Consider a hybrid orthogonal multiple access (OMA) and non-OMA (NOMA)—HMA multicasting for Internet-of-things (IoT) systems, where a group of interested IoT devices is assigned to subchannels, some using NOMA and the others using OMA. Since multicasting is heavily constrained by the weakest channel gain of involved IoT devices, an advanced resource allocation strategy is required to enhance energy efficiency. This paper investigates a simultaneous optimization of power allocation and subchannel assignment to maximize energy efficiency in HMA, not resorting to alternating optimization (AO). Specifically, an optimization problem is formulated as a mixed-integer nonlinear programming (MINLP) problem, posing a significant challenge in finding a globally optimal solution. To address this, we propose two novel algorithms: the first is designed for small search spaces, using the Big-M method for linearization of a bilinear term followed by an iterative and threshold-based rounding, and the second is tailored for large search spaces, utilizing a hybrid proximal policy optimization (HPPO) technique to train hybrid action space. The simulation results confirm that both the proposed algorithms surpass other AO-based benchmarks by 7.1% and 5.6%, respectively, in terms of energy efficiency.
Smart home automation, a component of the Internet of things (IoT), enables users to manage home functions with smart sensors and actuators, providing convenience, energy efficiency, and remote monitoring. The Z-Wave protocol, widely adopted in smart homes for lighting, security, appliance control, and power management, remains vulnerable to various external attacks, highlighting the need for effective attack detection tools. The existing single-layer artificial neural network (ANN) model (i.e., ZMAD) performs well on familiar data; however, it has limitations over datasets with different distributions and advanced attack vectors. This paper introduces ZENA, a lightweight protocol-aware anomaly-based intrusion detection model for Z-Wave networks, which employs a multilayer ANN, built from scratch, to improve detection accuracy and robustness. Using a dataset with several attack classes and adversarial generated vectors, the proposed model achieves a precision rate of 95%, significantly outperforming ZMAD and state-of-the-art deep learning neural networks on same dataset (i.e., 89–93%). The results indicate substantial improvements in advanced detection and resilience against adversarial attacks, enhancing security for Z-Wave smart home systems.
On-board unit (OBU) is the core component for connecting intra- and inter-vehicle networks, which is also the component most vulnerable to cyber attacks. The conventional countermeasures mainly include encryption authentication, intrusion detection, and privacy enhancement; however, the operating environment for vehicle measurement and control applications is also required to be strengthened since vulnerabilities in operating system and complex programs are unlikely to be completely cleared. In this work, a novel dynamic heterogeneous operating environment (DHOE) based security architecture of OBU is proposed, which is the first structure to use heterogeneous application containers to hold OBU application programs, with a dynamic model that is able to detect the security state and switch the online/offline application containers to enhance security. Through the heterogeneous processing units and dynamic switching scheme, DHOE is able to detect and mitigate security problems and provide a credible operating environment for OBU applications. The proposed DHOE architecture has been successfully integrated into a real-world OBU product and applied in the Yutong¯ test vehicle. The Markov chain based theoretical analysis proves the security performance of DHOE in a quantitative way; while the case study and performance evaluation results on the real-world OBU product show that the proposed DHOE architecture could greatly improve the OBU security level while only 16.3% extra CPU loads and 1.6% extra RAM memory costs are introduced.
Shannon's classical information theory has long underpinned communication systems by ensuring bit-level fidelity, yet it remains agnostic to the semantic content of transmitted data. Semantic communication (SemCom) introduces a paradigm shift beyond Shannon's model by enabling the transmission of meaning-relevant information, thereby enhancing communication efficiency. By leveraging artificial intelligence (AI), SemCom prioritizes task-relevant features, achieving significant reductions in bandwidth usage and latency, while improving throughput and system performance. Despite these advantages, SemCom introduces novel security vulnerabilities. The inherent open ness of wireless channels, high correlation between source and transmitted representations, and structural weaknesses in AI models render SemCom particularly susceptible to eavesdropping and privacy breaches. These risks necessitate a reassessment of traditional security paradigms. This paper presents a comprehensive survey of emerging security strategies for SemCom against eavesdropping, encompassing physical-layer security, crypto graphic frameworks, data-driven approaches, and semantic-level covertness. Each class of techniques is analyzed with respect to its methodological foundations, effectiveness, limitations, and its deviation from conventional communication security mechanisms. Distinctively, this survey provides a comparative analysis of these strategies under model inversion-based eavesdropping attacks, offering a unique perspective absent in existing literature. The paper concludes by identifying critical research challenges and outlining prospective directions for secure and trustworthy SemCom systems.
We consider a scenario in which an unmanned aerial vehicle (UAV) performs transmit antenna selection (AS) and is assisted by a reconfigurable intelligent surface (RIS) for integrated sensing and communication. We optimize the antenna subset and transmit beamformer at the UAV and phase shift at RIS to maximize the beampattern gain towards the target under a specific rate constraint for the communication. To solve this, we propose an alternating optimization based algorithm in which the optimization problem is split into a sequence of sub problems and presented as semidefinite programs. Our numerical results show that an improved average beampattern gain and outage performance is obtained by increasing number of transmit antennas or number of RIS elements. We show that by increasing the RIS elements, we can obtain the savings of radiated power while keeping the average beampattern gain fixed, a system with subset AS can obtain improved performance compared to a multi antenna system. Our results also show that an improved average beampattern gain and outage performance can be obtained even when UAV goes away from the user and comes closer to the RIS and target. A discussion on real-time feasibility challenges and potential directions for future work is also presented.
Accurate acquisition of downlink channel state in formation (CSI) at the base station (BS) is crucial in frequency division duplex (FDD) millimeter wave (mmWave) massive multiple-input multiple-output (mMIMO) systems. Although compressive sensing (CS) and deep learning (DL)-based CSI feedback methods demonstrate their advantages, the recovery accuracy of downlink CSI still faces severe challenges due to the facts of high path loss, significant user equipment (UE) estimation errors, and typical compression requirements, etc. To tackle these issues, improving the accuracy of downlink CSI recovery has become an urgent task. Inspired by sensing-assisted communication techniques, an echo sensing information-assisted CSI feedback with echo sensing information method is proposed in this paper. In the proposed method, the communication echo signals observed at the BS are utilized to extract the dedicated sensing prior information for the downlink CSI recovery. With the extracted sensing prior information, a CSI denoising method is developed to suppress the non-path entries of the downlink CSI matrix in the angular-delay domain, thereby improving the recovery accuracy of downlink CSI at the BS. The proposed method establishes an embedding framework for improving the recovery accuracy of downlink CSI in FDD mmWave mMIMO systems. In this framework, the echo sensing information-assisted CSI recovery algorithm is directly embedded in the BS receiver to enhance the recovery accuracy of downlink CSI without modifying the UE transmitter. Simulation results demonstrate that the proposed method improves the recovery accuracy of downlink CSI compared to the classic DL-based and the CS based CSI feedback methods. Furthermore, the proposed method exhibits its robustness against the impact of parameter variations.
Intelligent reflecting surface (IRS) has recently received considerable attention from the wireless communications research community. In this paper, we investigate a secure communication system aided by an IRS, comprising multi-user and a single eavesdropper. Specifically, under the unit modulus constraint at the IRS and the transmit power constraint at the access point (AP), we maximize the minimum secrecy rate by jointly optimizing the beamforming vectors at the multi-antenna AP and the phase shift matrix of the IRS. In order to solve the non-convex optimization problem, we propose an alternating optimization (AO) algorithm and obtain a suboptimal solution. Firstly, to optimize the beamforming vectors, the semidefinite relaxation (SDR) technique is employed. Secondly, with the aim of addressing the issue of phase shift matrix optimization at the IRS, the successive convex approximation (SCA) method is applied. Simulation results demonstrate that our proposed scheme performs better than the benchmark schemes in terms of both algorithm convergence and secrecy rate performance.
6G technology, integrated sensing and communication (ISAC) is an emerging approach that enhances communication efficiency by enabling simultaneous sensing tasks. This paper leverages the multitasking capabilities of deep learning to optimize beamforming selection and user localization. A convolutional neural network (CNN) is employed to extract features from channel state information (CSI) data, which are then processed through a fully connected neural network to identify the optimal beamforming configuration and estimate user location. A weighted loss function is introduced to balance the importance of each task, ensuring that the model effectively prioritizes its objectives. Experimental results show that the proposed model achieves a Top-1 beamforming classification accuracy of up to 78.2% and a Top-3 accuracy of 99.21% with 64 antennas, while reducing localization error to as low as 2.11 meters. Compared to traditional single-task models, our approach improves classification accuracy by up to 7% and reduces localization error by up to 81%. This study highlights the potential of multitask learning in advancing ISAC capabilities and provides valuable insights for practical deployment in 6G systems.
has gained significant attention for Internet of Things (IoT) applications that require low-power and long-range communication such as animal tracking. However, LoRa's spreading factor (SF) selection mechanism and 1-hop topology lack effective per-link adaptation to dynamics of the communication channel nor provide comprehensive coverage of moving animals. To address this problem, we present AURORA, an adaptive and distributed SF control scheme for low-power multihop LoRa networks. AURORA exploits the key tradeoff of SF-receiver sensitivity vs. data rate-with a focus on energy efficiency. AURORA accurately predicts the packet delivery ratio (PDR) for each SF using efficient probing and model fitting techniques, enabling rapid per-link SF adaptation to improve both PDR and duty cycle in multihop networks. Through real-world experiments conducted on both indoor and outdoor testbeds, we demonstrate that AURORA effectively reduces energy consumption while ensuring reliable communication and extending network coverage. Compared to state-of-the-art approaches such as ADR+, AURORA improves the PDR by 14% and reduces the duty cycle by 18%.
The advent of software-defined networking (SDN) and network function virtualization (NFV) represents significant advances in networking and telecommunications. Their emergence reveals a change in networking paradigms that are more programmable and adaptable. Nevertheless, these technologies present new difficulties, especially regarding security. Conventional security measures are less effective in these complex, dynamic environments. In fact, through using artificial intelligence (AI) abilities, we can develop more robust and adaptive intrusion detection systems that can effectively identify and overcome challenging issues. Combining AI with SDN/NFV may assure a more secure network architecture by significantly enhancing security measures. In this paper, we propose the integration of deep learning (DL) and machine learning-based (ML) detection algorithms with the concepts of SDN and NFV, to enhance intrusion detection. DL/ML-based techniques have notable success in detecting novel and emerging types of network intrusions when supplied with sufficient and relevant training data. Anomaly detection methods rely heavily on data, which includes features that are representative of system behavior. Data fusion is essential in this context, as it combines information from multiple sources to provide a more comprehensive and accurate understanding of network activities. To address the challenge of finding relevant training data, this study introduces the SDN-Net dataset, which is the result of combining two pre-existing SDN-oriented datasets. SDN-Net has 79 features, 11 categories of traffic, and more than one and a half million rows of observations. By providing a dataset including a wide range of normal and abnormal network behaviors, we facilitate the training and testing of DL/ML models capable of detecting network threats and improving the security and resilience of SDN/NFV networks. The results demonstrate how, in terms of accuracy and efficiency, our AI-based intrusion detection model surpasses more conventional methods. We reached 99.99% precision and 99.9% accuracy when using different DL and ML models.
Traditional one-time password (OTP) systems rely on fixed-length numeric codes, making them vulnerable to replay attacks, phishing, and social engineering. This paper proposes an adaptive OTP generation framework that integrates AI-based risk assessment with context-aware two-factor authentication. The proposed system evaluates contextual metadata-including device information, geolocation, and user behavior-to classify risk levels in real time using a lightweight machine learning model. Based on the assessed risk, the OTP generation policy dynamically adjusts both length and character complexity. It produces simple six-digit numeric OTPs in low-risk scenarios and transitions to alphanumeric or symbolic OTPs exceeding ten characters in high-risk conditions. This adaptive strategy enhances security while preserving usability. The architecture is modular and platform-independent, enabling seamless deployment without cloud dependency or specialized hardware. Experimental evaluations demonstrate that the proposed system increases resistance to targeted attacks while maintaining authentication speed and user convenience. This work offers a practical and scalable enhancement to conventional authentication systems, particularly for mobile and resource-constrained environments.
Increasing demand for wireless sensor networks (WSNs) raises challenges in energy constraints and communication sustainability, especially in remote or disaster-affected areas where resources are limited. To address the constraints, this study proposes an integration of unmanned aerial vehicle (UAV)-assisted radio-frequency energy harvesting (RF-EH), non-orthogonal multiple access (NOMA), and device-to-device (D2D) communication to help with WSN energy constraints. The proposed system employs a UAV to transfer an RF signal for the WSN to perform EH, which will be used as a power source. The WSN operates in a D2D communication framework using NOMA communication. The system employs an EH time-switching (TS) protocol and dynamic power allocation to mitigate the impact of imperfect successive interference cancellation (SIC). The proposed system utilizes a machine learning-based k-nearest neighbor (kNN) algorithm to perform node selection at the UAV. Simulation results indicate that the kNN-based user selection improves the energy harvested by the WSN through the RF-EH process, which can reduce the BER by approximately 22.8%. Overall, the proposed system demonstrates superior performance by achieving lower bit error rate (BER) compared to conventional scenarios without the RF-EH process.
the context of future 6G networks, the integration of non-orthogonal multiple access (NOMA) with unmanned aerial vehicles (UAVs) is proposed to significantly enhance high-speed and reliable communication capabilities in emergency scenarios. This study addresses the challenge of maximizing the total sum rate for both downlinks and uplinks in multi-UAV Full-Duplex NOMA emergency networks through the joint optimization of subcarrier scheduling and power allocation. The inherent complexity of this problem arises from the nonconvex nature of the objective function, interdependence of product terms, and combinatorial characteristics of the decision variables. To overcome these challenges, we employ the Big-M method to decouple the product terms, relax the combinatorial subcarrier assignment into continuous variables, and introduce a Pp-norm-based penalty term in the objective function to penalize non-binary subcarrier assignments. Although these transformations simplify the problem, it remains nonconvex. To address this challenge, we propose an iterative refinement method with threshold-based rounding (IRM-TR) algorithm. The IRM-TR algorithm iteratively refines subcarrier assignments and power allocation by combining successive convex approximations with a dynamically updating weight tensor of the Pp norm to drive the subcarrier assignments toward binary values. A post-convergence threshold-based rounding step is also applied to ensure binary subcarrier assignments. The simulation results validate the proposed approach, demonstrating its effectiveness and efficiency in enhancing communication performance under emergency conditions.
(CF) massive multiple-input multiple-output (mMIMO) is emerging as a key technology for sixth-generation (6G) communication systems, offering nearly uniform service for users across various areas while effectively managing interference compared to traditional mMIMO systems. However, data detection in CF-mMIMO environments requires sophisticated signal processing techniques. While both linear and nonlinear detectors have demonstrated strong performance, the exploration of iterative detection methods in CF-mMIMO has been limited. This paper addresses this research gap by examining the performance of five efficient iterative scalable CFmMIMO detectors based on approximate/avoid matrix inversion techniques: Newton iteration, Gauss-Seidel, Jacobi, accelerated over-relaxation, and successive over-relaxation. Additionally, we propose an efficient detector based on sphere decoding (CF-SD) for scalable CF-mMIMO systems. Simulation results indicate that the linear iterative methods can achieve performance that approximates that of the minimum mean square error detector, while also maintaining a lower computational burden. In addition, while the CF-SD detector demonstrates considerable performance enhancements, it requires higher computational complexity compared to its linear iterative counterparts.
Edge computing and integrated sensing and communication (ISAC) technologies offer promising prospects for intelligent transportation systems (ITSs) in which the sensing data of vehicles can be processed directly or be offloaded to a base station (BS) or to the surrounding vehicles. However, the inherent scarcity of communication resources becomes a crucial problem in ITSs, especially when ISAC is introduced. In this paper, we propose an ISAC-assisted vehicular edge computing networks (VECNs) architecture composed of two interconnected stages: resource management and task offloading. Vehicles perform sensing and dynamically offload sensing tasks to the BS or nearby vehicles based on the link conditions. A two-stage joint optimization problem is formulated to optimize the resource block (RB) allocation for V2I and V2V links, including communication and sensing power among multiple vehicles, so as to maximize the overall data transmission rate. Concurrently, the offloading decisions are optimized, aiming to minimize the weighted sum of the system task completion delay and energy consumption. Considering the complex, dynamic transmission environment, we reformulate these problems as Markov Decision Processes and propose a deep reinforcement learning-based dual-stage resource management and offloading decision strategy (DDROS). Simulation results demonstrate that the proposed DDROS achieves strong convergence and exhibits significant performance advantages over baseline strategies under various conditions.
this paper, we present a novel approach to enhance the throughput of 6G non-terrestrial networks (NTN) by incorporating deep learning-based channel estimation, Doppler pre-compensation, and compensation techniques. We propose a new framework for accurate and efficient channel estimation in 6G-NTN systems, leveraging neural networks to improve channel estimation performance, leading to enhanced throughput and link tion and compensation techniques to address the challenges posed by high mobility scenarios in 6G-NTN. Extensive simulations demonstrate the effectiveness of our approach, showing significant improvements in mean squared error (MSE), throughput, and robustness to Doppler effects under high mobility scenario in NTN systems. The training data for the convolutional neural network (CNN) model, developed specifically for DM-RS channel estimation, demonstrates a MSE of 1.4175 at a transonic speed of 1,000 km/h and an altitude of 10 km in the NTN environment. The implementation of both Doppler pre-compensation and compensation techniques effectively neutralizes the Doppler shift. This results in a comparable bit error rate (BER) performance, achieving link reliability with a spectral efficiency of 3.325 bps/Hz at an NTN mobility of 1,000 km/h and an altitude of 10 km. The proposed framework has the potential to significantly impact the performance of 6G-NTN systems, paving the way for reliable and efficient wireless communication in challenging environments.
develop adaptive frequency block allocation schemes to mitigate the interference between intelligent low Earth orbit (LEO) satellites. As satellite networks attract increasing attention, the demand for limited frequency resources is expected to surge, creating a need for more efficient frequency utilization techniques. In particular, intelligent and dynamic frequency allocation methods will be more popular, which underscores the necessity for novel frequency resource allocation algorithms that take these considerations into account. In this work, we introduce two resource allocation strategies that exploit multi-agent reinforcement learning: the unmodified terrestrial-to-satellite (UTS) strategy that extends previous terrestrial method to the satellite environment, and the adapted satellite-specific (ASS) strategy that is tailored to satellite communication systems. Through simulations in both controlled and interference-prone environments, we evaluate and compare their performance, showing that, compared to the UTS strategy, the proposed ASS strategy improves throughput by up to 38% and reduces collision rate by up to 89% across different interference scenarios. Our findings highlight the effectiveness of customized resource allocation strategies in dynamic LEO satellite environments, paving the way for more efficient and scalable satellite communication systems in 6G networks.
The evolutions in communication technologies demand high-performance processing units and reliable back-hauling lines for the management of vast data in wireless networks. A reliable low-latency network is, therefore, essential for efficient data transfer, system maintenance, and information dissemination. This paper analyzes a backbone network system, for consideration in the real-time deployment and analysis of touch technology interfacing middleware networks. The proposed layer-wise network deployed using graph theory underscores an ultra-reliable, low-latency network design for optimal network performance. The algorithm selects symmetric or asymmetric deployed networks based on the topology and application requirements, ensuring minimum latency. The network optimizes throughput, latency, and data transfer for efficient connectivity between sources and destinations. It connects to controllers and edge devices at the user end, ensuring reliable data transfer and efficient communication. The computational time of the deployed network path between the source and destination end is evaluated and compared with popular algorithms, determining the computational complexity of the deployed network. Finally, the computational complexities between existing network approaches and the proposed deployed network are compared. This paper thus outlines optimal network design for touch technology systems in 6G.
