Complex networks have a large number of nodes and edges, which prevents the understanding of network structure and the discovery of valid information. Community detection is an important issue in studying network structure and network characteristics. It has received widespread attention in many fields. Most existing community detection algorithms obtain the final community structure by analyzing the relationship between each node and surrounding nodes. Starting from a portion of nodes in each community, the corresponding community for each node can be obtained through expansion operations, thereby obtaining the entire community structure. Such strategy can improve the accuracy of community detection algorithms. When solving large- scale combinatorial optimization problems, the traditional ant colony algorithm has a slow convergence rate and tends to fall into local optima. More and more scholars propose relevant optimization algorithms on the basis of classical ant colony algorithm. To overcome premature convergence, adaptively adjusted the pheromone on the path according to the existing solution, which enabled it to escape the local optimal value. To address these challenges, this research proposes an improved optimization method. This approach integrates community detection, multi-group cooperation, pheromone feedback mechanisms and Hybrid Dynamic Pheromone Updating Mechanism to improve exploration efficiency and convergence speed in large-scale TSP problems.
The unique properties of terahertz (THz) waves were examined, particularly in contrast to lower-frequency bands, highlighting the effects of weather conditions and material surface roughness on THz signal propagation. To address modeling challenges, the study introduced a ray-tracing (RT) approach to refine the 3D environmental model and material electromagnetic properties using minimal channel measurement data. These calibrated parameters were then applied to broader scenarios, reducing reliance on extensive measurements. Key channel characteristics—such as path loss, shadow fading, Rician K- factor, delay spread, angular spread, and Doppler effects in mobile settings—were analyzed. The research explored two distinct 6G THz use cases: indoor desktop wireless links and outdoor vehicular communications, with the latter incorporating weather-related impacts. These findings offer valuable insights for designing and assessing THz communication systems
The rapid adoption of Software Defined Networking and Network Function Virtualization architectures has transformed modern communication infrastructure by introducing flexible and programmable network control. However, the same programmability has increased the exposure of these environments to sophisticated cyber threats, particularly the zero-day attacks that exploit previously unknown vulnerabilities. Traditional intrusion detection mechanisms rely heavily on signature-based or supervised learning models that require labelled attack data. Such approaches have limited capability when the network encounters unseen attack patterns. Consequently, an effective anomaly detection framework that can operate without extensive labelled datasets has become an important research requirement. This study has proposed a Self-Supervised Network Anomaly Representation Model (SS-NARM) for detecting zero-day attacks within SDN/NFV environments. The proposed approach has utilized self-supervised representation learning that has extracted latent behavioural patterns from network traffic without the need for manual annotation. The architecture has integrated a feature encoder that has learned intrinsic traffic characteristics and a contrastive learning module that has maximized the similarity between semantically related network flows while separating anomalous behaviour. During the training phase, the model has generated pseudo-labels from intrinsic traffic patterns, which have guided the representation learning process. The anomaly scoring mechanism has evaluated deviations between learned normal traffic embeddings and real-time observations within the SDN controller monitoring layer. The experimental evaluation demonstrates that the proposed SS-NARM framework significantly improves the detection capability for zero-day attacks in SDN/NFV environments. The model achieves 96.8% detection accuracy, 95.4% precision, 94.9% recall, and 95.1% F1-score, while achieving an AUC value of 0.98 that reflects strong discrimination capability between normal and malicious traffic flows.
Personalized speech enhancement (PSE) is a speech enhancement method to remove interfering speech, background noise, and reverberation based on a speaker embedding extracted from the target speaker such as d-vector and x-vector. In full duplex communication scenarios, when the microphone and far-end signal are coexisted together, it creates acoustic echoes. This echo is one of the major factors to the degradation of the sound quality of online communication systems, including video conferencing. Hence, Acoustic Echo Cancellation (AEC), a technique that can effectively remove these acoustic echoes, has been investigated. For full-duplex communications, which acoustic echoes are exist with background noises and interfering speech together, AEC and PSE must be combined. We study this combination. Our goal is to develop a causal model that can be applied to various model architectures to efficiently handle the tasks of AEC, PSE, and joint AEC-PSE. The features are extracted from the far-end signal and the near-end signal. The cross- attention alignment mechanism is used for feature alignment of the far-end signal and x-vectors are used as speaker embedding features. The proposed method is applied to PSE models such as E3Net and VoiceFilter-Lite. We present extensive experimental results. We demonstrate the effectiveness of the proposed method through the experiments in terms of various evaluation metrics with several standard audio and real recording datasets.
The rapid expansion of wireless communication systems has intensified the challenge of co-channel interference, which has significantly affected signal reliability and spectral efficiency. In dense communication environments, multiple transmitters often have shared the same frequency band, which has created overlapping signals at the receiver. Traditional blind source separation techniques, which have relied on statistical independence assumptions or matrix factorization strategies, have faced limitations when signals have exhibited complex temporal correlations and nonlinear distortions. These limitations have motivated the need for adaptive learning models that have captured deeper signal representations without extensive labeled datasets. This study has addressed the problem of separating mixed communication signals under severe co-channel interference conditions. Conventional supervised learning frameworks have required labeled mixtures and ground truth signals, which have remained difficult to obtain in real wireless deployments. As a result, a robust representation learning strategy has become essential for extracting meaningful signal structures from unlabeled observations. To overcome this limitation, the study has proposed a Self-Supervised Contrastive Representation Separation Network (SCRSN), which has utilized self-supervised representation learning for blind source separation. The proposed method has learned latent signal embeddings through a contrastive objective that has encouraged the model to distinguish between temporally consistent signal patterns and unrelated interference components. An encoder–decoder architecture has extracted hierarchical signal features, while a clustering-based separation module has reconstructed the independent source signals. The model has leveraged signal augmentation strategies that have generated positive and negative sample pairs without manual labeling, which has enabled efficient representation learning from raw signal mixtures. The experimental evaluation demonstrates that the proposed SCRSN framework achieves 94.1% signal separation accuracy at 20 dB SNR, which exceeds the performance of the ICA, NMF, and Deep Autoencoder BSS approaches. The method produces 23.4 dB Signal- to-Interference Ratio and 22.4 dB Signal-to-Distortion Ratio, which indicate strong interference suppression and signal reconstruction capability. The framework also reduces the reconstruction error to 0.021 Mean Squared Error, while maintaining an efficient computational time of 11.7 seconds for large signal inputs.
Antenna beamforming has emerged as a promising solution for the advanced wireless communication systems particularly where large number of antennas is used at the transmitter and receiver (multiple input multiple output-MIMO) systems. The beamforming techniques MIMO system provides improved signal quality, enhances capacity and reduces interference. Telecommunication service providers need to address several challenges for implementation of the beamforming techniques because of its different aspects; beamforming for suitable applications, complexity, implementation cost, power consumption, capacity etc. This paper provides a comprehensive analysis of three types of beamforming techniques (analog, digital and hybrid) over various fading channels and antenna configurations. The analysis explores effect of fading conditions on the performance of the beamforming techniques and crucial parameters such as spectral efficiency (SE) and energy efficiency (EE). Further paper investigates the influence of multiple antenna configurations. While adding more antennas gradually improves spectral efficiency by leveraging spatial diversity, but it introduces increased power consumption, leading to diminishing returns in energy efficiency. The simulations are also carried out with different channel models such as Rician, Rayleigh and Nakagami. Moving from Rayleigh to Rician (K=5) channel model, boosts spectral efficiency of hybrid beamforming by approximately 11% at 10 dB signal to noise ratio. These results validate hybrid beamforming architecture as a pivotal technology for future wireless systems, enabling the deployment of massive MIMO to meet escalating data demands in a power-efficient and cost-effective manner.
The rapid expansion of heterogeneous services in sixth generation (6G) communication networks has increased the complexity of resource orchestration within the network core. Emerging applications such as autonomous systems, immersive communication, and large-scale Internet of Things environments have required highly flexible and efficient resource slicing mechanisms. Conventional resource allocation techniques have relied on static or semi-dynamic policies that have limited adaptability to fluctuating traffic patterns and diverse quality of service requirements. As the network scale has grown and service diversity has intensified, these approaches have faced challenges in maintaining efficient utilization and service reliability. Consequently, the dynamic management of network resources has remained a critical issue in the evolving 6G infrastructure. This study has investigated a dynamic resource slicing mechanism that has utilized Multi-Agent Reinforcement Learning based Adaptive Resource Slicing (MARL-ARS) for the 6G network core environment. The proposed framework has introduced multiple intelligent agents that have interacted with the network environment and that have cooperatively optimized the allocation of bandwidth, computational capacity, and storage resources across different network slices. Each agent has learned an optimal allocation policy through continuous interaction with the system state, while the cooperative learning structure has enabled coordinated decision making among distributed agents. The reinforcement learning mechanism has incorporated reward optimization strategies that have considered network latency, resource utilization efficiency, and service reliability. Through iterative learning, the model has gradually refined its slicing policies and has achieved adaptive resource allocation under varying traffic loads and service demands. The experimental results demonstrate that the proposed MARL-DRS framework significantly improves the performance of dynamic resource slicing in the 6G network core. The system achieves 93% resource utilization under high network load conditions, while the baseline approaches achieve between 78% and 85% utilization. The proposed model also improves the network throughput to 8.6 Gbps, which exceeds the existing approaches that achieve 6.6–7.5 Gbps. The slice allocation accuracy reaches 94% after 35 training episodes, which indicates that the cooperative learning agents effectively interpret the network state and allocate resources accordingly. In addition, the framework reduces the network latency to 35 ms under heavy traffic conditions and maintains a 96% QoS satisfaction rate across heterogeneous service slices.
The deployment of agentic artificial intelligence (AI) applications across heterogeneous multi-cloud environments demands orchestration mechanisms that ensure both efficiency and privacy. This paper presents a framework that integrates adaptive scheduling with privacy-preserving techniques, including homomorphic encryption, trusted execution environments and differential privacy. The objective is to investigate how such integration can achieve scalable and low- latency execution without incurring prohibitive overhead. The proposed approach is implemented and evaluated against state-of-the- art orchestrators such as Docker Compose, Kubernetes and Karmada. Experimental results demonstrate that the framework sustains high throughput and efficient resource utilization while introducing only modest privacy-related overhead. These findings confirm the feasibility of embedding strong privacy guarantees into real-time multi-cloud orchestration for data-intensive AI workloads.
In the present wireless scenario, nonorthogonal multiple access (NOMA) and multiple-input multiple-output (MIMO) aims to achieve substantially improved spectrum efficiency and high performance. In this paper, we analyze the outage and sum-rate performance of MIMO- NOMA using fair and improved fair power allocation (PA) and compare it with MIMO-orthogonal multiple access (OMA). Also, we evaluate system performance over Nakagami-m fading scenario. The weak user is given priority while computing the power coefficient for fair PA. To meet the weak user’s target rate, the PA coefficients are computed. We observe that the outage probability steadily increases with the increase in weak user’s target rate in fair PA. The likelihood of a weak user reaching the target rate decreases as the target rate rises. Moreover, we make minor alterations to improve efficacy and minimize outages for strong users using improved fair PA. Simulation results show that MIMO-NOMA delivers lesser outages and larger sum rate for both the users using improved PA than MIMO-OMA in Nakagami- m fading.
Data center placement in a network plays a vital role for different online applications like VoIP, cloud computing, etc. However, disasters can affect their functionality leading to huge disruption in service. Not only this, network load balancing is another major concern nowadays, the improper management of which can hamper the network throughput and quality of service. In this paper, a new routing, spectrum and core allocation (RSCA) heuristic has been developed to balance the network load by imposing labels to the data centers based on their usage in dynamic space division multiplexing-based elastic optical network (SDM-EON). In this context, two data center selection strategies are introduced which are tested and analysed on two well- known topologies against different parameters, proving their efficacy over each other.
The most challenging problem of video conferencing systems is the degradation of sound quality due to various noise sources. Speech enhancement includes the reduction of background, acoustic echo cancellation, and dereverberation. A number of studies have been carried out to remove acoustic echo and background noise in video conferencing systems, and recently, DNN approaches have been applied to speech processing based on classical digital signal processing techniques, leading to great progress. We first propose a multi-input deep complex recurrent network (MIDCCRN) for noise suppression. Then, we propose a model for joint acoustic echo cancellation and background noise suppression in online voice communication systems, including video conferencing systems, using this network. The best performance of the proposed method is demonstrated by experiments with objective metrics including echo return loss enhancement (ERLE), signal-to-artifacts-ratio (SAR) and scale-invariant source-to-noise ratio (SI-SNR), mean opinion score (MOS) as a subjective metric, and AECMOS, real time factor (RTF), network size, and final score.
The convergence of sensing and automation technologies in agriculture has arisen as a promising means to counteract issues of water shortage, crop protection, and green farming. This work introduces the development of an embedded multi-sensor platform intended for precision farming purposes, with special emphasis on intelligent irrigation, greenhouse temperature control, and animal intrusion alert. The system utilizes a soil moisture sensor to monitor soil conditions and control an automatic water pump, irrigating only when the moisture percentage drops below 30%. The system includes a temperature sensor to control the microclimate plant, whereby a servo-controlled ventilation or shade cover adjusts whenever the temperature rises above 20 °C. In addition, the platform uses an ultrasonic sensor to sense object or animal approach towards the crop field, thus offering real-time protection against threats. Two LCD displays are used for ongoing observation of soil temperature, moisture levels, and object detection distance, providing user-friendly feedback to farmers. With these subsystems, the platform minimizes wastage of water, avoids heat stress in crops, and improves crop protection. The proposed system identifies the advantage of embedded control and multi-sensor integration in developing cost-effective, scalable, and farmer-friendly smart agriculture solutions that can aid sustainable and technology-enabled farming practices.
Wireless Sensor Networks (WSNs) are mainly used for continuous monitoring and reliable data transmission are essential. However, the limited battery capacity of sensor nodes poses significant challenges to long-term network operation. Clustering is an effective strategy to reduce communication overhead, but selecting an optimal Cluster Head (CH) remains a complex task due to varying node energy, distance, and network conditions. This study proposes a hybrid Machine Learning–Firefly Optimization–based Cluster Head selection (ML–FOA–CH) approach that combines predictive fitness evaluation with metaheuristic optimization. Machine learning models assess node suitability using key features, while FOA refines the search by maximizing brightness values. Experimental results show that ML–FOA–CH significantly improves CH selection accuracy, prolongs network lifetime, and delays the first node death compared to LEGN, TEGN, and traditional FOA-based methods. The proposed model demonstrates superior adaptability and energy efficiency, making it a promising solution for intelligent and sustainable WSN operations.
Group travel, including college trips, trekking trips, family trips, or bike rallies, often faces the challenge of keeping everyone together. When participants get separated, existing tools like Google Maps live sharing or WhatsApp live location only show individual positions, making group coordination difficult. This can lead to confusion, wasted time, and even safety concerns. To address this, we introduce SyncFleet, a lightweight real-time coordination system built specifically for group travel. Instead of tracking people one by one, SyncFleet displays all members on a single shared map and provides instant alerts for events such as stops, delays, or breakdowns. A modern stack powers the system: React for the interface, Node.js with Express for the back-end, WebSockets for live updates, and MongoDB for storage. Experimental simulations demonstrate that SyncFleet reduces coordination time By improving safety, enhancing efficiency, and keeping groups synchronized, it shows promise not just for trips but also for treks, rallies, marathons, and even emergency response situations.
The rapid growth of smart-city infrastructures has created an environment in which massive IoT deployments operated across dense, heterogeneous wireless networks. As device density increased, the communication channels have often experienced severe interference, unpredictable fading, and high noise levels that collectively limited estimation accuracy. Traditional estimation techniques relied on linear models that struggled to track the dynamic channel conditions of large-scale IoT environments. This scenario established the core problem: existing estimators have not maintained reliable performance when network density surged or when devices transmitted sporadic traffic. To address this, the study proposed an AI-driven channel estimation framework that has leveraged deep learning to extract latent channel characteristics from limited pilot signals. The method incorporated a hybrid convolutional–recurrent design that captured spatial variations while it tracked temporal fluctuations of each channel. The system also included an adaptive refinement block that has improved estimation accuracy when pilot contamination occurred. The architecture was trained with synthetic and real-world datasets that have represented typical smart-city IoT deployments, including traffic sensors, utility meters, and environmental monitoring nodes that operated under mixed mobility patterns. The evaluation demonstrates that the proposed framework consistently outperforms conventional estimators. The method achieves a 6.2% NMSE at 100 epochs compared with 10.4% for MMSE and 8.2% for CS, and reduces MAE to 4.0% compared with 7.2% for MMSE. Spectral efficiency increases to 6.9 bps/Hz, while pilot overhead is reduced by 25%, outperforming baseline methods. Computational time remains practical at 3.6 ms per batch, confirming that the AI-assisted estimation effectively enhances reliability and efficiency in large IoT smart-city deployments.
Wireless Sensor Networks (WSNs) have expanded substantial attention owing to their wide variety of applications in various fields. However, energy consumption remains a critical challenge in WSNs, as the nodes are typically powered by limited battery resources. This paper addresses the energy consumption problem in WSNs by proposing a novel approach that combines the Hybrid Firefly Glow-Worm Swarm Optimization (HF-GSO) algorithm, Dynamic Voltage and Frequency Scaling (DVFS) algorithm, and the duty cycling technique. The HF-GSO algorithm stands employed for the selection of effective cluster heads and routing in WSNs. It leverages the collective behavior of fireflies and glow-worms to achieve optimal energy utilization and network performance. By incorporating HF-GSO, the proposed approach optimizes the formation of clusters, minimizing the energy consumption associated with long-distance communication and data aggregation. Additionally, the DVFS algorithm is integrated into the system to energetically regulate the voltage and frequency levels of sensor nodes. This adaptive scaling mechanism allows the nodes to operate at lower power levels during periods of low activity, effectively reducing energy wastage. The DVFS algorithm further contributes to energy efficiency without compromising the network’s overall performance by scaling up the voltage and frequency only when necessary. Furthermore, the proposed approach utilizes duty cycling, a technique that enables the nodes to alternate between active and sleep modes. By effectively scheduling the node’s active and sleep durations, duty cycling significantly reduces idle listening and idle transmission, minimizing unnecessary energy consumption. The usefulness of the proposed method is demonstrated through extensive simulations and performance evaluations. The results indicate notable improvements in energy efficiency, network lifetime, and overall system performance compared to existing approaches. In conclusion, this research paper gives a complete solution to the energy consumption problem in WSNs. By integrating the HF-GSO algorithm, DVFS algorithm, and duty cycling, the proposed approach achieves significant energy savings and extends the lifetime of WSNs, making it highly suitable for energy-constrained WSN applications.
The rapid integration of wireless sensor networks in healthcare monitoring has created strong opportunities for continuous patient assessment. However, the distributed nature of these networks has exposed sensitive medical data to significant privacy and security risks. Traditional centralized learning models have struggled to protect patient information, particularly when the data has/have been transmitted across heterogeneous devices. This study addressed these concerns by evaluating an enhanced secure federated learning framework that has/have reduced communication overhead and strengthened protection against model-level threats. The problem emerged when conventional federated models failed to defend aggregated parameters against inference attacks that targeted the intermediates shared during training. To overcome this limitation, the proposed system integrated authenticated encryption, differential privacy, and a lightweight blockchain layer that/which supported tamper-proof logging. The method followed a three-stage design that/which included secure client selection, privacy-preserved gradient update, and decentralized model validation. The wireless nodes operated with an adaptive update schedule that/which minimized energy use while maintaining stable model convergence. The evaluation demonstrates that the proposed secure federated learning framework achieves a classification accuracy of 96.0%, outperforming Encrypted Aggregation FL (93.0%), Differential Privacy FL (90.2%), and Blockchain-Assisted FL (94.2%). The communication cost has/have been reduced to 17.2 MB from 22.0 MB, 18.1 MB, and 23.5 MB, respectively. Energy consumption per node is lowered to 1.95 J, compared to 2.45 J, 2.68 J, and 2.63 J in the existing methods. The system achieves a privacy preservation score of 0.94, higher than 0.75– 0.87 in baseline approaches, and maintains strong model robustness at 94.2% under adversarial conditions. These results validate that the proposed framework provides reliable, energy-efficient, and secure federated learning suitable for real-time healthcare monitoring applications.
In this paper, we propose a method to solve the instantaneous frequency estimation problem at each point of the digital signal sequence obtained by digitizing the chirped signal by combining the projection method with the stochastic resonance method. This problem is chosen as the estimation of the instantaneous frequency (instantaneous frequency at the midpoint of a frame) at the center of the frame, and then we can solve the above problem through the orthogonal projection of a space of which dimension is equal to the frame length into a two-dimensional subspace, combined with stochastic resonance theory. Assuming this digital signal frame as a vector lying in a space of which dimension is equal to the length of the frame, we estimate the instantaneous frequency of the frame center point by finding the basis vector of the corresponding frequency that constitutes the two-dimensional subspace in which the vector is placed. By moving the center point of the frame onto each point of the digital signal sequence corresponding to a period of frequency modulation, we can obtain the overall frequency curve accurately. At this time, the basis vectors that constitute the two-dimensional subspace are constructed by reflecting the frequency modulation characteristics. The estimation results obtained in the simulation are compared with the results using the short-time Fourier transform, Wigner-Bill distribution, which are often used in the presence of noise and Doppler effects such as Doppler radar and sonar, which shows that the proposed method has very high accuracy.
The rapid growth of next-generation networks has created a strong demand for communication systems that have delivered high reliability, low latency, and resilience under harsh channel conditions. Although classical error correction codes have improved many wireless links, their performance has proved insufficient as data rates increased and channel dynamics became more unpredictable. This study has explored a quantum-inspired error correction framework that has combined structural principles from quantum stabilizer codes with the efficiency of classical block codes. The aim was to provide an adaptive mechanism that has reduced noise effects and supported ultra-reliable communication targets. The problem has emerged from the gap between existing coding techniques and the reliability requirements of mission-critical services. Classical codes have struggled when the channel has exhibited fast fading or burst noise, and quantum codes, while powerful, have required complex hardware. The proposed approach has addressed this gap by adopting quantum-inspired parity structures that have retained the lightweight processing of classical codes while mimicking the robustness observed in quantum systems. The method has employed a hybrid coding model that has integrated a modified stabilizer-like generator with a classical low-density parity check backbone. The encoder has produced redundant qubit-analogous syndromes that have allowed the decoder to infer error patterns with higher confidence. A sequential belief-propagation algorithm has been used, which has adjusted decoding weights according to channel variation. Simulations have been performed over Rayleigh and Rician channels, and the system has been tested under high mobility. The results of the proposed framework demonstrate substantial improvements over conventional coding methods. The hybrid stabilizer-LDPC structure reduces the bit error rate from 22.3% to 0.5% across an SNR range of 0–20 dB and lowers the frame error rate from 45.7% to 1.1%. Throughput improves from 4.2 Mbps at 0 dB to 13.8 Mbps at 20 dB, while the average decoding iterations decrease from 42 to 5, indicating reduced computational complexity. Under high-speed mobility, BER and FER remain low at 2.3% and 4.6%, respectively, while throughput stays above 10.2 Mbps and convergence requires only 12 iterations. These numerical results confirm that the proposed method provides highly reliable, efficient, and adaptive error correction suitable for next-generation networks.
The IoT (Internet of Things) networks face various types of security issues and threats at the physical layer because of massive connectivity, a large number of interconnections, and resource-constrained devices. In this research article, we have discussed the security issues, challenges, and threats at the physical layer in IoT networks, including eavesdropping, jamming, spoofing, unauthorized access, and pilot contamination. In this research, we have also highlighted various techniques and approaches to overcome the threats and issues at the physical layer of IoT networks and ensure secure connectivity, authenticity, authorization, and confidentiality. The noise aggregation and anti-eavesdropping techniques provide initial defense against unauthorized access and eavesdroppers. Radio Frequency Fingerprinting (RFF), Multiple Input Multiple Output (MIMO) systems, Non-Orthogonal Multiple Access (NOMA) technique, Secret Key Generation (SKG), and Reconfigurable Intelligent Surfaces (RIS) improve channel randomness, ensure secure connectivity, and optimize resource allocation. Cooperative techniques (jamming and beam- forming) enhance physical layer security through spatial diversity against various attacks and threats. Deep Learning-based Intrusion Detection Systems (DL-based IDS) detect and mitigate security threats, while Quantum Computing and Federated Learning solve cryptographic and privacy issues in distributed IoT networks. This research presents a comprehensive review, comparison, and analysis of physical layer security techniques for IoT networks.
