This paper proposes a method for extracting information from the parameters of a single point incremental forming (SPIF) process. The measurement of the forming force using this technology helps to avoid failures, identify optimal processes, and to implement routine control. Since forming forces are also dependent on the friction between the tool and the sheet metal, an innovative solution has been proposed to actively control the friction forces by modulating the vibrations that replace the environmentally unfriendly lubrication of contact surfaces. This study focuses on the influence of mechanical properties, process parameters and sheet thickness on the maximum forming force. Artificial Neural Network (ANN) and different machine learning (ML) algorithms have been applied to develop an efficient force prediction model. The predicted forces agreed reasonably well with the experimental results. Assuming that the variability of each input function is characterized by a normal distribution, sampling data were generated. The applicability of the models in an industrial environment is due to their relatively high performance and the ability to balance model bias and variance. The results indicate that ANN and Gaussian process regression (GPR) have been identified as the most efficient methods for developing forming force prediction models.

The technological maturity of AI solutions has been consistently increasing over the years, expanding its application scope and domains. Smart home systems have evolved to act as proactive assistants for their residents, autonomously detecting behavioral patterns, inferring needs, and making decisions pertaining to the management and control of various home subsystems. The implementation of explainable AI (XAI) solutions in this challenging domain can improve user experience and trust by providing clear and understandable explanations of the system’s behavior. The article discusses the increasing importance of explainable artificial intelligence (XAI) in smart home systems, which are becoming progressively smarter and more accessible to end-users, and presents an agent-based approach for developing explainable Internet of things (IoT) systems and an experiment conducted at the Centre of Real Time Computer Systems at the Kaunas University of Technology. The proposed method was adapted to build an explainable, rule-based smart home system for controlling light, heating, and ventilation. The results of this study serve as a demonstration of the feasibility and effectiveness of the proposed theoretical approach in real-world scenarios.

Tax authorities face the challenge of effectively identifying companies that avoid paying taxes, which is not unique to European Union countries. Limited resources often constrain tax administrators, who traditionally rely on time-consuming and labour-intensive tax audit tools. As a result of this established practice, governments are losing a lot of tax revenue. The main objective of this study is to increase the efficiency of the detection of tax evasion by applying data mining methods in the East European country Lithuania, which has a rapidly developing economy, by applying data mining methods concerning affluence-related impacts. The study develops various models for segmentation, risk assessment, behavioral templates, and tax crime detection. Results show that the data mining technique can effectively detect tax evasion and extract hidden knowledge that can be used to reduce revenue losses resulting from tax evasion. This study’s methods, software, and findings can assist decision-makers, experts, and scientists in developing countries in predicting tax fraud detection.

The safety of ultrasound exposure is very important for a patient’s well-being. High-frequency (1–10 MHz) ultrasound waves are highly absorbed by biological tissue and have limited therapeutic effects on internal organs. This article presents the results of the development and application of a low-frequency (20–100 kHz) ultrasonic transducer for sonication of biological tissues. Using the methodology of digital twins, consisting of virtual and physical twins, an ultrasonic transducer has been developed that emits a focused ultrasound signal that penetrates into deeper biological tissues. For this purpose, the ring-shaped end surface of this transducer is excited not only by the main longitudinal vibrational mode, which is typical of the flat end surface transducers used to date, but also by higher mode radial vibrations. The virtual twin simulation shows that the acoustic signal emitted by the ring-shaped transducer, which is excited by a higher vibrational mode, is concentrated into a narrower and more precise acoustic wave that penetrates deeper into the biological tissue and affects only the part of the body to be treated, but not the whole body.

This paper focuses on several aspects extending the dynamical efficiency of a cantilever beam vibrating in the third mode. A few ways of producing this mode stimulation, namely vibro-impact or forced excitation, as well as its application for energy harvesting devices are proposed. The paper presents numerical and experimental analyses of novel structural dynamics effects along with an optimal configuration of the cantilever beam. The peculiarities of a cantilever beam vibrating in the third mode are related to the significant increase of the level of deformations capable of extracting significant additional amounts of energy compared to the conventional harvester vibrating in the first mode. Two types of a piezoelectric vibrating energy harvester (PVEH) prototype are analysed in this paper: the first one without electrode segmentation, while the second is segmented using electrode segmentation at the strain nodes of the third vibration mode to achieve effective operation at the third resonant frequency. The results of this research revealed that the voltage generated by any segment of the segmented PVEH prototype excited at the third resonant frequency demonstrated a 3.4–4.8-fold increase in comparison with the non-segmented prototype. Simultaneously, the efficiency of the energy harvester prototype also increased at lower resonant frequencies from 16% to 90%. The insights presented in the paper may serve for the development and fabrication of advanced piezoelectric energy harvesters which would be able to generate a considerably increased amount of electrical energy independently of the frequency of kinematical excitation.

Digital technologies that continuously quantify animal behavior, physiology, and production offer significant potential for the early identification of health and welfare disorders of dairy cows. In this study, a multimodal artificial intelligence (AI) framework is proposed for real-time health monitoring of dairy cows through the integration of physiological, behavioral, production, and thermal imaging data, targeting veterinarian-confirmed udder, leg, and hoof infections. Predictions are generated at the cow-day level by aggregating multimodal measurements collected during daily milking events. The dataset comprised 88 lactating cows, including veterinarian-confirmed udder, leg, and hoof infections grouped under a single ‘sick’ label. To prevent information leakage, model evaluation was performed using a cow-level data split, ensuring that data from the same animal did not appear in both training and testing sets. The system is designed to detect early deviations from normal health trajectories prior to the appearance of overt clinical symptoms. All measurements, with the exception of the intra-ruminal bolus sensor, were obtained non-invasively within a commercial dairy farm equipped with automated milking and monitoring infrastructure. A key novelty of this work is the simultaneous integration of data from three independent sources: an automated milking system, a thermal imaging camera, and an intra-ruminal bolus sensor. A hybrid deep learning architecture is introduced that combines the core components of established models, including U-Net, O-Net, and ResNet, to exploit their complementary strengths for the analysis of dairy cow health states. The proposed multimodal approach achieved an overall accuracy of 91.62% and an AUC of 0.94 and improved classification performance by up to 3% compared with single-modality models, demonstrating enhanced robustness and sensitivity to early-stage disease.