George V. Tsoulos

As bandwidth-intensive applications proliferate and the usage of wireless devices surges, fifth-generation (5G) and beyond (B5G) networks are challenged to enhance coverage, reduce latency, and improve efficiency. The application of machine learning (ML) models for received signal strength (RSS) estimation is a powerful tool. This study evaluates various ML models—categorical boosting (CatBoost), extreme randomized trees (ETs), light gradient boosting machine (LGBM), and extreme gradient boosting (XGBoost)—for effective estimation of RSS. Additionally, we apply explainable artificial intelligence (XAI) methodologies, especially the Shapley additive explanation (SHAP) framework. Our investigation reveals the sophisticated mechanisms within these models, notably highlighting the exceptional accuracy of the ET model. We further introduce SHAPRP, in which SHAP attributions reduce the input space and sparse regression selects a compact subset of the ET ensemble. The results are obtained from a single-operator rural/semi-rural campaign and constitute a case study, where the trained estimator is deployment-specific and is not a pre-trained model applicable to arbitrary 5G/B5G scenarios, so what transfers is SHAPRP itself.

Smart-building systems need reliable sensing, long-term storage, and remote control, but many studies mix up fast network response with slow physical changes in the building. We report measurements from an openHAB deployment at University Lab 1 and University Lab 2 in Greece over 19 months (April 2024–October 2025), with Z-Wave and ZigBee devices and Parquet exports for offline analysis. We treat sensing availability, Message Queuing Telemetry Transport (MQTT) command latency, and heating, ventilation, and air conditioning (HVAC) behavior as separate questions. In March–June 2024, raw record-level temperature field availability was 79.7% and 74.2% at University Lab 1 and University Lab 2, respectively, increasing to 99.7% and 98.5% after the <30 min linear field interpolation used in golden-dataset construction; assessable device-month cadence-normalized reading-count ratios were far lower (7.9% temperature and 13.0% humidity under the 300 s assumption at University Lab 1; humidity 7.8–26.0% across 180–600 s), reflecting heterogeneous archival participation rather than a fully populated expected-cadence denominator over the full archive span. In May–June 2024, 1247 MQTT commands yielded 1246 successful correlated responses; the archived application log contained no repeated correlation identifiers or duplicate response records (broker DUP flags are not archived); median, p90, and p99 latencies were 287 ms, 487 ms, and 1.66 s. These times describe the MQTT–edge-agent–openHAB path only, not physical device action. HVAC figures are illustrative; we do not infer settling times. The results show why availability metrics, archive denominators, and response-time boundaries must be defined separately in smart-building evaluations.

Despite the promising benefits, the integration of 5G, particularly through Non-Standalone (NSA) architectures that rely on existing 4G infrastructures, introduces challenges in maintaining optimal radio network performance and service quality. This study evaluates 4G and 5G radio network performance through empirical field trials across highway operational scenarios, a domain less scrutinized compared with urban environments. By conducting simultaneous measurements and comparing all available networks, this research focuses on dissecting the performance of critical quality indicators to gauge the intricacies of radio network behavior in high-speed travel conditions. Our findings illuminate distinct behavioral patterns, highlighting the unique challenges and optimization opportunities in these scenarios. Variability in signal strength across the highway routes underscores the influence of geography and infrastructure on coverage, while the relative stability in signal quality suggests the networks’ capability to maintain signal quality amid fluctuating strength. Interference results indicate effective management of signal interference, crucial for high-quality links, whereas latency and throughput metrics highlight a lag behind anticipated goals for reduced latency but promising data rates. This study not only showcases the variances in network quality and performance, thus pinpointing areas for operator-specific enhancements, but also emphasizes the comparison between the robustness of 4G infrastructures and the challenges in optimizing 5G networks.

There was an error in the original publication

This paper presents the design and implementation of a versatile IoT testbed utilizing the openHAB platform, along with various wireless interfaces, including Z-Wave, ZigBee, Wi-Fi, 4G-LTE (Long-Term Evolution), and IR (Infrared Radiation), and an array of sensors for motion, temperature, luminance, humidity, vibration, UV (ultraviolet), and energy consumption. First, the testbed architecture, setup, basic testing, and collected data results are described. Then, by showcasing a typical day in the laboratory, we illustrate the testbed’s potential through the collection and analysis of data from multiple sensors. The study also explores the capabilities of the openHAB platform, including its robust persistence layer, event management, real-time monitoring, and customization. The significance of the testbed in enhancing data collection methodologies for energy assets and unlocking new possibilities in the realm of IoT technologies is particularly highlighted.

This paper presents the design and implementation of a versatile IoT testbed utilizing the openHab platform along with various wireless interfaces, including Z-Wave, ZigBee, WiFi, 4G-LTE, and IR, and an array of sensors for motion, temperature, luminance, humidity, vibration, UV, and energy consumption. First, the testbed architecture, setup, basic testing, and collected data results are described. Then, by showcasing a typical day in the laboratory, we illustrate the testbed's potential through the collection and analysis of data from multiple sensors. The study also explores the capabilities of the openHab platform, including its robust persistence layer, event management, real-time monitoring, and customization. The significance of the testbed in enhancing data-collection methodologies for energy assets and unlocking new possibilities in the realm of IoT technologies is particularly highlighted.

The usage of unmanned aerial vehicles (UAVs) as flying base stations (FBSs) for expanding coverage and assisting the terrestrial cellular networks constitutes a promising technology for 5G and beyond. A crucial parameter affecting cellular network design is path loss prediction. An alternative to the accurate, though time-consuming, propagation prediction with deterministic Ray Tracing models could be Machine Learning (ML) based predictions. Ensemble Learning techniques are used in order to optimally combine the predictions of standalone models. That is, they combine the best-performing individual models into a better-performing meta-model. Our proposed method of the evolutionary tuned stacked ensemble optimizes the ensemble as a whole, instead of optimizing its individual base learners. To the best of our knowledge, this is the first time that an evolutionary technique is applied in order to mutually tune an ensemble’s base learners for a path loss modeling problem in electromagnetics. Moreover, we present a model that works in more than one frequency. As opposed to the standard implementation of ensemble learning, our method offers a significant performance boost with low complexity.

Building Information Modeling (BIM) is a critical element for the “digitalization” of the construction industry and can be exploited for energy-driven renovation procedures of existing residences. Advancing beyond a BIM with data-capturing capabilities that are limited to building static information only requires sensor data streams related to indoor/outdoor ambient conditions, as well as to energy-consumption parameters of the residences. The data streams require the deployment of robust Wireless Sensor Networks (WSNs) that are able to capture and transmit real-time data to appropriate cloud-based renovation toolkits. The technology and topology of such networks are addressed herein. The paper sets the lines for similar installations that are required by the construction industry for collecting dynamic data, since it is based on the outcome of real-world WSN installations in pilot sites in three European countries, carried out in the context of a major collaborative BIM research project. An application example of the WSN data is also provided in the context of training occupant behavior models in order to demonstrate the use of the measured data.