Journal of RF and Microwave Communication Technologies

An Overview of Transforming IoT with Millimeter-Wave

Heena T. Shaikh — 2026-03-31

The Internet of Things (IoT) is expanding rapidly, connecting billions of devices and transforming industries. However, this increase is pushing the boundaries of traditional wireless communication technology. This study introduces millimeter wave (mmWave), a high-frequency band that has the potential to revolutionise IoT connectivity by offering significantly more capacity and lower latency. Even while mmWave is still in its early stages, it has the potential to play a significant role in future IoT ecosystems. mmWave are defined as frequencies in the range of 30 GHz to 300 GHz. Compared to the lower frequencies used in Wi-Fi and cellular networks, mmWave frequencies offer a far greater range. Applications that require high-throughput and real-time communication can send data at significantly faster speeds because of its massive capacity. The IoT is expanding rapidly, necessitating ever-increasing bandwidth to support the enormous number of linked devices and their data-intensive applications. The demand for faster data rates and lower latency has drawn attention to mmWave technology, even if traditional sub-6 GHz frequencies have proven beneficial. mmWave runs in the 30 GHz to 300 GHz range, offering the possibility of multi-gigabit speeds and vast spectrum resources. However, using mmWave in IoT presents unique design challenges. This study looks at the key steps for getting beyond these challenges and making use of mmWave’s potential for better IoT connectivity.

Machine Learning-based RF Signal Classification for Cognitive Radio Networks: A Deep Learning Approach

Sareena A. Mulani — 2026-04-13

Machine Learning (ML) has become an essential enabler for improving the intelligence and adaptability of Cognitive Radio Networks (CRNs), especially in environments where spectrum availability is limited and highly dynamic. This study suggests a deep learning-driven structure for Radio Frequency (RF) signal classification to enhance spectrum sensing and decision-making processes within CRNs. Conventional signal classification methods largely depend on manually engineered features and prior domain knowledge, which restrict their effectiveness in complex and noisy conditions. The suggested approach uses DNNs, especially CNNs, to automatically extract discriminative and significant features from raw In-phase and Quadrature (I/Q) signal data to overcome these limitations. A variety of RF datasets with different modulation formats and SNR levels are used to train and test the model. To increase the model’s resilience and capacity for generalisation, preprocessing techniques including data augmentation and standardisation are used. According to experimental results, the proposed framework performs well even in noisy situations and provides superior classification accuracy, making it appropriate for real-time applications. When it comes to processing high-dimensional RF data, deep learning algorithms outperform more conventional ML techniques like Support Vector Machines (SVMs) and K-Nearest Neighbours (KNN). Moreover, integrating the proposed classification system into CRNs facilitates better spectrum utilisation, reduces interference, and supports dynamic spectrum access. Overall, this work highlights the potential of deep learning-based RF signal classification to significantly improve the efficiency and reliability of cognitive radio systems, offering a scalable solution for future wireless communication challenges.

A Conceptual Unified RF Front-end Architecture for SATCOM-MANET Integration in VTOL Uncrewed Aircraft Systems

Settapong Malisuwan — 2026-04-06

Vertical takeoff and landing (VTOL) uncrewed aircraft systems are being used increasingly in missions that demand flexible deployment, extended operational reach, and dependable communications in environments where infrastructure may be unavailable, damaged, or unreliable. As these platforms are expected to support both beyond-line-of-sight satellite communications and local tactical networking, integrating satellite communication (SATCOM) and mobile ad hoc network (MANET) subsystems on a single airframe has become an important RF engineering challenge. This challenge involves not only communication performance, but also size, weight, and power limitations, restricted antenna placement space, and the risk of electromagnetic interference among closely integrated RF subsystems. This study proposes a conceptual unified RF front-end architecture for SATCOM-MANET integration in VTOL uncrewed aircraft systems. Using a qualitative engineering approach, the study examines subsystem coordination, antenna placement strategy, self-interference mitigation, and communication resilience in dynamic RF environments. The proposed framework is intended to provide a conceptual foundation for future simulation, hardware prototyping, and experimental validation of multi-band airborne communication architectures.

Design and Study Analysis of a Patch Antenna for Two Different Substrates

M. S. Sethsanadi — 2026-01-14

This study investigates the effects of substrate dielectric constant (εr) and thickness (h) on the impedance and radiation performance of a microstrip patch antenna operating in the 2.3–2.4 GHz bands. According to simulation results, substrate parameters significantly affect return loss, gain bandwidth, and voltage standing wave ratio. Low-permittivity substrate ɛ_r = 2.2 (RT-duroid) antennas with a return loss of at least −24.19dB and a voltage standing wave ratio of 1.07 at substrate height h = 2.8 mm demonstrate good radiation efficiency and impedance matching. Fringing fields are improved by moderate substrate thickness (2.8–3.2 mm), which improves matching and gain without appreciably exciting surface waves. In contrast, antennas with substrates like FR4 with higher-permittivity ɛ_r= 4.4, offer return loss of at least −22.91dB and higher bandwidth (≈90 MHz), but their matching is acceptable but inferior with reduced gain (≈3.2–3.6 dB). In every scenario, resonance stays constant in the 2.3–2.4 GHz range with only slight frequency shifts brought on by changes in effective permittivity. Overall, the study confirms that ɛ_r = 2.2 with h ≈ 2.8–3.2 mm is an ideal configuration for effective ISM band antenna design since it offers the best trade-off between return loss, gain, directivity and matching.

Deep Learning-based Microwave Signal Classification for Intelligent Spectrum Management: A Review

Gadade Bhanudas Mahadeo — 2026-04-10

The rapid expansion of wireless communication systems and the proliferation of connected devices have significantly increased demand for efficient spectrum utilisation. Microwave frequency bands, widely used in radar, satellite, and wireless communications, are becoming increasingly congested, necessitating intelligent spectrum management strategies. Deep Learning has emerged as a transformative method for analysing complex microwave signal environments and enabling automated decision-making. A thorough examination of deep learning-based microwave signal categorisation methods for intelligent spectrum management is provided in this review study. Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Long short-term memory (LSTM) networks, and hybrid models are among the deep learning architectures that are examined in this paper, and their efficacy in extracting significant characteristics from unprocessed microwave signals is highlighted. Different signal representations, such as in-phase and quadrature (I/Q) data, spectrograms, and time-frequency features, are also discussed. The review further explores key applications, including spectrum sensing, interference detection, modulation classification, and dynamic spectrum allocation in cognitive radio systems. Additionally, the paper identifies major challenges such as data scarcity, low Signal-to-Noise Ratio (SNR) conditions, computational complexity, and real-time deployment constraints. Emerging trends, including transfer learning, federated learning, and edge-based AI, are also examined as potential solutions. Comparative analysis of existing methods demonstrates that deep learning significantly outperforms traditional signal processing techniques in terms of accuracy and adaptability. In order to facilitate effective and intelligent spectrum management in next-generation wireless communication systems, this review attempts to give researchers and practitioners a comprehensive understanding of recent developments and future research directions in deep learning-based microwave signal classification.