This work focuses on detecting defects in welding seams using the most advanced You Only Look Once (YOLO) algorithms and transfer learning. To this end, the authors prepared a small dataset of images using manual welding and compared the performance of the YOLO v5, v6, v7, and v8 methods after two‐step training. Key findings reveal that YOLOv7 demonstrates superior performance, suggesting its potential as a valuable tool in automated welding quality control. The authors’ research underscores the importance of model selection. It lays the groundwork for future exploration in larger datasets and varied welding scenarios, potentially contributing to defect detection practices in manufacturing industries. The dataset and the code repository links are also provided to support our findings.
This study evaluates five control methods for the Fischertechnik Training Factory Industry 4.0–24 V system, focusing on accuracy, efficiency, and performance in managing voltage differences and encoder pulse detection. The proposed method introduces a novel approach by integrating a dedicated pulse-counting Arduino with the TBD62783APG chip, enhancing precision and reliability beyond traditional methods. Despite requiring additional hardware, this method demonstrated exceptional control and dependable operation, showcasing the efficacy of specialized integrated circuits in industrial automation applications. The presented evaluation enhances control strategies for industrial automation, emphasizing efficient voltage conversion and precise encoder pulse detection to enhance system performance and reliability, laying the groundwork for future advancements in remote control capabilities and complex manufacturing systems.
Implicit neural networks and the related deep equilibrium models are investigated. To train these networks, the gradient of the corresponding loss function should be computed. Bypassing the implicit function theorem, we develop an explicit representation of this quantity, which leads to an easily accessible computational algorithm. The theoretical findings are also supported by numerical simulations.
Operating a factory near a war zone or in a country with an unstable political environment poses significant risks. These risks can be mitigated by managing production lines remotely. To preserve technical knowledge and specific algorithms, it is essential to implement secure, precise, and efficient remote control in industrial applications, especially considering the risks associated with storing operational code on controllers. To address these challenges, we propose a novel technique where executable commands are dynamically transmitted from a Python script to an ESP32-WROOM-32 microcontroller. Unlike conventional methods that preload code onto the controller, this approach interprets, executes, and erases commands immediately after execution, thereby enhancing security and precision. The proposed system was evaluated through a comparative analysis of eleven distinct methods, which varied in command transmission strategies and employed dual-core processing for performance optimization. VPN technology was integrated to enable remote control from geographically distant locations, demonstrating the system’s adaptability for global industrial operations. The results indicate that this method significantly outperforms the traditional on-site approach, where operational code is preloaded onto the microcontroller. Specifically, bundled command transmission combined with dual-core processing proved particularly effective, reducing latency and improving reliability. These findings highlight the robustness of the proposed approach as a secure and flexible solution for modern smart factories.