The goal of this paper is to evaluate the performance of an adaptive beamforming approach in fifth-generation millimeter-wave multicellular networks, where massive multiple-input multiple-output configurations are employed in all active base stations of the considered orientations. In this context, beamforming is performed with the help of a predefined set of configurations that can deal with various traffic scenarios by properly generating highly directional beams on demand. In parallel, a machine learning (ML) beamforming approach based on the k-nearest neighbors (k-NN) approximation has been considered as well, which is trained in order to generate the appropriate beamforming configurations according to the spatial distribution of throughput demand. Performance is evaluated statistically, via a developed system level simulator that executes Monte Carlo simulations in parallel. Results indicate that the achievable spectral efficiency (SE) and energy efficiency (EE) values are aligned with other state of the art approaches, with reduced hardware and algorithmic complexity, since per user beamforming calculations are omitted. In particular, considering a two-tier cellular orientation, then in the non-ML approach EE and SE can reach up to 5 Mbits/J and 36 bps/Hz, respectively. Both metrics attain the aforementioned values when the ML-assisted beamforming framework is considered. However, beamforming complexity is further reduced, since the ML approach provides a direct mapping among the considered throughput demand and appropriate beamforming configuration.

The goal of this paper is the performance evaluation of a deep learning approach when deployed in fifth-generation (5G) millimeter wave (mmWave) multicellular networks. To this end, the optimum beamforming configuration is defined by two neural networks (NNs) that are properly trained, according to mean square error (MSE) minimization. The first network has as input the requested spectral efficiency (SE) per active sector, while the second network has the corresponding energy efficiency (EE). Hence, channel and power variations can now be taken into consideration during adaptive beamforming. The performance of the proposed approach is evaluated with the help of a developed system-level simulator via extensive Monte Carlo simulations. According to the presented results, machine learning (ML)-adaptive beamforming can significantly improve EE compared to the standard non-ML framework. Although this improvement comes at the cost of increased blocking probability (BP) and radiating elements (REs) for high data rate services, the corresponding increase ratios are significantly reduced compared to the EE improvement ratio. In particular, considering 21.6 Mbps per active user and ML adaptive beamforming, the EE can reach up to 5.3 Mbps/W, which is significantly improved compared to the non-ML case (0.9 Mbps/W). In this context, BP does not exceed 2.6%, which is slightly worse compared to 1.7% in the standard non-ML case. Moreover, approximately 20% additional REs are required with respect to the non-ML framework.

We present a comprehensive analysis of pulse compression in adiabatically tapered silicon photonic wire waveguides (Si-PhWWGs), both at telecom (λ ∼ 1.55 μm) and mid-IR (λ ≳ 2.1 μm) wavelengths. Our theoretical and computational study is based on a rigorous model that describes the coupled dynamics of the optical field and photogenerated free carriers, as well as the influence of the physical and geometrical parameters of the Si-PhWWGs on these dynamics. We consider both the soliton and non-soliton pulse propagation regimes, rendering the conclusions of this study relevant to a broad range of experimental settings and practical applications. In particular, we show that by engineering the linear and nonlinear optical properties of Si-PhWWGs through adiabatically varying their width, one can achieve more than 10× pulse compression in millimeter-long waveguides. The inter-dependence between the pulse characteristics and compression efficiency is also discussed.

We study the generation of parabolic self-similar optical pulses in tapered Si photonic nanowires (Si-PhNWs) at both telecom ($\lambda =1.55\mathrm{\mu m}$) and mid-infrared ($\lambda =2.2\mathrm{\mu m}$) wavelengths. Our computational study is based on a rigorous theoretical model, which fully describes the influence of linear and nonlinear optical effects on pulse propagation in Si-PhNWs with arbitrarily varying width. Numerical simulations demonstrate that, in the normal dispersion regime, optical pulses evolve naturally into parabolic pulses upon propagation in millimeter-long tapered Si-PhNWs, with the efficiency of this pulse-reshaping process being strongly dependent on the spectral and pulse parameter regime in which the device operates, as well as the particular shape of the Si-PhNWs.

We present a comprehensive analysis of the imaging characteristics of a scanning microwave microscopy (SMM) system operated in the transmission mode. In particular, we use rigorous three-dimensional finite-element simulations to investigate the effect of varying the permittivity and depth of sub-surface constituents of samples, on the scattering parameters of probes made of a metallic nano-tip attached to a cantilever. Our results prove that one can achieve enhanced imaging sensitivity in the transmission mode SMM (TM-SMM) configuration, from twofold to as much as 5× increase, as compared to that attainable in the widely used reflection mode SMM operation. In addition, we demonstrate that the phase of the S₂₁-parameter is much more sensitive to changes of the system parameters as compared to its magnitude, the scattering parameters being affected the most by variations in the conductivity of the substrate. Our analysis is validated by a good qualitative agreement between our modeling results and experimental data. These results suggest that TM-SMM systems can be used as highly efficient imaging tools with new functionalities, findings which could have important implications to the development of improved experimental imaging techniques.

Long Range wide area network (LoRaWAN) has emerged as a key enabler for Internet of Things (IoT) applications in smart cities, offering long-range connectivity with low-power consumption. This study evaluates the performance of a large-scale operational LoRaWAN network deployed in Pafos, Cyprus, consisting of over 30000 smart water meters. Leveraging twelve months of real-world data, we apply regression-based machine learning models (decision trees and XGBoost) to predict estimated signal power (ESP), enabling enhanced network reliability and efficiency. Unlike prior studies that primarily rely on simulations or small-scale testbeds, this work is the first to analyze large-scale real-world LoRaWAN data, uncovering key environmental factors influencing signal propagation. Furthermore, we introduce a novel software tool that integrates ML-driven ESP predictions, allowing for automated network optimization. Unlike conventional models, this tool dynamically refines signal power predictions based on empirical data, making it a valuable asset for smart city planning and large-scale IoT deployments. To the best of our knowledge, this study represents the first comprehensive analysis focused on the features influencing and predicting ESP. The results demonstrate that our approach effectively tailors propagation predictions to urban environments, achieving high-predictive accuracy across diverse conditions. This work contributes to the optimization of large-scale IoT deployments, paving the way for scalable and reliable smart city applications.