To address the growing congestion of the radio frequency (RF) spectrum, Cognitive Radio (CR) systems employ Automatic Modulation Classification (AMC) to dynamically optimize spectrum utilization without introducing protocol overhead. In modern Orthogonal Frequency Division Multiplexing (OFDM) standards, effective AMC requires advanced signal-processing techniques capable of accurately identifying modulation schemes under dynamic channel conditions. Therefore, maintaining robust performance under realistic environments remains a fundamental challenge. This paper evaluates how dataset scale, synthetic impairments, and hardware-induced signal impairments affect the cross-domain generalization of a Convolutional Neural Network (CNN) architecture for OFDM Automatic Modulation Classification (AMC), using 2D amplitude-phase histograms for signal representation. To assess these effects, the CNN is trained on five distinct datasets, encompassing both synthetically generated signals with varying scales and synchronization impairments, as well as a conducted hardware dataset. The cross-domain generalization of the trained models is assessed by evaluating them on a completely unseen indoor Over-The-Air (OTA) dataset collected across 13 distinct positions. Statistical analysis demonstrates that the large-scale synchronization-impaired synthetic dataset achieves the best generalization performance, reaching a mean indoor OTA accuracy of 93.36% and outperforming the limited-size conducted hardware dataset. Overall, this study demonstrates the critical role of data-generation strategies and establishes a robust baseline for achieving reliable cross-domain generalization of CNN-based AMC.

An intelligent reflective surface (IRS)-assisted cognitive radio (CR) multiple-input multiple-output (MIMO) communication system is considered. Incorporating cognitive radio and IRS capabilities into such a system yields significant improvements in system performance, including energy efficiency (EE) and receiver quality of service (QoS). For enhancing the attainable rate of secondary users (SU) without exceeding the interference temperature limit (IT) on the primary users (PU), a non-convex optimization problem is formulated, which is usually solved by means of alternative optimization (AO) methods such as block coordinate descent (BCD) algorithms. In this paper, we focus on deep reinforcement learning (DRL) approaches, specifically, the soft actor-critic (SAC) algorithm, to solve this optimization problem. For comparison, all simulation figures will be composed of a BCD benchmark beside the SAC curves. In addition, a 16-element MIMO antenna array for the secondary transmitter (ST) base station is proposed, designed, fabricated, and tested, yielding a 90% radiation efficiency with perfect impedance matching and acceptable return losses.

The article presents a theoretical analysis of spatial-frequency methods for filtering and jamming suppression in electronic warfare countermeasure systems. The main types of electronic jamming and their impact on the operation of modern communication, radar, and navigation systems are considered. The principles of spatial and frequency signal selection, as well as the features of their combined application to enhance jamming resistance, are analyzed. Special attention is given to adaptive spatial-frequency algorithms that provide effective suppression of both narrowband and broadband jamming under conditions of intensive electronic interference. A comparative analysis of spatial-frequency methods with individual spatial and frequency approaches is conducted, highlighting their advantages, limitations, and theoretical efficiency boundaries. Prospective directions for the development of spatial-frequency signal processing in the context of cognitive radio systems and integration with artificial intelligence methods are outlined. The obtained results can be used for further theoretical research and improvement of electronic warfare countermeasure systems.

FCAN: A frequency-aware cross-scale attention network for automatic modulation recognition in satellite communications

The purpose of the work: to analyze innovative technologies in telecommunications, to identify the most popular ones and to assess their impact on military communications.Research method: the methodological basis of the research is the general theory of systems using the methods of system analysis and the sampling method based on the correspondence of parameters that satisfy military communication and military command and control systems.Results of the study: ten information and telecommunication technologies are considered and selected, which, in the authors' opinion, will have a significant impact on the military communication system.These include: 5G/6G mobile communications, in which a significant transformation and transition to new formats are being carried out; cloud computing, which will ensure the development of not only military communications; Internet of Things technologies, which ensure the interaction of physical objects with each other and/or with the external environment, with the use of specialized equipment, software; satellite communications in low Earth orbit, providing broadband Internet access, the development of IoT systems and cellular communication services using Direct-to-Device (D2D) technology; quantum technologies that implement new methods of infotelecommunications and computing; Wi-Fi 7/Wi-Fi 8 wireless technologies.promising a maximum data transfer speed of 13 times faster for Wi-Fi 5 and almost five times faster for Wi-Fi 6. Wi-Fi 7; Networks with artificial intelligence; FTTx high-speed access technologies of a broadband telecommunications data transmission network, which uses fiber optic cable as the last mile in its architecture to provide all or part of the subscriber line; Software-defined networks are changing the traditional approach to network management, making them more flexible, manageable and adaptable to the requirements of communication consumers; cognitive radio systems.

Abstract This study presents a UWB antenna for spectrum sensing and a reconfigurable communication antenna with dynamic frequency selection designed for cognitive radio applications. The UWB antenna (ANT1) is used for spectrum sensing while the PIN diode-integrated frequency-reconfigurable antenna (ANT2) is used for communication. The antenna is designed on a FR-4-based dielectric material with dimensions of 25 × 35 × 1.6 mm 3 (0.175 λ ₀ × 0.245 λ ₀ × 0.011 λ ₀) and sized at the wavelength ( λ ₀) corresponding to the lowest resonant frequency in free space. The UWB antenna (ANT1) operates at a bandwidth of 2.1–12 GHz and the communication antenna (ANT2) operates at 3.8–6.3 GHz and 2.4–3.5 GHz; 8.6–12 GHz, respectively, depending on the pin diode off/on states. To increase the isolation between the antenna components and the impedance bandwidth, a T-shaped extension is integrated into the ground plane, which significantly enhances the isolation levels. The proposed antenna structure is transformed into a 4 × 4 8-port MIMO antenna architecture. A total of four PIN diodes are integrated into the MIMO antenna structure to provide frequency reconfigurability. The simulated and fabricated antenna structures are comprehensively evaluated in terms of MIMO performance parameters, including mutual coupling, diversity gain (DG), and envelope correlation coefficient (ECC).

A major challenge in enhancing the performance of multi-user cognitive radio networks lies in accurately characterizing the dynamic service arrivals of secondary users (SUs) for optimal spectrum utilization. To enhance protocol adaptability in complex traffic environments, this paper proposes a Traffic Pattern-Adaptive Allocation (TPA) protocol. Integrating Markov modeling with queueing theory, TPA employs multi-scale windows to concurrently measure SU traffic arrivals. By incorporating an adaptive window weight adjustment mechanism, the protocol achieves a granular characterization of the SU arrival process. Furthermore, it constructs a Probability Allocation Vector to dynamically map traffic states to channel allocation strategies, enabling automatic adjustment of resource policies in response to traffic fluctuations. Experimental results demonstrate that, compared to the Maximum Throughput Allocation protocol, TPA delivers higher throughput and lower packet rejection rates under complex traffic dynamics. This approach thus offers a robust solution for addressing the stochastic nature of user demands in next-generation cognitive radio systems.

The research work introduces a compact, frequency reconfigurable trident-shaped antenna suitable for cognitive radio (CR) applications. The antenna features a compact dimension of 30 × 30 × 1.6 mm3 and incorporates four PIN diodes, namely S 1 t , S 2 t , S 3 t and S 4 t to toggle between various operational frequency bands. This innovative design supports multiple states, enabling both wide-band (WB) sensing and narrowband communication modes across the spectrum. In the ON state, the PIN diodes exhibit low inductance and resistance, allowing current to flow with minimal impedance. In the OFF state, they demonstrate high resistance and low capacitance, effectively acting as insulators. When all four PIN diodes are activated, the antenna achieves a wide operational frequency range from 3.6 GHz to 12.8 GHz as wideband spectrum. By varying the switching combinations of all four diodes, the antenna can dynamically reconfigure to support 16 possible states with 23 different narrowband and wideband frequencies, providing full spectral coverage.

This work presents a complete development framework that begins with a new fork-shaped ultra wideband (UWB) antenna. The antenna is designed, optimized, fabricated, and experimentally validated. The prototype achieves a very wide bandwidth extending from 2.4 to 8 GHz, with stable radiation behavior and high efficiency. Building on this design, a 4 × 4 UWB MIMO array is developed. The four elements are arranged orthogonally to enhance isolation and provide strong pattern diversity. Next, the UWB antenna is transformed into a frequency-reconfigurable filtering antenna (filtenna). A single varactor diode is embedded in a modified radiator to enable continuous tuning from 2.45 to 3.48 GHz. A stepped ground, inset feed, and RF-choke-based biasing network are added to achieve stable tuning and low-loss filtering. The fabricated prototype shows a clear frequency shift with excellent matching and good radiation efficiency. To extend the concept, 2 × 2 and 4 × 4 MIMO filtenna configurations are also developed. Each stage introduces further structural refinement, including inter-element decoupling lines, L-shaped ground extensions, π-shaped shared ground sections, and pairwise high-impedance biasing networks. These features significantly enhance isolation and suppress surface-wave coupling. The proposed MIMO designs provide outstanding diversity performance. An envelope correlation coefficient (ECC) of approximately 10−2, a diversity gain close to 10 dB, and a channel capacity loss of less than 0.1 bits/s/Hz are accomplished. Additionally, the antenna exhibits deep total active reflection coefficient (TARC) nulls near − 15 dB, along with mean effective gain (MEG) values that are well-balanced around − 3 dB. Taken as a whole, the results confirm that the developed UWB antenna, its reconfigurable filtenna derivative, and their 2 × 2 and 4 × 4 MIMO extensions form a compact, low-loss, and highly efficient solution for next-generation 5G and cognitive radio systems.

The growing shortage of radio frequency spectrum, driven by the explosive increase in the number of wireless devices and data traffic volumes, makes Cognitive Radio (CR) technologies critically important for the future of telecommunications. This research addresses the challenge of dynamic spectrum management by developing a simulation model of a cognitive radio communication system based on the LTE network architecture. In contrast to existing solutions, the proposed model features a modular structure, allowing for the flexible integration and evaluation of various frequency resource occupancy prediction algorithms. The model is implemented in the MatLab environment and comprises three key modules: an LTE signal generation and processing module, which creates Radio Environment Maps (REM); a predictive neural network model training module; and a frequency resource occupancy prediction module. Particular focus is placed on utilizing the advanced Kolmogorov-Arnold Network (KAN) architecture for predicting unused Scheduling Blocks (SBs) within an LTE frame. Simulation results, encompassing 10,000 frames, demonstrated the high efficiency of the proposed approach. The KAN model achieved a prediction accuracy of 92.23% for identifying free frequency resources in a 10 ms frame. Comparative testing revealed that the KAN architecture outperforms the traditional LSTM network in accuracy by approximately 10% given an equal number of trainable parameters, and also converges faster during training. The practical significance of this work lies in providing a tool for accurate spectrum occupancy assessment and secondary user access planning, leading to a significant increase in the spectral efficiency and reliability of next-generation wireless networks.

Abstract Automatic modulation classification (AMC) plays a central role in adaptive and cognitive wireless communication systems by enabling reliable signal interpretation under varying channel conditions. This study presents a comprehensive evaluation of deep learning based AMC using constellation diagram representations under additive white Gaussian noise (AWGN) across six signal to noise ratio levels ranging from $$-5$$ - 5 , 0, 5, 10, 15, 20 dB. A balanced dataset consisting of 24,000 samples from eight modulation formats, namely BPSK, QPSK, 8PSK, 16PSK, 32PSK, 16QAM, 64QAM, and 256QAM, is employed to benchmark four convolutional architectures. These include the classical AlexNet, two lightweight modern backbones, MobileNetV2 and EfficientNet-B0, and the proposed Squeeze–Excite Convolutional Attention Lightweight Network (SCALNet). SCALNet integrates depthwise separable convolutions, squeeze–excite channel reweighting, and residual connections to enhance feature discrimination with minimal computational cost. Experimental results demonstrate that SCALNet consistently achieves the best robustness at low and mid SNRs, obtaining the highest F1-score at $$-5$$ - 5 dB (0.6004) and stable accuracy up to 20 dB, while requiring only 0.37M parameters and 1.36 MB of memory. Compared with AlexNet, SCALNet reduces parameter count and FLOPs by more than %99, while MobileNetV2 and EfficientNet-B0 provide intermediate performance–efficiency trade-offs. Efficiency analysis shows that SCALNet delivers the lowest inference latency (4.49 ms at 20 dB) and the smallest computational footprint across all SNRs, confirming its suitability for embedded or resource-constrained AMC deployments. These results position SCALNet as a lightweight yet highly reliable architecture, with future work planned to extend the evaluation toward fading, phase noise, and nonlinear RF impairments for emerging 6G communication scenarios.

ABSTRACT Integrated sensing and communication (ISAC) is a key technology for next‐generation cognitive radio (CR) systems that allows for the smooth coexistence of several wireless services and effective spectrum utilization. A comprehensive review of current developments in ISAC antenna configurations specifically designed for CR applications is provided in this article. The significant representations of CR architecture are dielectric resonator antenna (DRA)‐based CR systems and microstrip patch antenna (MPA)‐based CR systems. A communication and sensing antenna enables CR antennas to perform multiple purposes. The communication antenna, which is combined with a reconfigurable filter, offers narrowband operation for dependable data transmission, while the sensing antenna, which is usually constructed as an ultrawideband (UWB) structure, allows spectrum awareness by collecting a wide range of frequencies. The study covers key design approaches, reconfigurability strategies, and trade‐offs related to ISAC antennas in CR systems, focusing on their use in adaptive wireless communication and dynamic spectrum access. The review paper also addresses the design, challenges, and performance parameters of CR.

In cognitive radio (CR) systems, the accurate detection of spectrum holes is a cornerstone for efficient spectrum utilization. However, the increasing complexity of CR environments, particularly those with multiple primary users (PUs), has made precise spectrum sensing a paramount challenge. To address this challenge, this study introduces the ATC model, a novel deep learning architecture that integrates a parallel combination of attention mechanism-based networks and a Convolutional Neural Network (CNN). This hybrid design enables the model to capture both spatial and temporal features from the distinct statistics of sensing signals, thereby enhancing the accuracy of spectrum state detection. The model employs a Graph Attention Network (GAT) to extract complex topological features from graph-structured data derived from received signal strength, dynamically highlighting the most relevant information. To complement this, a CNN processes the sample covariance matrix of sensing signals, unlocking localized statistical correlations and hierarchical feature representations by treating the matrix as an image. Temporal dynamics, such as PU activity patterns, are modeled using a Transformer encoder, which leverages a self-attention mechanism to learn sequential features effectively. The proposed model is evaluated using both simulated and real-world datasets. For the simulated datasets, the model is assessed and compared with baseline methods under multi-PU scenarios across different channel models. For the real-world dataset, the experimental setup is configured for a single-PU scenario due to practical data collection limitations. In both cases, the ATC model demonstrates improved performance over the benchmarked spectrum sensing methods, exhibiting higher accuracy and robustness within the respective evaluation settings.

The rapid growth of wireless communication devices has intensified the need for efficient spectrum utilization and sustainable energy consumption. Cognitive Radio (CR) systems offer a promising solution by dynamically accessing underutilized frequency bands, but energy consumption remains a critical concern. This paper proposes an AI-enabled framework for cognitive radio networks that intelligently balances energy efficiency with communication performance. Leveraging machine learning algorithms, the system optimizes spectrum sensing, power allocation, and transmission scheduling to minimize power usage while maximizing throughput. Simulation results demonstrate that the proposed approach significantly reduces energy consumption without compromising data rates, highlighting its potential for green wireless communications. The integration of AI in CR networks paves the way for more adaptive, energy-aware, and high-performance communication systems.

Hidden radio frequency (RF) eavesdroppers pose a growing threat to wireless communication security due to their passive and covert nature. This paper presents a non-invasive framework for detecting such passive receivers by leveraging controlled backscattering characteristics of wireless devices. The proposed approach modulates the power supply states of suspected eavesdroppers to induce distinguishable variations in their backscatter signatures, enabling detection without requiring the target to transmit. Two complementary detection schemes are developed: 1) an energy detection (ED) scheme utilizing signal amplitude fluctuations, and 2) a cyclostationary feature detection (CFD) method exploiting inherent signal periodicity. Analytical models characterize detection, false alarm, and error probabilities under practical channel conditions. The framework mitigates the critical distance limitations of prior art by eliminating dependence on path-loss estimation and relying on intrinsic hardware-dependent backscattering properties. Sensitivity analyses provide operational guidelines on observation time and environmental stability. The framework reduces dependence on path-loss estimation by exploiting intrinsic, hardware-dependent backscattering characteristics. It offers analytical bounds and implementation criteria for next-generation wireless security and cognitive radio systems.

Cognitive radio (CR) has emerged as a promising solution to address spectrum scarcity in modern wireless communication networks. Non-parametric spectrum estimation approaches, including periodogram, kernel density estimation (KDE) and histogram-based methods, provide robust statistical tools for characterizing power spectral density (PSD) in heterogeneous stochastic environments. This paper presents a comprehensive exploration of these non-parametric strategies, highlighting their theoretical foundations in probability theory and statistical inference, as well as their practical significance for spectrum analysis. Simulation-driven evaluations demonstrate the statistical reliability of these methods under varying signal-to-noise ratio (SNR) conditions. Notably, KDE consistently outperforms other methods in minimizing mean squared error (MSE) and improving detection probability, underscoring its effectiveness as a density estimation technique. The results emphasize the importance of selecting suitable non-parametric methods for spectrum analysis in CR systems. The discussion concludes by outlining prospective research pathways, including the integration of non-parametric inference with advanced machine learning paradigms. Additionally, extending these methodologies to high-dimensional, time-varying and non-stationary signal models central to 5G and Internet of Things (IoT) ecosystems holds significant promise. By exploring these avenues, researchers can further enhance the performance and efficiency of CR systems, ultimately mitigating spectrum scarcity and enabling more efficient wireless communication networks. Further research in this area can lead to significant advancements in the field.

This paper presents a compact planar microstrip bandpass filter with reconfigurable triple- and quadruple- passband responses for dynamic spectrum access in cognitive and software-defined radio systems. Unlike conventional multiband filters that employ fixed resonant paths or complicated switching networks, the proposed topology enables selective switching between two distinct multiband operating modes using a simple and compact circuit configuration. A half-wavelength stepped open stub is strategically incorporated to generate multiple transmission zeros near the passband edges, thereby enhancing stopband suppression and passband selectivity. Reconfigurability is achieved by integrating only two PIN diodes into the filter structure, resulting in low switching complexity and minimal circuit overhead. The proposed filter is fabricated on a Rogers RO4003C substrate and occupies a compact area of 0.33λg × 0.28λg. In the triple-passband mode, measured center frequencies of 2.53, 3.37, and 4.37 GHz are obtained, whereas in the quadruple-passband mode, center frequencies of 2.57, 3.10, 3.89, and 4.45 GHz are achieved. Measured results are in good agreement with full-wave electromagnetic simulations, confirming the effectiveness and practical feasibility of the proposed reconfigurable multiband filter for modern wireless front-end applications.

Rotatable antenna (RA) technology has recently drawn significant attention in wireless systems owing to its unique ability to exploit additional spatial degrees-of-freedom (DoFs) by dynamically adjusting the three-dimensional (3D) boresight direction of each antenna. In this letter, we propose a new RA-assisted cognitive radio (CR) system designed to achieve efficient spectrum sharing while mitigating interference between primary and secondary communication links. Specifically, we formulate an optimization problem for the joint design of the transmit beamforming and the boresight directions of RAs at the secondary transmitter (ST), aimed at maximizing the received signal-to-interference-plus-noise ratio (SINR) at the secondary receiver (SR), while satisfying both interference constraint at the primary receiver (PR) and the maximum transmit power constraint at the ST. Although the formulated problem is challenging to solve due to its non-convexity and coupled variables, we develop an efficient algorithm by leveraging alternating optimization (AO) and successive convex approximation (SCA) techniques to acquire high-quality solutions. Numerical results demonstrate that the proposed RA-assisted system substantially outperforms conventional benchmark schemes in spectrum-sharing CR systems, validating RA’s capability to simultaneously enhance the communication quality at the SR and mitigate interference at the PR.

In recent years, reconfigurable intelligent surface (RIS) has been introduced into cognitive radio (CR) systems for signal enhancement and interference suppression. However, the existing RIS design for CR relies on the instantaneous channel state information (CSI) of all individual channels, which incurs high signal processing complexity and substantial channel estimation overhead. To overcome this challenge, in this paper, we propose a two-timescale (TTS) resource allocation scheme for RIS-aided CR, in which the long-term phase shifts at the RIS remain fixed during a frame, while the short-term transmit power at the secondary user transmitter (SU-TX) is optimized in each time slot within a frame. Based on this scheme, we formulate an optimization problem to maximize the ergodic rate of the SU subject to the average interference temperature (IT) constraint at the primary user receiver (PU-RX) and the average transmit power constraint at the SU-TX. To solve the non-convex optimization problem, we propose a two-stage algorithm. In the first stage, a constrained stochastic successive convex approximation (CSSCA) based algorithm is proposed to optimize the RIS phase shifts by exploiting the statistical channel state information (S-CSI) of both reflecting link and direct link channels. Once the phase shifts are obtained, in the second stage, the Lagrange dual-decomposition method is adopted to optimize the transmit power by using instantaneous CSI (I-CSI) of the composite channels, which represent the summation of the cascaded reflecting link and the direct link channels. Furthermore, three low-complexity algorithms are proposed for the long-term phase shift design. Simulation results have validated the effectiveness of our proposed TTS scheme with lower signal processing complexity and channel estimation overhead.

Automatic Modulation Classification (AMC) is a core technology for future wireless communication systems, enabling the identification of modulation schemes without prior knowledge. This capability is essential for applications in cognitive radio, spectrum monitoring, and intelligent communication networks. We propose an AMC system based on a hybrid Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) architecture, integrated with a Software Defined Radio (SDR) platform. The proposed architecture leverages CNNs for spatial feature extraction and LSTMs for capturing temporal dependencies, enabling efficient handling of complex, time-varying communication signals. The system’s practical ability was demonstrated by identifying over-the-air (OTA) signals from a custom-built FM transmitter alongside other modulation schemes. The system was trained on a hybrid dataset combining the RadioML2018 dataset with a custom-generated dataset, featuring samples at Signal-to-Noise Ratios (SNRs) from 0 to 30 dB. System performance was evaluated using accuracy, precision, recall, F1 score, and the Area Under the Receiver Operating Characteristic Curve (AUC-ROC). The optimized model achieved 93.48% accuracy, 93.53% precision, 93.48% recall, and an F1 score of 93.45%. The AUC-ROC analysis confirmed the model’s discriminative power, even in noisy conditions. This paper’s experimental results validate the effectiveness of the hybrid CNN-LSTM architecture for AMC, suggesting its potential application in adaptive spectrum management and advanced cognitive radio systems.