Mobile and IoT devices increasingly use Multipath TCP (MPTCP) to maximize throughput by leveraging multiple network interfaces. However, these throughput gains often come at the cost of significant energy drain, as naive multi-path scheduling can be highly inefficient. This paper addresses the challenge of energy-efficient MPTCP scheduling by introducing a novel distributed optimization approach. Unlike prior solutions that rely on centralized control, sacrifice performance, or require extensive training (e.g., reinforcement learning methods), we formulate the scheduling problem as a distributed constraint optimization problem (DCOP) and solve it using the Max-Sum message-passing algorithm. This is the first application of DCOP to MPTCP scheduling, enabling fully distributed optimization of energy use and throughput without centralized coordination or offline training. The proposed Max-Sum-based scheduler guarantees optimal decisions whenever the network graph is acyclic (42 × higher throughput than round-robin (and 50 × that of round-robin, ∼ 80 ∼ 25 × faster than the RL approach (400 vs. 10,000 episodes) with no training required. All improvements are statistically significant ( p < 0.001 , Cohen’s d > 3.0 ). These results demonstrate that energy efficiency and high performance need not be conflicting goals. By intelligently orchestrating multi-path transmissions via distributed optimization, the proposed solution offers a new paradigm that can extend device battery life and reduce network energy usage without sacrificing user experience.
This paper studies programmable-network transport scheduling for Multipath TCP (MPTCP) under overlap contention using LEO-inspired emulation conditions. We present GRLI, a controller-centric framework that combines in-band network telemetry, magnetic-attention graph encoding, and proximal policy optimization for joint path activation and rate allocation. The framework is implemented in a P4/Mininet prototype and evaluated against heuristic and learning-based baselines under homogeneous and heterogeneous contention settings. The results show that GRLI improves the throughput–latency tradeoff and operates within 4
Massive machine-type communications (mMTC) impose stringent requirements on uplink random access due to massive connectivity, limited uplink resources, heterogeneous service demands, and dynamic access conditions in dense deployment scenarios. Although existing studies have made important progress in improving access efficiency from different perspectives, two important limitations remain insufficiently addressed. First, differentiated access support for heterogeneous services is still limited, making it difficult to guarantee reliable access for traffic with different quality-of-service (QoS) requirements. Second, practical mMTC uplink access is governed by the coupled effects of channel conditions, traffic dynamics, and short-packet transmission, which jointly shape contention, reliability, and resource efficiency. Under these conditions, optimization from an isolated perspective is often inadequate. To address these issues, this paper proposes a priority-aware Q-learning-based uplink random access framework for non-orthogonal multiple access (NOMA)-based mMTC networks. The proposed framework incorporates service differentiation into the access process by introducing priority-aware resource reservation and a weighted reward design, while integrating NOMA transmission, short-packet communication (SPC), and Q-learning-based adaptation to improve uplink access efficiency and service reliability in dense and dynamic environments. Simulation results demonstrate that the proposed method achieves higher access success probability for medium- and high-priority devices while also improving system throughput and resource utilization under different traffic loads. Compared with conventional slotted ALOHA, traditional NOMA-based random access, and sparse code multiple access (SCMA)-based access schemes, the proposed method provides more stable access performance, especially for medium- and high-priority devices under dense mMTC traffic conditions.
Truth discovery is a key component of crowdsensing systems, where reliable truths must be inferred from heterogeneous and potentially unreliable worker reports. Super Fast Privacy-Preserving and Reliable Truth Discovery (FPTD)-style methods protect the downstream computation process, but they still rely on the quality of the worker-object matrix supplied to the protocol. When malicious uploads contaminate this matrix before protocol execution, privacy-preserving computation may faithfully process a biased input. This paper addresses this input-level vulnerability through a low-intrusion robust preprocessing framework placed before, rather than inside, the unchanged Original FPTD procedure. The framework prepares a cleaner matrix X through two data-type-aware branches. For sequential numerical reports, it combines robust current-round quality estimation, historical worker reputation, and probation-aware restriction to suppress persistently suspicious injected identities while avoiding premature removal of workers. For categorical reports, it introduces a compact collusion-aware gateway that uses cross-task worker agreement, group evidence, and participation masking to identify highly coordinated false-label collusion before FPTD execution. The proposed design keeps reputation, filtering, and support handling at the platform-side input-preparation layer, preserving the Original FPTD truth and worker-weight iterations. Experiments on Weather and Duck show that the preprocessing layer substantially mitigates the tested numerical injected-identity and categorical false-label-collusion attacks, recovers truth quality close to clean or high-accuracy settings, and maintains low intrusion into the privacy-preserving FPTD core.
Software-defined networking (SDN) is the promising network architecture of the near future, yet the network congestion in SDN forces a choice between processing load added to the controller and an efficient avoidance of the congestion. Most of the studies in this regard are based on the general-purpose devices of the data plane, which puts more processing load on controller. The controller’s operation is critical, the load on the controller must be optimized to affect only the most important network decisions. To resolve this, we leveraged the P4 programmable data plane to transfer a portion of the controller’s processing power to the data plane. In order to avoid congestion, leveraging the queueing theory and modeling the performance of the network nodes will aid us to mathematically measure each network node’s performance. Taking this into account, we leveraged the P4 programmable switches to gather the necessary information for queueing theory model instead of the controller. In this study we represent P4QT, with the aim of improvement on the network quality of service (QoS) and routing mechanisms through congestion-aware routing. We conducted this study with baseline implementations in order to analyze the effectiveness of our proposed method. The results show that P4QT was 94.9
The application of the Internet of Things (IoT) enables advancements in many areas worldwide by making everyday objects smart and creating intelligent spaces where users can interact seamlessly with IoT devices. These devices are equipped with sensors and applications that provide their functionality through various services. However, IoT devices are often characterized by mobility, low memory and computational capacity, and limited energy resources, which make the discovery process challenging. Traditional IoT discovery solutions are often inefficient, highlighting the need for dynamic resource management mechanisms. Although many architectures have been proposed, current service discovery approaches fail to account for node mobility, which can overload nodes and degrade network performance. In this paper, we provide a novel distributed-based mobility-aware service discovery mechanism using the Ant Colony Optimization algorithm, named MDSDA, for global search that helps mitigate flooding, thereby reducing the risk of overloading the network and minimizing the associated increase in network load. Simulation results clearly demonstrate that the proposed solution significantly enhances the discovery process, achieving a high query resolution efficiency, reducing network load by 40.5
Backscatter communication technology is used for optimizing energy efficiency in wireless communication networks. It leverages existing signals in the environment rather than generating new ones, allowing devices to communicate with minimal power consumption. Energy efficiency of wireless networks is a key goal in futuristic communication systems, the growth of IoT, and 5G beyond. Various methods have been proposed to enhance energy efficiency in wireless networks, focusing on reducing energy consumption while maintaining network performance. In this paper, the suggested technique combines energy harvesting with non-orthogonal multiple access technology to achieve optimal energy efficiency in wireless networks. The proposed technique is compared with the orthogonal multiple access technology of the network. Results show the efficacy of the proposed technique compared to the state-of-the-art approaches in terms of energy efficiency, distance of backscatter, and power splitting ratio.
Web Authentication (WebAuthn), standardized by the World Wide Web Consortium (W3C) and supported by major platforms, aims to eliminate password-based vulnerabilities by enabling cryptographic, passwordless authentication. However, its security model implicitly trusts the client device, exposing it to threats from local malware. This paper presents a Trojan-based proof-of-concept attack that compromises WebAuthn sessions in real-world settings. Our experiments achieved a 100
The growth of WiFi-enabled smart devices and high-bandwidth applications has generated a huge demand for intelligent wireless network management mechanisms that can offer Quality of Service (QoS), seamless mobility and resource efficiency. The conventional WiFi association mechanisms are primarily based on the Received Signal Strength Indicator (RSSI) selection, leading to uneven utilization of Access Points (AP), congestion, excessive handovers and degraded network performance in dense wireless environments. To overcome these challenges, in this paper we propose a multi-objective AP load balancing and handover scheme based on Software-Defined WiFi load balancing (SDW-LB) mechanism for the overlapping wireless coverage regions. The proposed framework leverages the centralized intelligence and global network view of Software Defined Networking (SDN) for dynamic management of user association and handover decision. When the user enters the overlapped area of multiple APs, the SDN controller will initiate an intelligent handover process based on a multi-objective optimization model in terms of RSSI strength, AP load, number of connected users, available bandwidth, channel utilization, packet delay, user mobility and throughput conditions. A weighted utility-based objective function is presented to select the most suitable AP while reducing network congestion and superfluous handovers. Moreover, a centralized SDN controller i.e., Open Network Operating System (ONOS), constantly observes the network environment and makes dynamic adjustments to the AP association decisions for the purpose of balanced resource allocation and enhanced QoS. The proposed scheme can enhance the network throughput, reduce the packet loss, relieve the AP congestion, and improve the user experience in dense WiFi deployments. The experimental results of our proposed approach show the effectiveness of the proposed scheme compared with benchmark methods with respect to QoS metrics such as fairness, delay, throughput, and packet loss ratio.
The rapid evolution of sixth-generation (6G) wireless networks demands antenna systems capable of adapting to highly dynamic communication environments characterized by user mobility, severe signal blockage, complex propagation conditions, and increasingly heterogeneous service requirements. Conventional fixed antenna architectures are unable to provide the flexibility, spatial adaptability, and intelligence required to support these emerging scenarios. Consequently, adaptive antenna technologies (AATs), including reconfigurable antennas, beam-steerable antennas, fluid antenna systems, movable and rotatable antennas, and artificial intelligence (AI)-enabled adaptive antennas, have attracted significant research attention. Unlike existing surveys that primarily focus on individual AATs, this review presents a unified framework that integrates electromagnetic reconfiguration, spatial mobility, and AI-native intelligence within a single multidimensional taxonomy. The proposed taxonomy systematically classifies AATs according to hardware architecture, reconfiguration mechanism, intelligence level, and spatial adaptability, while providing a comprehensive comparative analysis of their operating principles, enabling technologies, advantages, limitations, and representative applications. Furthermore, the review examines key implementation challenges associated with energy efficiency, security and privacy, terahertz (THz) communications, standardization, interoperability, and compatibility with existing fifth-generation (5G) infrastructure, together with a phased deployment roadmap toward practical AI-native 6G networks. Emerging research directions, including AI-native antenna ecosystems, digital twin-assisted optimization, reconfigurable intelligent surface-integrated adaptive antennas, semantic-aware communications, integrated sensing and communication, and spatial intelligence networks, are also discussed. The findings demonstrate that the future evolution of AATs will rely on the synergistic integration of advanced electromagnetic design, spatial reconfiguration, and AI-driven optimization, providing a comprehensive reference and practical roadmap for researchers and engineers developing intelligent, autonomous, and sustainable 6G wireless communication systems.
This study introduces a multi-objective clustering framework for UAV networks derived from the Secretary Bird Optimization Algorithm (SBOA). The proposed approach incorporates adaptive search behavior to balance exploration and exploitation during cluster-head (CH) selection in three-dimensional Flying Ad Hoc Networks (FANETs). To increase network scalability, stability, and energy-efficient routing, the proposed model jointly considers residual energy, intra-cluster separation, communication limitations, and load distribution. SBOA differs from existing bio-inspired methods such as Fire Hawk Optimization Algorithm (FHOA), Portia Spider Optimization Algorithm (PSOA), multi-objective sperm fertilization procedure-based optimization (MOSFP), and Draco lizard optimizer (DLO). SBOA employs a biologically motivated diversification strategy that preserves population variability, reduces the risk of premature convergence, and responds effectively to topology changes in dense UAV environments. Large-scale simulations were conducted on the population of UAVs (60–120 based) and the communication range of 100–900 m in large 3D grids. Findings indicate that SBOA always excels in benchmark algorithms on all significant FANET performance metrics. SBOA can maximize the optimization fitness by 16
Device-to-device (D2D) communication is one among awful notable technologies of fifth-generation (5G) networks that uplift the network’s reliability and spectral efficiency. Interference management and sum rate maximization are the key aspects in D2D communication networks. Various researchers have given solutions to resolve the aforementioned issues based on convex optimization, machine learning, deep learning, game theory, and graph theory. State-of-the-art approaches are managed to mitigate interference and sum rate maximization. But, time complexity and the number of iterations taken by algorithms are not significantly optimized till now for the state-of-the-art approaches. So, motivated by these, this paper proposes a maximum matching algorithm for channel allocation with a faster convergence rate. The proposed approach divides the entire set of cellular users (CUs) and D2D groups into overlapping clusters based on channel gains. After clustering, the proposed approach utilizes the Kuhn-Munkres algorithm for the best channel allocation to the D2D groups within the same cluster. The performance of the proposed clustering-based matching approach is compared with the baseline approach considering sum rate and secrecy capacity parameters and found superior with fixed and varying numbers of CUs and D2D groups.
In this article, we examine the performance of an intelligent reflecting surface (IRS)-assisted single-input multiple-output (SIMO) wireless communication system over Rician fading channels. The system comprises a single-antenna transmitter, an IRS, and a multi-antenna receiver employing maximal ratio combining to maximize the desired signal gain and effectively combine the multi-path components to enhance signal quality. We derive novel closed-form expressions for the average symbol error probability (ASEP) for both rectangular quadrature amplitude modulation (QAM) and cross-QAM schemes using a computationally efficient moment-generating function-based approach. Further, we provide asymptotic ASEP expressions, revealing the system’s diversity order as a function of the number of IRS elements and receiving antennas. Our analysis examines the impact of key parameters such as the number of reflecting elements, receiving antennas, and fading severity, on ASEP performance. Results indicate that increasing the IRS elements and diversity branches improves ASEP performance, while severe fading degrades performance. Monte Carlo simulations validate the accuracy of our derived analytical expressions as well as the asymptotic expressions.
Electrocardiogram (ECG) signals have emerged as a promising biometric modality owing to their inherent uniqueness, temporal stability, and intrinsic liveness characteristics. However, the increasing use of biometric data in authentication systems raises critical concerns regarding privacy, secure storage, revocability, and the ethical handling of sensitive physiological information. To address these challenges, this paper proposes a privacy-preserving ECG-based biometric authentication method based on cancelable biometric template generation. The proposed approach employs the Lorenz chaotic system, a well-established nonlinear dynamic model, to mask the original ECG features and generate protected templates. Subject-specific morphological features are extracted from ECG segments centered around the R-peaks within the QRS complex and are subsequently transformed using a Lorenz-based chaotic masking mechanism. This transformation enhances template privacy and revocability while preserving the discriminative characteristics required for reliable authentication. Experimental results obtained on two widely used ECG datasets, ECG-ID and MIT-BIH, demonstrate that the proposed method effectively protects user privacy without degrading recognition performance, achieving an accuracy of up to 99.5
Wireless Sensor Networks (WSNs) have emerged as a focal point of research and practical application in the field of computer networks and telecommunications. One of the foremost challenges in WSNs is the development of precise and efficient localization methods. In this research, we propose an enhanced DV-Hop localization algorithm. First, the neural network approach is applied to enhance localization accuracy by minimizing distance errors. The work is extended to cover energy efficiency by using optimization techniques for WSN-enabled DV-Hop. To solve the non-convex optimization for the WSN-enabled DV-Hop scheme, Alternating Direction Method of Multipliers (ADMM) is employed, which decomposes the global optimization task into three coordinated sub-problems—local variable update, global variable consensus, and dual variable update, allowing for efficient distributed computation. Results show that there is a 25.9
The rapid urbanization and the increased need for sustainability have placed the Green Internet of Things (G-IoT) at the center as the key driver of green smart cities. In this paper, a systematic literature review (SLR) of G-IoT technologies and strategies applied in urban settings is provided following PRISMA guidelines. Integrating 130 studies from 2015 to 2025, the review consolidates findings across three interconnected areas: energy-saving technology, communication protocols, and security concerns with corresponding countermeasures. Structuring these findings, the study proposes a three-category technology model and a three-layer security framework. The review highlights the most recent advances, trade-offs, and contextual suitability of G-IoT solutions while identifying gaps that require further investigation. With the merging of these perspectives, the review presents a comprehensive understanding of G-IoT's role in improving energy efficiency, adaptability, and sustainability in smart cities as well as outlining avenues for future research towards achieving greener, adaptive urban environments.
Device-to-device (D2D) communication offers an infrastructure-independent solution for post-disaster connectivity, but conventional clustering methods often suffer from load imbalance, premature energy depletion, and unstable connectivity. This paper proposes an Adaptive Fuzzy–PSO Unequal Clustering (AFPUC) framework that integrates fuzzy-logic-based cluster-head selection, feedback-driven reclustering, event-triggered PSO reselection, and energy-aware multi-hop routing within a closed-loop adaptive framework. Unlike conventional approaches that treat clustering and adaptation separately, AFPUC jointly optimises resilience, load balancing, and energy sustainability. Simulation results under sparse and dense deployments, benchmarked against six state-of-the-art clustering algorithms, show that AFPUC achieves up to 89.5
In the Cloud, intrusion detection is vital for analyzing and monitoring system activities and network traffic within cloud environments for detecting malicious behavior. Nonetheless, the prevailing detection of cloud intrusion models faces challenges like detecting sophisticated attacks, handling large-scale data, and ensuring low false positives. Accordingly, a new approach, namely, Convolutional Neural Network-based Transfer Learning with Ablation Frigate-Fairy Hybrid Optimization (CNN_TL_AFFHO), is presented for detecting cloud intrusion in Federated Learning (FL). FL protects privacy by storing all local logs on individual client devices and transmitting only model updates to the central server, allowing multiple parties to train a model together without revealing private data. The FL framework has a server and several local models. In the local model, Weitendorf’s Linear (WL) normalization is applied to normalize the log data. This method is applied to standardize log features at local clients. Then, key features are selected utilizing AFFHO, which is designed by combining Superb Fairy-wren Optimization Algorithm (SFOA) and Frigate Ablation Optimization Algorithm (FAOA). N4.1.2.4ext, the Synthetic Minority Over-sampling Technique (SMOTE) is used to achieve data augmentation. The SMOTE is used to augment minority attack samples and address class imbalance. Finally, intrusion detection is performed using CNN_TL, which utilizes hyperparameters derived from the trained Deep Xception convolutional Forward Harmonic Network (DXcov-FH Net). The DXcov-FH Net is developed by DSA, harmonic analysis, and XCovNet. The DXcov-FH Net is used to extract rich hierarchical features for intrusion detection. Lastly, local aggregation and updating at the server are established by averaging. The proposed model enables improved privacy, efficient feature selection for reduced false positives, and scalable deployment across distributed cloud nodes. Here, the Bot-IoT dataset and Network Intrusion Detection dataset are used for the evaluation of the devised model. Moreover, proposed CNN_TL_AFFHO achieves a better accuracy of 97.89
This paper investigates the integration of Simultaneous Transmitting and Reflecting Reconfigurable Intelligent Surfaces (STAR-RIS) with adaptive transmit power control and magnetic energy harvesting in Underlay Cognitive Radio Networks (UCRN). The proposed framework aims to enhance spectrum utilization and energy efficiency by enabling secondary users to coexist with primary networks without violating interference constraints. In the considered system, the STAR-RIS assists signal propagation by dynamically adjusting transmission and reflection coefficients, while the secondary transmitter adapts its power based on instantaneous channel and interference conditions. Moreover, magnetic induction-based energy harvesting is employed to provide sustainable energy for the secondary network devices. Simulation results demonstrate that the proposed STAR-RIS-aided UCRN significantly improves throughput and energy efficiency compared to conventional non RIS-assisted schemes.
In this paper, we present the design, implementation, and experimental validation of a novel low-power RF switch using varicap diodes, especially optimized for the 868 MHz ISM band and integrated on a LoRa-based IoT node. Unlike typical RF switches, which frequently use PIN diodes or Micro-Electro-Mechanical Systems (MEMS) technology, the proposed design uses the tunable capacitance of varicap diodes to achieve a compact, low-bias, and energy-efficient solution. Circuit modelling, PCB layout, and biasing approach are all part of a full design process. A prototype was made and tested to show that it had moderate insertion loss, good isolation, and very low-power consumption, which is strongly needed for LoRa-based IoT nodes. We present a proof-of-concept implementation in which the switch serves to physically manage the antenna connection. In this setup, the antenna is disconnected from the LoRa transceiver, and it is only activated during the LoRaWAN transmission and reception windows. This approach reduces RF noise exposure from idle sources, provides an additional layer of hardware security during reprogramming or maintenance, and increases resistance against powerful adjacent signals.
