False Data Injection Attacks (FDIAs) present a significant threat to smart grids by manipulating measurement data, which may lead control centers to make incorrect operational decisions. Accurate and efficient detection of FDIAs is critical for ensuring reliable grid operation. Existing deep learning approaches often fail to capture both short-term local features and long-term dependencies in power grid data, and they typically show weak correlations with past and future time series information, reducing the trustworthiness of detection results. Similarly, conventional Intrusion Detection Systems (IDS) struggle to detect advanced FDIAs due to their reliance on predefined signatures and rule-based mechanisms. To overcome these limitations, we propose IG-APSO-DNN, a two-stage deep learning model for detecting FDIAs in smart grids. The first stage employs Information Gain (IG) and Adaptive Particle Swarm Optimization (APSO) for feature selection, reducing data dimensionality and improving model efficiency. The second stage uses a Deep Neural Network (DNN) to effectively capture both spatial and temporal patterns in smart grid measurements. The proposed model is evaluated on the Industrial Control System (ICS) Cyber Attack Power System Dataset, which simulates various FDIA scenarios. Results demonstrate that IG-APSO-DNN significantly outperforms traditional methods, improving key performance metrics including detection accuracy, precision, recall, and F-measure, while ensuring reliable operation of the smart grid. This study presents a robust anomaly-based IDS framework and highlights future directions, such as real-world validation, adaptive learning, exploration of novel optimization algorithms, and addressing scalability and real-time processing challenges.

Unmanned aerial vehicles (UAV)-mounted simultaneous transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) systems can provide full-dimensional coverage and flexible deployment opportunities in future 6G-enabled IoT networks. However, practical imperfections such as jittering and airflow of UAV could affect the phase shift of STAR-RIS, and consequently degrade network security. In this respect, this paper investigates the impact of phase shift errors on the secrecy performance of UAV-mounted STAR-RIS-assisted IoT systems. More specifically, we consider a UAV-mounted STAR-RIS-assisted non-orthogonal multiple access (NOMA) system where IoT devices are grouped into two groups: one group on each side of the STAR-RIS. The nodes in each group are considered as potential Malicious nodes for the ones on the other side. By modeling phase estimation errors using a von Mises distribution, an analytical closed-form expressions for the ergodic secrecy rates under imperfect phase adjustment are derived. An optimization problem to maximize the weighted sum secrecy rate (WSSR) by optimizing the UAV placement is formulated and is then solved using a linear grid-based algorithm. Monte Carlo simulations are provided to validate the analytical derivations. The impact of phase estimation errors on system’s secrecy performance is analyzed, providing critical insights for the practical realisation of STAR-RIS deployments for secure UAV-enabled IoT networks.

Ensuring the secure communication of unmanned aerial vehicle-assisted reconfigurable intelligent surface (UAV-RIS) is crucial in maintaining seamless connection in next-generation Internet of Things (IoT) networks. For this purpose, intelligent beamforming is essential to ensure secure data transmission from IoT devices to UAV-RIS and optimize communication while preventing adversarial attacks. This paper proposes a novel framework of federated learning for long short-term memory-based deep deterministic policy gradient with an attention mechanism (F-DDPG-AM). The proposed algorithm aims to improve security and mitigate potential threats in UAV-RIS-assisted IoT networks. The F-DDPG-AM combines the federated LSTM’s power to capture long-term dependencies in sequential data with the attention mechanism to focus on key network states and improve decision-making efficiency. The F-DDPG-AM framework improves learning efficiency, accelerates convergence, and enhances resilience against adversarial attacks by selectively prioritizing crucial network information and focusing on insecure scenarios. In addition, federated learning in the proposal ensures secure decision-making through local training for UAV-RIS-enabled IoT networks. The F-DDPG-AM enhances system scalability, trustworthiness, and compliance with secure machine learning principles by decentralizing the training process. The simulation results demonstrate the superior performance of the proposed F-DDPG-AM framework in defending against attacks, significantly outperforming traditional security approaches and other existing reinforcement learning models.

The growth of the Internet of Medical Things (IoMT) has intensified electromagnetic-interference (EMI) concerns in the 5.8 GHz ISM band. Frequency-selective surfaces (FSS) offer a practical alternative to rigid enclosures for wearable and conformal medical devices, providing lightweight, integrable shielding. This article presents a polarization-insensitive, conformal FSS designed for broadband EMI shielding in 5.8 GHz ISM-band for IoMT applications. The proposed structure employs a complementary patch configuration printed on opposite sides of a thin dielectric substrate to achieve a broadband band-stop response. The top-layer resonator defines the lower stopband withS21 < −10 dB and a compact unit-cell size of 0.17λ₀, where λ₀ corresponds to the resonance frequency. The FSS achieves a <−10 dB stop band from 4.0 GHz to 7.7 GHz, centered at 5.8 GHz with the bandwidth of 63.793% that fully covered the ISM spectrum. Owing to its fourfold symmetric geometry, the design demonstrates polarization independence and maintains angular stability up to 75°. To validate the real time performance, A fabricated 6×6 array prototype is experimentally evaluated under both planar and conformal conditions, showing strong agreement with simulation results. The proposed lightweight and flexible FSS provides an effective solution for broadband EMI suppression, enabling reliable electromagnetic compatibility (EMC) and signal integrity in compact and wearable biomedical IoT systems.

Auditory stimuli strongly shape emotional and physiological states, making them central to affective computing and mental health technologies. We present the study of three auditory conditions, spiritual meditation (SM), music (M), and natural silence (NS), using acoustic time-series signals. To support this, we introduce the Spiritual, Music, Silence Acoustic Time Series (SMSAT) dataset, a benchmark of controlled acoustic recordings with demographic diversity. We develop a contrastive learning-based SMSAT encoder that learns discriminative embeddings from ATS data, achieving 99% accuracy. In addition, we propose the Calmness Analysis Model (CAM), integrating multi-domain features for affective state classification, achieving a 99% accuracy in the three-stimulus classification task. Inter & intra-class feature space separability, calmness evaluation using Temporal Segmented Response Profiling (TSRP) confirm significant physiological differences across auditory conditions, with SM showing stronger effects on cardiac response characteristics (CRC).WaveGAN is used to generate additional dataset. Under subject-wise evaluation, CAM reached$98.4\%$accuracy, and the SMSAT Encoder achieved$96.5\%$accuracy. This work provides a validated dataset and scalable deep learning framework for stress monitoring, well-being, and therapeutic audio interventions.

This paper investigates the secure communication capabilities of a non-orthogonal multiple access (NOMA) network supported by a STAR-RIS (simultaneously transmitting and reflecting reconfigurable intelligent surface) deployed on an unmanned aerial vehicle (UAV), in the presence of passive eavesdroppers. The STAR-RIS facilitates concurrent signal reflection and transmission, allowing multiple legitimate users-grouped via NOMA-to be served efficiently, thereby improving spectral utilization. Each user contends with an associated eavesdropper, creating a stringent security scenario. Under Nakagami fading conditions and accounting for phase shift inaccuracies in the STAR-RIS, closed-form expressions for the ergodic secrecy rates of users in both transmission and reflection paths are derived. An optimization framework is then developed to jointly adjust the UAV's positioning and the STAR-RIS power splitting coefficient, aiming to maximize the system's secrecy rate. The proposed approach enhances secure transmission in STAR-RIS-NOMA configurations under realistic hardware constraints and offers valuable guidance for the design of future 6G wireless networks.

Reconfigurable intelligent surfaces (RISs) have gained considerable interest because of their inherent passive and energy-efficient design. Integrating unmanned aerial vehicles (UAVs) with reconfigurable intelligent surfaces (RIS), known as UAV-RIS, can significantly improve network performance and serve as a crucial enabler for advancements in 6G mobile networks. However, ensuring security in UAV-RIS systems poses notable challenges, particularly in the presence of imperfect channel state information (CSI) and beamforming complexities. In this paper, we identify the critical security requirements for UAV-RIS beamforming in practical scenarios. To address these challenges, we introduce a novel deep deterministic policy gradient with a distributional critic (DDPG-DC)-based beamforming approach aimed at securing UAV-RIS systems while improving the overall secrecy rate. Our proposed secure beamforming solution achieves up to a 48% performance improvement compared to existing state-of-the-art algorithms.

Unmanned aerial vehicles (UAVs) in integration with reconfigurable intelligent surfaces (RIS) play a crucial role in improving wireless communication coverage and enhancing overall performance. However, optimizing the beamforming for the RIS and base station (BS) is critical in improving coverage and ensuring connectivity in densely populated areas. The primary challenge in this process arises from the diverse Quality of Service (QoS) requirements set by user equipment (UEs). To address this complexity, machine learning algorithms are employed to predict the optimal beamforming configurations for both the BS and RIS. However, the traditional supervised learning methods are becoming less effective due to the ever-changing demands of UEs, as these methods rely on fixed data patterns that struggle to adapt to the fluctuating QoS requirements of UEs. Thus, in this paper, we propose an unsupervised learning-based deep learning (DL) approach to jointly predict the optimal beamforming matrix for RIS and BS, enhancing communication coverage and maximizing QoS satisfaction of UEs. The proposed DL-based beamforming adaptively predicts the beamforming matrix, facilitates efficient data exploration during the initial learning phase, and seamlessly scales as the process advances, thereby enhancing overall performance. Numerical results demonstrate that the proposed DL-based RIS and BS beamforming outperforms by up to 89%, compared to the state-of-the-art methods.

The evolution of Artificial Intelligence (AI) integrated with the Sixth-generation ($\mathbf{6 G}$) framework poses significant challenges to low-latency applications. Recently, semantic communication has emerged as a promising technique for future intelligent applications. However, the resource management problem combined with semantics is not fully explored. In this paper, we present a deep reinforcement learning-based twin-delayed deep deterministic policy gradient (TD3) for task-centered semantic communication. The proposed TD3 algorithm optimizes bandwidth, and semantic information and prioritizes data with maximum signal-to-noise ratio (SNR) for the efficient transmission of useful information. Simulation results demonstrate the effectiveness of the proposed TD3 scheme compared to state-of-the-art work in terms of transmission efficiency by up to $36 \%$ for varying users and up to $33 \%$ for varying SNR.

Unmanned aerial vehicles (UAVs) have been adopted in various applications, including agriculture, public safety, surveillance, and crucial military missions. However, alongside their advantageous nature, UAVs have also been employed for malicious activities, leading to an increased requirement for timely detection and identification. Despite significant progress in UAV detection, challenges persist, particularly concerning various types of UAVs, the payload carried by UAVs, and the traits of their flight. Employing single machine learning for detection and identification has limitations due to the inability to handle diverse datasets and acquire complex relationships. Therefore, in this paper, we introduce a novel integration of the Visual Geometry Group-based convolutional neural network (VGG-CNN) framework employed for detection with the Mask Region-based convolutional neural network (MR-CNN) for identification of UAVs (jointly termed MR-DCNN). For efficient deployment of MR-DCNN, we add diversity to the dataset by performing data augmentation of new images in the training dataset for the detection of various types of UAVs, payload categories, and flight characteristics. The performance evaluation of the MR-DCNN approach was conducted via simulations, revealing superior detection capabilities for malicious UAVs compared to existing methods.

In this paper, a new approach has been proposed and investigated with the help of variational auto-encoder (VAE) as a probabilistic model to reconstruct the transmitted symbol without sending the data bits out of the transmitter. The novelty of the proposed End-to-end (E2E) wireless system is in representing the symbol as a image hot vector (IHV) that contains the features of the shape such as spikes, closed squared frame, pixels index location and pixels grey-scale colours. The previously mentioned features are inferred by latent random variables (LRVs). The LRVs are used for fronthaul and backhaul data representation. The LRVs parameters have only been transmitted through the physical wireless channel instead of the original bits as in the classical modulations or the hot vectors in the Auto-encoders (AE) E2E systems. The new proposed VAE architecture achieved the reconstruction of the symbol from the received LRV. The results show that the VAE with a simple classifier can provide a better symbol error rate (SER) than both AE baseline and classical Hamming code with hard decision decoding, especially at high$E_{b}/N_{o}$.

Deep learning (DL) is playing an increasingly important role in the design of next-generation communication systems. In this paper, we apply DL algorithms to enhance signal detection and performance of multiple-input-multiple-output (MIMO) based orthogonal frequency-division multiplexing (OFDM) systems with index modulation (IM). The proposed detector termed DLIM is used as fully connected layers of a deep neural network (DNN) and adopted to achieve minimum bit error rates in IM-MIMO-OFDM over Rayleigh wireless channels. To show the enhancement of the proposed algorithm, the DL model is trained initially offline using data generated from simulation based on common statistical wireless channel models. DLIM is then adopted to recover the online transmitted data. Simulation results confirm that the proposed DLIM can detect the transmitted symbols with a performance comparable to near-optimal BER in a shorter runtime than required by the existing classical detectors.

In this paper, we are re-modelling the intelligent reflecting surfaces (IRS) assisted communication systems using the auto-encoder (AE) deep learning (DL) technique to represent the classical IRS system as an end-to-end communication system. The cascaded channels from source to sink through the IRS have been transformed to a deep neural network (DNN) that learns how to reduce the wireless environment impairments effect by optimizing the representation of transmitted symbols. The proposed system design shows superior symbol error rate (SER) performance under the AWGN channel compared to both classical IRS and conventional AE end - to-end systems. The relation between improvement of performance and the capability of the proposed AE to learn optimized presentation for transmitted symbols is being explained through observing and comparing the baseline AE constellations learning with the ones that the proposed model learned.