management in distributed sixth-generation (6G) networks. In practical CF-MIMO deployments, independent access-point (AP) oscillators and unequal propagation delays introduce AP-dependent phase variations that reduce coherent combining and degrade precoding performance. This paper develops a unified RIS-RSMA optimization framework for asynchronous downlink CF-MIMO. The RSMA common precoder, private precoders, and passive RIS phase shifts are jointly optimized within a sum-rate maximization problem subject to total transmit-power, per-user quality-of-service, and RIS unit-modulus constraints. The resulting non-convex problem is addressed through a weighted minimum mean-square error (WMMSE)-based alternating optimization framework with Gauss–Seidel coordinate optimization for the RIS phase subproblem. An analytical common-stream power expression is derived to explain the limited benefit of decoupled common-precoder designs. The analysis shows that independent oscillator phase drifts at the APs attenuate coherent cross-AP combining terms and can restrict the common rate through the weakest-user decoding constraint. In the default overloaded CF-MIMO configuration, the tested heuristic common-precoder schemes provide less than 0.2% gain over the corresponding optimized space-division multiple access (SDMA) baseline when combined with WMMSE-optimized private precoders. In contrast, the proposed joint RSMA design achieves approximately 12–15% sum-rate gain at moderate-to-high transmit powers and retains a similar high-power advantage in a larger-scale validation. Targeted comparisons using WMMSE and regularized zero-forcing (RZF) private precoders, together with the common-power allocation results, support the need for joint common/private precoder design. Additional results show that the gain persists across the tested RIS sizes, oscillator phase-noise variances, Rician factors, loading ratios, channel realizations, and multi-RIS deployments. Distributed multi-RIS deployment also improves spatial coverage uniformity under a fixed total element budget.
Low Earth Orbit (LEO) nano-satellite communications are playing an increasingly vital role in supporting ground-based Internet of Things (IoT) sensing applications. A key challenge is designing grant-free uplink communication access schemes that are fair and scalable, with minimal feedback to the ground terminals. This paper proposes a novel weighted attempt rate (WAR) scheme. We formulate an ordinary differential equation (ODE) to find the optimal set of terminal weights that achieve max–min fairness. The ODE’s convergence to an optimal point is rigorously analyzed, and an ODE solver is developed. We also address the feedback overhead problem for dynamic IoT applications where terminal weights need to be changed on a per satellite pass basis. We propose a quantized time-grid clustering method with a corresponding feedback mechanism, which delivers the weights to terminals in real-time during the satellite pass. The number of weights depends only on the time-grid dimensions, thus ensuring that downlink signalling is independent of the terminal population size. Extensive numerical studies validate the convergence, robustness, and reliability of the proposed scheme, while also illustrating the inherent trade-off between feedback overhead and fairness performance.
In this paper, we propose a novel digital post-distortion (DPoD) technique designed to improve the received signal quality in orthogonal frequency-division multiple (OFDM)-based multi-user multiple-input multiple-output (MU-MIMO) systems suffering from severe coexisting transmitter impairments in the form of power amplifier (PA) nonlinearities and in-phase/quadrature (I/Q) imbalance. The proposed one-shot or non-iterative processing approach combines augmented memory polynomial-based joint impairment modeling and cancellation with practical parameter estimation using 3GPP 5G NR standard-compatible demodulation reference signal (DMRS) structures. Numerical evaluations conducted in extensive 5G NR MU-MIMO uplink scenarios encompassing both 2.1GHz and 28GHz frequency bands as well as LOS and NLOS channel conditions reveal significant gains in receiver performance, measured through effective error vector magnitudes (EVMs), while similar EVM gains are shown also with measured 6G mid-band PA responses at 7GHz. The results show that the proposed method maintains its effectiveness even under aggressive transmitter impairments and high-order modulation schemes including 64-QAM, 256-QAM, and 1024-QAM, thus offering a promising path for improved power-efficiency and coverage or link-distances in future wireless networks toward the 6G era. The numerical results also show and highlight that the proposed Joint-DPoI approach achieves lower receiver EVM values compared to prior-art methods where coexisting impairments are neglected. Finally, it is also demonstrated that the proposed DPoD approach is robust against residual carrier frequency offsets, reasonable user mobilities, frequency-selective impairments, and different types of PA nonlinearities.
Uninterrupted 5G Standalone (SA) coverage is essential for time-critical applications, particularly in Connected and Automated Mobility (CAM) scenarios. A key challenge in this context is minimizing Cross-Border Interruption Time (CBIT), which can significantly impact the performance of time-critical connected driving and high-bandwidth infotainment applications such as platooning, safe overtaking, and virtual reality immersive services. Although several EU-level initiatives have proposed solutions for reducing CBIT, many existing approaches rely on non-standard procedures, have limited scalability, or lack sufficient empirical validation. This paper presents a standard-compliant and scalable cross-border roaming solution based on a multi-domain architecture that includes a compact core network, and an inter-PLMN handover mechanism enhanced with a novel two-DNN switching solution. The system is validated through field trials conducted at the Estonian–Latvian border. Furthermore, this paper presents a measurement methodology and a comprehensive analysis of CBIT, providing detailed insights into the individual timing components of roaming procedures. Results from CAM use cases, including See-Through and Cooperative platooning, demonstrate the effectiveness of our approach in reducing CBIT and minimizing service disruption across borders.
The continued growth of Healthcare Internet of Things (HC-IoT) applications has transformed patient monitoring, diagnosis, and treatment from traditional practices into intelligent, data-driven processes that support a wide range of healthcare tasks. However, the effectiveness of these systems depends on reliable, secure, and context-aware communication among heterogeneous medical devices, communication networks, and clinical infrastructures. Large Language Models (LLMs) have recently been explored to improve data interpretation, coordination, and decision support in HC-IoT applications. However, most existing solutions use LLMs for perception and reasoning, while device control, routing, bandwidth allocation, and resource management still depend on predefined rules. Therefore, they cannot adapt the communication process in real time when network conditions, device status, or patient needs change. Large Action Models (LAMs) offer a possible path beyond this limitation by connecting reasoning with autonomous action. Through tools and control interfaces, they can reconfigure network routes, allocate bandwidth, coordinate devices, and manage communication resources according to changing system goals. In this survey, we examine how LAMs can support communication intelligence in HC-IoT by combining perception, reasoning, and real-time control. For this, we cover the HC-IoT application communication architectures, protocols, and standards published from 2017 to 2026, with particular attention to 5G/6G networks, edge and cloud collaboration, adaptive routing, and quality-of-service management. We also present a unified taxonomy and architectural framework for HC-IoT applications using LAMs to manage latency, reliability, throughput, and resource allocation across medical networks. Finally, we outline open challenges and future directions for tactile and ultra-reliable healthcare communications to guide the development of efficient and intelligent LAM-enabled frameworks for next-generation HC-IoT applications.
This paper serves as a foundation for researchers and practitioners to explore the key concepts, primary applications, and potential impacts of xApps in the open radio access network (O-RAN) architecture. We provide background information and examples related to xApps and discuss several fundamental concepts to familiarize readers with the key notions presented throughout the paper. In addition, we propose an architectural framework for the development and deployment of xApps within the near-real-time RAN intelligent controller (Near-RT RIC), offering a concise examination of its features, components, and interactions through internal and external open interfaces. Moreover, we explore three major aspects of the Near-RT RIC: the Near-RT RIC platform, the Near-RT RIC-related application programming interfaces (APIs), and the E2 service model (E2SM). Our primary objective is to provide a broader perspective on the theoretical foundation and practical implementation of xApps. Specifically, (a) we examine the Near-RT RIC platform to understand how its functionalities facilitate the deployment and monitoring of xApps, (b) investigate the APIs to comprehend the interconnections and data exchanges between xApps, the Near-RT RIC platform, and other components within O-RAN, and (c) analyze the E2SM, which governs the interactions and actions of xApps with E2 Nodes. Furthermore, we shed light on the management and orchestration of a collection of xApps within the Near-RT RIC, as well as present the lifecycle management of an individual xApp, detailing the phases from design to termination. Subsequently, we demonstrate the practical implementation of an open-source xApp (the key performance indicator [KPI] Monitoring xApp) in our O-RAN testbed. To accomplish this, we offer a detailed overview of the xApp, discuss the setup and architecture of our testbed (including the hardware and software components used), and provide a step-by-step guide to implementing this xApp in the testbed. Following that, we explain xApp-related conflict mitigation, addressing a range of topics, with a particular focus on the types of xApp-level conflicts and corresponding mitigation strategies. We then present several lessons learned from our research on xApps and their practical deployment. Finally, we identify several research and engineering challenges that need further effort to develop novel xApps and ensure their seamless deployment in O-RAN.
Confronted with the stringent demands of the sixth-generation (6G) wireless networks, such as ultra-high data rates, minimal latency, and high energy efficiency, conventional fully-digital massive multiple-input multiple-output (MIMO) systems face significant challenges in scalability, hardware complexity, and computational overhead. To address these challenges, stacked intelligent metasurfaces (SIMs) emerge as a promising technology that leverages multi-layer programmable meta-atoms to perform complex signal processing tasks directly in the wave domain at the speed of light. This paper provides a comprehensive survey of SIM, beginning with its physical principles and hardware design. We then review novel SIM-enabled signal processing frameworks and summarize their reported performance gains across representative applications, including holographic MIMO, direction-of-arrival estimation, multi-user beamforming, satellite communications, as well as integrated sensing and communications. We also discuss the integration of SIM with machine learning methods, including reinforcement learning and semantic communications, highlighting the advantages of low latency and high energy efficiency inherent to wave-based computation. Additionally, we identify key practical challenges including hardware imperfections, inter-layer coupling effects, high-dimensional optimization complexity, and system-level integration, while proposing viable solutions such as model calibration, robust training frameworks, and optimization strategies powered by artificial intelligence. Emerging evidence suggests that SIM technology possesses significant potential to reduce reliance on conventional power-intensive digital baseband processing modules, thereby establishing a scalable, sustainable, and intelligent infrastructure for integrated sensing, communication, and computing networks in the 6G era.
Cell-free massive multiple-input multiple-output (CF-mMIMO) is a promising architecture for future sixth-generation (6G) wireless networks because cooperation among geographically distributed access points enables seamless connectivity, more uniform quality of service, and improved spectral and energy efficiency. However, most existing performance evaluations rely on idealized or simulated channel models. This paper presents a measurement-driven system-level evaluation of downlink CF-mMIMO orthogonal frequency-division multiplexing (OFDM) using urban channel measurements acquired at 5.89 GHz. The considered framework combines user-centric clustering, distributed partial zero-forcing (PZF) precoding, and local peak-to-average power ratio (PAPR)-aware precoding under nonlinear power-amplifier constraints. Using the measured channels, we investigate the spectral-efficiency and energy-efficiency tradeoffs associated with access-point density, the number of antennas per access point, and the power-amplifier input back-off. Increasing the number of antennas per access point significantly improves spectral efficiency, although it progressively shifts the deployment away from the cell-free paradigm by reducing the number of distributed access points. In the considered fixed antenna budget, the configuration with 32 antennas per access point and 13 access points achieves the highest energy efficiency among the evaluated configurations. PAPR-aware techniques, with power amplifiers operating at an input back-off of 3 dB, provide up to a 2.5 times improvement in spectral efficiency and a 120 percent increase in energy efficiency, indicating the potential of CF-mMIMO under realistic propagation and hardware constraints for 6G wireless deployments under real-world propagation and hardware conditions.
Open-RAN is transforming 5G and Beyond-5G networks by promoting openness, flexibility, and interoperability across network components. However, this architectural shift introduces new security challenges due to expanded attack surfaces, distributed control mechanisms, and exposed interfaces. Traditional Intrusion Detection Systems (IDS) struggle to address the dynamic threat landscape of O-RAN environments. To address this challenge, we propose ORANGuard, an Artificial Intelligence-driven dual-tier intrusion detection and mitigation framework for O-RAN environments. ORANGuard comprises two core modules: Real-Time Guard (RT-Guard), a real-time threat detection model deployed in the Near-Real-Time RAN Intelligent Controller (Near-RT RIC), and Non-Real-Time Guard (NRT-Guard), a periodic anomaly analysis model operating within the Non-Real-Time RIC, together enabling multi-level threat detection. The framework is built on an end-to-end 5G O-RAN testbed using srsRAN and Open5GS, capturing E2 measurements under both benign and attack conditions. Custom scripts simulated anomalous scenarios, including DDoS, producing a dataset of 17,052 E2 measurement records across a testbed of 4 gNodeBs and 12 UEs. RT-Guard, based on Random Forest, achieved 96.7% accuracy, while NRT-Guard, using an Autoencoder, reached 95.6%. The hierarchical dual-tier decision rule achieved 99.5% accuracy on the held-out test set. RT-Guard required a median detection latency of 5.2 ms and a median detect-and-enforce latency of 17.9 ms, whereas NRT-Guard required 283.6 ms after closure of its one-second analysis window. These results demonstrate the complementary low-latency and anomaly-analysis capabilities of ORANGuard for O-RAN security.
In next-generation wireless networks, electromagnetic signals are envisioned to be radiated from large-scale radio-frequency transmitters operating in the millimeter-wave and sub-terahertz bands. The combination of electrically large radiating apertures and high-frequency transmission extends the radiative near-field region around the transmitter. In this region, unlike in the far field, the wavefront is nonplanar, which provides additional degrees of freedom to shape and steer the transmitted beam as desired. In this paper, we focus on Airy beams, which may exhibit several highly desirable properties in the radiative near-field region. In their ideal form, these beams can follow self-accelerating (curved) trajectories, exhibit resilience to perturbations through self-healing, and maintain a shape-preserving intensity profile in a co-moving transverse frame, making them effectively diffraction-resistant. Specifically, starting from diffraction theory as the foundational propagation model for radiative near-field free-space beam manipulation, we first present the underlying principles of self-accelerating beams radiated by continuous aperture field distributions. We then address several challenges regarding the generation of Airy beams, including their exponential decay due to finite energy constraints and spatial truncation of the aperture. Moreover, we examine their free-space propagation characteristics, focusing on a generalized link budget formulation and a polychromatic representation. The second part of the paper focuses on the propagation behavior of Airy beams in non-line-of-sight (NLoS) scenarios, which are particularly relevant for radiative near-field wireless communication applications. We also present a comparison between Airy beams and Gaussian beams, which represent the most common solution for focused beam transmission in the near field, evaluating their performance in terms of received energy. Our theoretical and numerical results show that Airy beams may offer a performance advantage over Gaussian beams in certain NLoS channels, provided that their key properties are largely preserved, specifically, self-acceleration along a parabolic trajectory and diffraction-resistant propagation. In the presence of an obstacle, this requires that the portion of the transmit aperture with a clear line-of-sight to the receiver is sufficiently large. These findings underscore the intriguing potential of Airy beams in radiative near-field wireless links, while also highlighting the importance of a concurrent electromagnetic and telecommunication design to fully harness their advantages in practical systems.
Energy imbalance remains a key challenge in Wireless Sensor Networks (WSNs), as nodes near the base station deplete their energy faster due to heavy forwarding loads. While mobile agents (MAs) have been employed for either data collection or sensor charging, existing approaches lack adaptability and fail to integrate both functions under realistic hardware constraints. This paper introduces a unified mobile agent framework that performs both data collection and wireless charging sequentially under single-antenna limitations. The agent’s decision-making is formulated as a two-layer Hierarchical Reinforcement Learning (HRL) problem, where the upper layer optimizes movement planning and the lower layer determines the appropriate service based on real-time network states. This hierarchical structure enables the agent to learn adaptive task scheduling policies without predefined rules. Extensive simulations demonstrate that the proposed method achieves up to 15% longer network lifetime and more balanced energy distribution compared with state-of-the-art mobile agent and deep RL approaches.
Reliable transmission of compressed images over error-prone channels remains challenging because compressed bitstreams are highly sensitive to channel impairments. Although existing quantum-inspired encoding schemes exploit superposition to improve transmission reliability, they remain vulnerable to channel noise, limiting reconstructed image quality. To address these limitations, this paper proposes a chirp transform-based multi-qubit encoding framework for robust compressed image transmission. Unlike conventional transform-based approaches, including the Hadamard transform and Quantum Fourier Transform (QFT), the proposed framework introduces nonlinear (quadratic) phase modulation into the multi-qubit representation, producing a high-dimensional superposition that enhances noise resilience through improved phase separability. This nonlinear phase variation improves error resilience through phase-induced decorrelation, reducing noise-induced distortions. The system integrates Joint Photographic Experts Group (JPEG) and High Efficiency Image File Format (HEIF) source encoding, followed by channel coding, multi-qubit encoding, and chirp transformation before transmission through a simulated quantum channel. At the receiver, inverse chirp transformation, measurement-based decoding, channel decoding, and source decoding reconstruct the transmitted image. Performance is evaluated using Bit Error Rate (BER), Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), and Universal Quality Index (UQI). Experimental results demonstrate that the proposed framework achieves up to a 3 dB signal-to-noise ratio (SNR) gain over QFT-based and Hadamard-based systems in the single-qubit configuration. For the eight-qubit configuration, gains of up to 2 dB over QFT-based systems and 7 dB over Hadamard-based systems are achieved. These results demonstrate the effectiveness and scalability of the proposed framework for robust compressed image transmission over noisy quantum communication channels.
The Open Radio Access Network (O-RAN) introduces openness and disaggregation to cellular networks, enabling innovation and multi-vendor interoperability. This article provides a comprehensive examination of O-RAN security with emphasis on two distinct contributions. First, we classify the O-RAN threat surface into three domains: infrastructure, open interfaces, and Radio Access Network (RAN) intelligence. This categorization provides a structured framework for analyzing vulnerabilities and attacks across the O-RAN architecture. Second, we survey state-of-the-art testing tools, including both open-source and commercial solutions, and map their capabilities to the identified threat surfaces. In addition, we review mitigation strategies, ongoing standardization efforts, and emerging defense mechanisms. Unresolved challenges and future research directions are highlighted to guide further research and development. This dual focus on systematic threat surface classification and security testing methodologies differentiates this article from prior work and provides a roadmap for researchers and practitioners securing O-RAN deployments.
