Coal remains a vital energy resource and plays a key role in national development. Ensuring the safety of underground mining personnel is essential, and intelligent algorithms are increasingly used to detect miners in surveillance footage. However, complex underground environments—characterised by poor lighting, occlusions, irregular postures, and reflective gear—make accurate detection difficult. This study proposes improvements to the YOLOv10-N object detection model for miner detection. Using 37,463 annotated images from real mining environments, we propose three main enhancements: a Coordinate Attention (CA) mechanism to highlight important spatial features, a Dynamic Head (DyHead) module to improve multi-scale feature fusion, and the Efficient IoU (EIOU) loss function to enhance bounding box regression and speed up convergence. While CA, DyHead, and EIOU are established methods, their synergistic integration for asymmetric miner detection (e.g., occluded limbs, uneven lighting) presents a novel application-specific optimisation. Experimental results confirm that the enhanced model significantly outperforms the original. It achieves 92.69% accuracy, 87.53% recall, and an average accuracy of 89.9%, with a practical detection effect of 68.24%. These findings show that the proposed method improves both accuracy and robustness in challenging mining conditions while maintaining processing efficiency.

This paper focuses on three different aspects of a multi-agent approach implementation applied in smart manufacturing. Information systems and their internals are considered as a domain of the research. Among a variety of approaches and implementation techniques the following and presented: networks of self-organising heterogeneous agents are used to address problems of increasing server loads, AI techniques (POMDPs) are used to ensure highest possible availability via integrated reasoning, special optimization routines are designed to increase performance of historical access to time series data. Described approaches aim on advanced scalability, high fault-tolerance and low latency execution of a resulting system respectively. Each technique defines a mathematical model that is used to optimize corresponding criteria. For each technique implementation peculiarities and details are included

In industrial scenarios, bearing fault diagnosis often suffers from data scarcity and class imbalance, which significantly hinders the generalization performance of data-driven models. While generative adversarial networks (GANs) have shown promise in data augmentation, their efficacy deteriorates in the presence of multi-category and structurally complex fault distributions. To address these challenges, this paper proposes a novel fault diagnosis framework based on a Multi-Domain Feature Transformer GAN (MDFT-GAN). Specifically, raw vibration signals are transformed into 2D RGB representations via joint time-domain, frequency-domain, and time–frequency-domain mappings, effectively encoding multi-perspective fault signatures. A Transformer-based feature extractor, integrated with Efficient Channel Attention (ECA), is embedded into both the generator and discriminator to capture global dependencies and channel-wise interactions, thereby enhancing the representation quality of synthetic samples. Furthermore, a gradient penalty (GP) term is introduced to stabilize adversarial training and suppress mode collapse. To improve classification performance, an Enhanced Hybrid Visual Transformer (EH-ViT) is constructed by coupling a lightweight convolutional stem with a ViT encoder, enabling robust and discriminative fault identification. Beyond performance metrics, this work also incorporates a Grad-CAM-based interpretability scheme to visualize hierarchical feature activation patterns within the discriminator, providing transparent insight into the model’s decision-making rationale across different fault types. Extensive experiments on the CWRU and Jiangnan University (JNU) bearing datasets validate that the proposed method achieves superior diagnostic accuracy, robustness under limited and imbalanced conditions, and enhanced interpretability compared to existing state-of-the-art approaches.

Nowadays more and more attention is given to energy development problems. Modern ways of producing energy are bounded by planet’s resources of coal, natural gas and petroleum, and harmful effects of waste produced by energy industry forced the increasing growth of alternate ways of energy development, such as wind farms, geothermal and solar energy, etc. The article considers a solution the problem of creating energy technologies for autonomous decentralized energy supply, using intelligent automated control systems.

Digitalisation is currently developing in many sectors of the economy. The success of a manufacturing enterprise requires a transition to digital control of production and business processes, to the greatest extent possible without human intervention using artificial intelligence technologies. This research applies a Manufacturing Execution System (MES) with Predictive Analysis to an automatic production line (Chemical Line) which contains sensors, actuators and pumps controlled by four Programmable Logic Controllers, linked together, and being monitored through a Supervisory Control and Data Acquisition system. The production line composed of four different subsystems, responsible for filtration to bottling the chemical product. This paper tries to join the MES system with Artificial Neural Networks (ANN) in order to not only monitor the system but having predictive analysis to plan the future. In such way we will take advantage of the benefits of the ANN, such as Long-Short Term Memory architectures. The experimental data will be compared with other usual platforms, such as the Master SCADA itself, through the course of this research.