Residential consumption represents one of the most important percentages of total electricity consumption. A considerable number of household appliances consume energy even when they are not in operation, i.e., they are in the so-called standby state, thus producing additional costs, which become significant over time. In this context, one method to solve this problem is to develop a smart system capable of severing the power connection to devices in standby mode, thereby conserving energy and reducing the energy costs. The first step in the design of this system consists of the identification and accurate measurement of the standby state, which was carried out for three of the most common household appliances. Then, by using an ESP32 microcontroller, a system was designed to manage the operation of a relay module based on the current consumption of the connected equipment. Control over the system was achieved through a web application that works across all devices equipped with a web browser, offering functionalities to adjust current value time delays and to manually switch the system on or off. Finally, the deployment of this system across the three appliances studied led to a reduction in the energy consumption in standby mode of 26.68 kWh per month.

Standby power consumption in household electrical and electronic equipment remains a persistent source of energy waste worldwide. Despite regulatory measures and ongoing technological developments, a considerable amount of electricity is still consumed by devices in standby or “off-mode”, resulting in higher utility costs and carbon emissions. This review synthesizes the latest research to clarify the scale of standby energy consumption, discusses relevant policies and regulations, and explores intelligent technologies and behavioral strategies that minimize energy consumption. Starting from the theoretical analysis and modeling of equipment consumption in standby mode to the implementation of intelligent systems to reduce it, the paper highlights heuristic optimization methods, smart grid integration, and occupant-centered interventions, all of which demonstrate tangible energy savings. This research was carried out in close connection with current policies regarding energy consumption and sustainable development, respectively, with the implementation of new technologies. Thus, in accordance with the latest European directives, the intelligent systems used have reduced the energy consumption of some common household appliances by 26.68 kWh. Additionally, knowledge gaps, particularly regarding user behavior, data granularity, and the integration of advanced analytics that limit the efficacy of current solutions, are identified. Recommendations for future research, emphasizing the importance of harmonized policies, precise data measurement, and artificial-intelligence-driven approaches for further reducing standby loads, are finally presented.

The mandatory introduction of Romania’s national e-invoicing system, ANAF e-Factura, in January 2024 has reshaped B2B document exchange in the retail sector, but suppliers still operate in parallel with two proprietary electronic data interchange (EDI) platforms, EDINET and DocProcess, which increases integration complexity. This paper presents the architecture, implementation, and evaluation of a custom Laravel-based B2B platform developed to automate commercial workflows across these three channels. The system supports XML purchase order ingestion and normalization, product identifier resolution, unified order persistence, platform-specific invoice XML generation, and ANAF SPV submission via SmartBill and Oblio REST APIs. A comparative analysis of real production XML documents showed full field-level overlap across 21 invoice data dimensions, with the main differences between systems related to entity identification schemes rather than business information content. During 2025, the platform processed 1247 EDI purchase orders and achieved an 87.30% fully automated processing rate, reaching 94.60% by year-end through progressive product catalog enrichment. The results indicate that ANAF e-Factura is technically capable of covering the core invoice exchange function currently duplicated by proprietary EDI platforms, while their coexistence continues to impose additional integration effort and slows SME digital transformation, particularly for small and medium-sized suppliers.

This paper presents the design and implementation of a Wi-Fi mesh network system for distributed monitoring of environmental parameters (temperature, humidity, atmospheric pressure, and air quality/VOC) in a residential setting. The system comprises one ROOT node acting as gateway and web server, and three SLAVE nodes for data acquisition. Built on the ESP32 platform using the painlessMesh library, the system operates fully locally without Internet dependency. A web interface accessible via captive portal enables real-time visualization, configurable alert thresholds, topology mapping, and CSV data export. Hardware integration includes BME680 sensors, OLED displays, autonomous 18650 battery power, and 3D-printed enclosures. Test results confirm coherent multi-room data differentiation and stable mesh.

Machine learning-based network intrusion detection systems are often optimized using aggregate accuracy, although operational security depends on the reliable detection of rare, high-impact attacks. This paper proposes a data-preserving intrusion detection framework that combines hybrid feature selection, heterogeneous ensemble learning and cost-sensitive optimization for imbalanced multi-class attack detection. The method first applies Mutual Information filtering and Recursive Feature Elimination to reduce the NSL-KDD feature space from 122 one-hot encoded attributes to 25 discriminative features. Four classifiers, Random Forest, XGBoost, Support Vector Machine and K-Nearest Neighbors, are evaluated individually, and a stacking ensemble is constructed using Logistic Regression as a meta-learner. Class imbalance is addressed by balanced class weighting rather than by synthetic oversampling, preserving the original minority-class observations. Experiments on the NSL-KDD benchmark show that the proposed cost-sensitive configuration improves rare attack recognition, most notably increasing U2R recall from 0.00% to 35.82% (24 of 67 test instances) for the stacking ensemble; this improvement, together with the accompanying weighted F1-score change from 0.7120 to 0.7214, is statistically significant under the Wilcoxon signed-rank test across repeated random seeds, and both values are reported with their variability rather than as single point estimates. SVM obtains the largest global gain, with a 7.06 percentage point improvement in weighted F1-score. The results show that cost-sensitive learning is a simple and practical mechanism for improving rare-attack visibility, but also reveal a remaining limitation for R2L detection, where feature overlap with Normal traffic remains substantial. The revised validation design explicitly includes direct resampling baselines, repeated-seed evaluation, statistical significance testing, feature-subset sensitivity analysis, and absolute true-positive counts for R2L and U2R in order to avoid overinterpreting marginal point-estimate gains. All experiments, including the resampling comparison, the component ablation, the feature-subset sensitivity analysis and the repeated-seed statistical evaluation, are executed on the complete KDDTrain+ training set of 125,973 instances under a single unified protocol, so that every reported per-class value refers to the same experimental setting. The revised study additionally reports probability-level evaluation for the primary model, including class-level PR-AUC, precision-recall curves and a U2R threshold and alert-budget analysis, and validates the framework externally on the UNSW-NB15 benchmark, where balanced class weighting raises the recall of the rarest categories (Worms, Shellcode, Backdoor) from near-zero baseline levels to 69–96% under an identical protocol.