This paper presents the procedure for calculating the electric field level in the vicinity of base stations (BS) implemented in 5G technology. The specificity of the 5G system is the application of dynamic routing of traffic radiation beams directly towards active users (beamforming). In the paper, the following characteristics are calculated: 1. maximum field level; 2. mean field level without application of power control and 3. mean field level with application of radiated power control in the BS for traffic channels depending on the distance from the BS, taking into account only the field created by the considered BS, not the neighbouring ones. In this calculation, a uniform distribution of user density in the BS cell and an equal traffic load of all users during a sufficiently long-time interval were assumed. The parameters that change in the graphs are the characteristics of the applied antenna (width of its main radiation beam, gain in the main and side radiation beams), BS height and radius of the BS cell. The most interesting result is that the mean level of radiation in the case of application of control of the emitted power can also increase with increasing distance from the BS, whereby the shape of the obtained characteristics depends on the radius of the BS cell. However, in an absolute sense, the level of the electric field in this case is lower than the values allowed by regulations and can pose a danger only through the aggregate value with other sources of radiation.
This paper presents a modified approach to implementing Sigma Point Kalman Filters (SPKFs), in which the Singular Value Decomposition (SVD) of the covariance matrix is used to generate sigma points for the unscented transform. The advantage of the proposed approach is that it requires just (n+1) instead of the usual (2n+1) sigma-points with only a minor reduction of precision. The proposed method for selecting the sigma-points to adjust the normal process parameters during the update phase of the filter can be intuitively visualized in the 2D case by plotting the covariance ellipsoid. The proposed reduced sigma-point filter achieved satisfactory performance in a case study involving state-of-charge (SoC) estimation for vehicular batteries when compared to the original SPKF implementation.
A voltage differencing buffered amplifier (VDBA) and a current differencing buffered amplifier (CDBA) are used in the proposed compact meminductor emulator design. The proposed emulator reduces the passive components number by utilizing parasitic components of CDBA to obtain a compact layout. This feature also aids in the realized circuit's ability to operate at a high frequency. The element in the suggested design has memory of its previous state, since the effective meminductance is dependent on the total charge stored on the capacitors. By altering the capacitance values, pinched hysteresis loops between magnetic flux (Phi) and current (i) are shown throughout a broad frequency range (1 kHz to 3 MHz). Both incremental and decremental hysteresis assessments, non-volatility measures, and transient response tests verify that the suggested emulator is operating correctly. The theoretical predictions are closely matched with LTspice simulations utilizing 0.18 & micro;m process characteristics. The non-ideal analysis of the suggested configuration verifies the accuracy of the derived expressions and simulation results. Lastly, the meminductor practical application in actual circuits is confirmed by incorporating it into a chaotic oscillator.
This work presents a compact floating configuration of the memtranstor (MT), a recently introduced memory element defined by the direct relationship between magnetic flux (phi) and charge (q). Unlike prior designs, the proposed emulator eliminates the need for multipliers or other complex circuitry, resulting in a simplified and power-efficient architecture. The circuit employs a single voltage differential transconductance amplifier (VDTA) and one voltage differential current conveyor (VDCC) as active devices, together with three grounded capacitors and one grounded electronic resistor as passive components. The emulator successfully reproduces the fundamental phi-q relationship and exhibits origin-crossing pinched hysteresis loops under sinusoidal excitation - a defining characteristic of memtranstive systems. It operates reliably at a supply voltage of +/- 0.9V and supports electronic tunability through adjustment of the VDTA and VDCC transconductance parameters), ensuring adaptability across a wide range of operating conditions. Extensive validation was carried out through mathematical modeling and LTSpice simulations based on a 180-nm CMOS process. The evaluation includes demonstration of memory effects, Monte Carlo analysis, temperature sensitivity studies, and characterization of pinched hysteresis loops under variations in DC control voltage, excitation frequency, and transconductance values. A full-custom layout was implemented, occupying a silicon area of 2529.49 mu m & sup2;. Non-ideal effects, including parasitics at active device terminals, were also thoroughly analyzed to ensure functional robustness. Distinguished by its compact structure, low component count, and ease of integration, the proposed design provides a robust and efficient platform for MT-based applications. Its demonstrated performance highlights significant potential for neuromorphic computing, chaos-based systems, nonlinear dynamics, and other emerging analog memory-oriented domains.
With the ever-increasing realization of the limitations of the radio frequency (RF) spectrum in the evolution towards 5G and 6G communication systems, VLC has been recognized as a high-speed and secure communication system that supports gigabit-per-second data rates. However, in most cases, VLC communication systems are often hampered by issues related to line-of-sight (LOS) propagation and signal shadowing. This paper proposes the optimization of indoor VLC communication systems through the lighting geometry.Using the simulation environment set up in the 5 m & times; 5 m & times; 3 m indoor environment and carried out for different LED array patterns including single-centred patterns, square patterns, and circular patterns of 12-LEDs with a radius of 1.5 m, it was determined that the 12-LED array pattern with a radius of 1.5 m lead to an average SNR of 36.7 dB, a BER of 2.5 & times;10-9, and Fa of 8.24. The superiority of the proposed system was strictly observed in terms of significant improvement over the standard array patterns. Also, from the research carried out, it was ascertained that the proposed system can attain high-order modulation constellations of 256-QAM. The findings of the study clearly highlight that the collaboration of geometric organization and intelligent surfaces is critical in the design of future IoT-enabled communications and smart buildings requiring high spectral efficiency.
With the rise of Industry 4.0, wireless sensor networks (WSN) are increasingly vital, particularly in indoor positioning applications using various positioning techniques. Radiocommunication-based technologies such as WiFi and Bluetooth Low Energy (BLE), Ultra-Wideband (UWB) technologies are receiving growing attention in this area. The Free Space Path Loss (FSPL) model establishes a relationship between signal attenuation during propagation and distance. Previous studies have demonstrated that considering the angle between individual modules, i.e., their orientation, can significantly impact the accuracy of distance estimation. In this paper, an orientation-based positioning method is proposed. The appropriate FSPL model is selected based on the orientation of both the anchors and the station unit to compute the distance, followed by position determination using trilateration. This method was compared to a generalized FSPL model approach. In the tested scenario, the proposed orientation-based method demonstrated superior performance, achieving a 13% relative improvement in the average of the Mean Absolute Error (MAE).
This paper presents on FBG Code Supervision into Single Core Underground Cable Power Line for Health Condition. Fiber Bragg Grating (FBG) is a novel technique for network health monitoring, offering numerous advantages such as high precision, multiplexing capabilities, resistance to electromagnetic interference, and excellent repeatability. The FBG sensor integrated into the network has been employed in numerous applications to monitor system and environmental conditions. Any interruption on the system or environment can be sensed and monitored through the status of network health. A consistent Fiber Bragg Grating (FBG) array situated before and after the splitter will be utilized to reflect the optical signal that has been transferred along the fiber. As the network's FBGs use different wavelengths to represent each line connection, FBG codes will serve as an indicator between one connection to another. In this paper, a new technique is shown to monitor network health for single core underground cable by developing fault branch identification FBG configuration designs (FBG codes) up to 36 branches while maintaining quality of power transmission. Its capabilities are confirmed by simulations using real model data and it shows the potential to solve the various problems presently we have.
Stress is one of the phenomena characterizing today's modern human society. High stress is also reflected in voice and speech in many ways. This paper presents a detailed analysis of voice period fluctuations when a speaker is exposed to psychological stress. The initial features determined are the length of the fundamental period and local jitter extracted from the voiced regions of the speech signal. Statistical parameters of their distribution and other derived features were taken into account in the experimental analysis. Multiple parameters were compared pairwise for normal speech and speech under stress, and parameters with significant differences caused by stress were searched for. The best reliability of 100% was achieved for two parameters. The test speakers were a group of nine university students under real stress, which was induced by oral exams in front of an examination board.
The ultra-lift Luo converter is a fourth-order DC-DC step-up converter capable of providing high voltage gain with low voltage and current ripples. However, its nonlinear dynamics, switching behaviour, and parametric uncertainties make controller design a challenging task. This paper proposes a Lyapunov-stable interval type-2 fuzzy PID controller for output voltage tracking of the ultra-lift Luo converter. The controller performance is investigated under step reference voltage changes, input voltage variations, and output capacitor uncertainties. A conventional PID controller is considered for comparison. Performance evaluation is carried out in terms of rise time, settling time, recovery time, overshoot/undershoot, and integral absolute error using a MATLAB/Simulink-based interval type-2 fuzzy logic toolbox. Simulation results demonstrate that the proposed controller achieves superior transient performance and provides enhanced robustness against parameter uncertainties compared with the conventional PID controller.
Finite element modelling (FEM) has become an important tool for analysing biomechanical behaviour in rehabilitation systems. However, the optimization of model parameters is often performed empirically, resulting in limited reproducibility and increased computational cost. This work proposes a statistical optimization framework for FEM-based rehabilitation systems using the Taguchi method combined with analysis of variance (ANOVA). An L9 orthogonal array was employed to investigate the influence of material, geometric, and loading parameters while reducing the number of simulations required compared with full factorial approaches. The biomechanical response of the system was analysed through three-dimensional FEM simulations implemented in COMSOL Multiphysics. Experimental measurements acquired using a deformation gauge-based electronic acquisition system were used for model validation. ANOVA enabled the identification of the most influential parameters affecting deformation behaviour and mechanical response stability. The optimized configuration showed reduced deformation variability and improved agreement between numerical and experimental results. In particular, the optimization procedure reduced the average modelling error from 12.4% to 4.1%, while improving response consistency under different loading conditions. The proposed methodology provides a computationally efficient and physically interpretable approach for the optimization of rehabilitation-oriented biomechanical systems.
Securing RPL-based Internet of Things (IoT) networks remains a critical challenge due to the limited computational and energy resources of sensor nodes, making them vulnerable to routing attacks. This paper proposes a light-weight Intrusion Detection System (IDS) for the detection of five prominent RPL-specific attacks, namely Version Number Attack, DIS Flooding, Reduced Rank, Worse Parent, and Local Repair Attack. Unlike previous work that employed traditional datasets, a novel dataset is generated by simulating actual real-world RPL-based IoT scenarios, encompassing both normal and malicious behaviors. The dataset contains 13 features representing network behavior that were carefully selected to reflect the characteristics of RPL communications. Five machine learning models consisting of K-Nearest Neighbors (KNN), Decision Tree (DT), Random Forest (RF), XGBoost, and LightGBM were trained and evaluated using both the entire feature set and a subset of the top 7 most informative features. LightGBM performed the best among them, with 99.47% accuracy, 99.58% precision, 99.37% recall, and 99.47% F1-score. Even with fewer features, it still yielded robust results, confirming its robustness and viability in constrained environments. The proposed architecture offers a feasible and efficient solution to enhancing RPL security in real-world IoT networks. The proposed framework highlights the potential of tailored machine learning models and original datasets to detect protocol-specific threats while respecting the limitations of low-power networks.
Distributed adaptive filtering over heterogeneous networks is challenging due to inconsistent local update directions, varying sensing quality, and unequal reliability across nodes. Conventional diffusion-based adaptive algorithms typically promote cooperation through parameter exchange, which may be insufficient to regulate discrepancies in local learning dynamics under heterogeneous conditions. This paper proposes a reliability-aware coupled-gradient distributed adaptive filtering framework in which cooperation is introduced through two complementary mechanisms: reliability-aware gradient coupling before adaptation and reliability-aware post-update smoothing after intermediate estimation. The proposed approach enables nodes with more reliable observations to exert stronger influence during both adaptation and estimate refinement. A unified network recursion is developed, and the mean as well as mean-square behavior of the resulting algorithm is analyzed using a Lyapunov-based framework. Explicit conditions for stability are established, and the analysis characterizes how gradient coupling and post-update smoothing jointly influence convergence dynamics, disagreement reduction among neighboring update directions, and steady-state estimation accuracy. Simulation results under heterogeneous sensing conditions demonstrate that the proposed strategy improves robustness relative to purely local adaptation and achieves competitive distributed estimation performance compared with conventional diffusion LMS methods. Furthermore, close agreement between theoretical predictions and simulated steady-state behavior validates the developed analytical model.
The subject of the study is the process of developing motion models on a plane for a two-wheeled experimental balancing prototype (TEBP) considered as a control object. The aim of the study is to develop an approach for constructing mathematical models of the translational and rotational motions of the TEBP, taking into account mechanical and electromechanical processes. A physical model of the TEBP was developed and its nonlinear mathematical description was obtained. The application of analytical linearization made it possible to derive linear approximation models in the form of differential equations with constant coefficients. Using the Laplace transform, transfer functions were obtained, enabling the construction of block diagrams in both time and frequency domains. The motion models of the TEBP are also represented in the state-space form in the controllability canonical form. The simulation results confirm the feasibility of using the developed mathematical descriptions as dynamic models of the control object. The scientific novelty lies in the development of a comprehensive approach to modeling the motions of the TEBP, taking into account the acting forces and moments.
Deep learning (DL) models require significant computational resources, making their deployment on edge devices with limited power and hardware capabilities challenging. Field-programmable gate arrays (FPGAs) provide an effective platform for accelerating such workloads because of their inherent parallelism and energy efficiency. This study investigates the impact of layer-wise parallelization levels, represented by folding coefficients, on the resource utilization and performance of FPGA-based DL accelerators, with a specific focus on convolutional (CONV) and fully connected (FC) layers. A LeNet-based accelerator model was implemented using the Xilinx FINN framework with W1A2 quantization (1-bit weights and 2-bit activations). Three folding coefficients, namely low (L), medium (M), and high (H), were defined for both the CONV and FC layers, yielding nine unique parallelization configurations. These accelerators were deployed on the PYNQ-Z1 platform and evaluated using the Fashion-MNIST dataset. A comprehensive evaluation quantifies key metrics, such as throughput (frames per second, FPS), resource utilization (look-up tables (LUTs), flip-flops (FFs), and block RAMs (BRAMs)), and power consumption. The results show that lower folding levels, corresponding to higher parallelism, significantly enhance the throughput, reaching up to 6809 FPS in the L-M and L-L configurations. This represents a 13-fold improvement over the baseline H-H configuration (C1) at the cost of increased resource usage. This study extends prior research on quantized neural networks (QNNs) by analyzing layer-specific parallelization strategies through adjustable folding factors and their effects on performance and resource trade-offs, offering valuable insights for optimizing FPGA-based deep learning (DL) inference in resource-constrained environments.
This manuscript presents the realization of low noise amplifier (LNA) for the sub 6 GHz applications having frequency range of 5.7 to 7 GHz with the bandwidth 1.3 GHz. The realized LNA uses the common gate topology decreasing the reflections at the input, i.e., to decrease the S11 parameter, cascode topology is used to enhance the gain of the circuit while source follower topology is used to for the output impedance matching and transformer matching is employed to further enhance the S22 parameter. The body biasing for the connected MOS is done through the RL circuit connected between body and drain. The realized LNA has a maximum gain value of 11.35 dB, minimum S11 value of -22.35 dB, minimum NF value of 2.08 dB and minimum S22 value of -10.06 dB at 6.03 GHz. The process corner simulation for the slow-slow and fast-fast corners are also done and observed that LNA parameters values are under the range. The realized LNA consumes VDD=1.8 V, with current consumption of 11.40 mA, i.e., having power consumption of 20.52 mW.
Printed circuit board (PCB) based power semiconductors are essential elements in modern electrical systems by their fast and reliable capabilities within compact miniaturized sizes. However, due to the recent trends of increased operating voltage and frequency ranges, the dielectric strength of insulation systems in PCBs has been faced a lot of challenges regarding degradation and acceleration of insulation properties, initiating partial discharges (PD). This paper proposes a PCB based microstrip patch-type PD sensor to detect high frequency PD signals generated from insulation defects. Four typical types of PCB defect models were fabricated to simulate the insulation defects within the power semiconductor. Phase-resolved partial discharge (PRPD) patterns were obtained at 120% of each partial discharge inception voltage (PDIV) level, and various statistical PD features were extracted to establish PD datasets. Four representative machine learning (ML) algorithms were comparatively analyzed, and the random forest (RF) model achieved the highest performance with an accuracy of 96.9%.
This paper presents the design and implementation of a compact FPGA-based (Field-Programmable Gate Array) controlled ultrasonic transducer system. Our aim was to achieve ultrasound beam steering using three elementary ultrasonic transducers. The transducers generated a steerable beam through ultrasound interference, which was controlled by adjusting the phase differences between them. The system's performance was demonstrated using a radar-style beamforming plot. Using elementary transmitters with an initial aperture angle of 50 degrees-130 degrees, an effective aperture angle of 74 degrees-101 degrees was achieved. The core of the device was implemented on a ZCU102 development panel, which had a considerable impact on the following development phase. Our proposed approach offers a compact and cost-effective solution for Non-Destructive Testing (NDT) systems; its applicability was demonstrated on a wood inspection example. Finally, potential directions for further development toward a fully functional device are discussed.
Accurate material recognition with low computational overhead is critical for edge applications such as autonomous drones, mobile robots, and smart manufacturing systems. Direct fine-tuning of deep backbones often leads to early saturation in validation accuracy due to overfitting on small, domain-specific datasets. To address this, we propose a structured multi-phase fine-tuning strategy for EfficientNetV2-S, progressively unfreezing layers over four stages with adaptive learning rate scheduling. The approach also incorporates label smoothing, dropout, and data augmentation to enhance generalization. We evaluated the method on a curated dataset of 1,730 images across four material classes: glass, metal, paper, and plastic. The resulting model achieves a validation accuracy of 95.66%, demonstrating that the proposed pipeline effectively balances accuracy and computational efficiency, making it suitable for real-time deployment on resource-constrained edge devices.
Cell Free (CF) model, with the advantages of improved performance, high load capacity, and uniform communication quality, is a potential solution for 5G and B5G networks. Furthermore, the combination of the CF and advanced Unmanned Aerial Vehicle (UAV) technology further enhances the overall service quality and flexibility of the system. In this work, we examine the Multi-ARS CF model, where a large number of UAVs acting as Aerial Relay Stations (ARSs) are coordinated by the Ground Base Station (GBS) and cooperate to serve a large number of Mobile Stations (MSs) within the same frequency and time resources. The communication protocol utilizes the time division duplex (TDD) mechanism, with the implementation of the Minimum Mean Square Error (MMSE) method for estimating the uplink channel. A closed-form expression for MS throughput is derived. Moreover, we propose a pilot assignment algorithm at the master ARSs focusing on minimizing "pilot contamination" in the MS-master ARS communication link. Additionally, a clustering algorithm based on pilots is implemented. The system's performance is assessed through the Cumulative Distribution Function (CDF) of the MS throughput and compared with other pilot assignment algorithms, such as random and the Greedy algorithm. The results reveal that the pilot assignment and clustering algorithm successfully address the pilot contamination issue and enhance the overall system performance.
Multiuser detection in multicarrier-code-division multiple access (MC-CDMA) systems is a critical problem, especially in a situation where the user densities are high, channel conditions are dynamically changing, and therefore, labelled data is scarce. The current paper introduces AdaptGNN, a self-supervised graph neural network (GNN) receiver that learns the MC-CDMA uplink by treating the actors as a heterogeneous graph of users and subcarriers, meaning that the interference topology is explicitly represented in the receiver. The self-supervised tasks, including masked subcarrier reconstruction, interference-edge prediction, and contrastive representation learning, facilitate the direct learning of interference-aware embeddings from the received waveforms, without the need for manually annotated labels. To improve operational robustness, a self-supervised test-time adaptation (TTA) system is integrated; the system adapts a limited set of model parameters during inference on unlabeled test examples and therefore mitigates distributional change caused by changes in user load and channel statistics. Monte Carlo simulations under Rayleigh fading conditions demonstrate that AdaptGNN significantly reduces bit-error rate (BER) and outperforms traditional multiuser detection methods, particularly in highly congested interference environments. Besides, the method reduces the latency of detection and has a high resistance to channel estimation errors compared to traditional detectors and graph-based models. These results highlight that AdaptGNN is well-positioned to serve as a scalable and efficient receiver for deployment in dense and dynamic wireless environments.
