Augmenting cobots for sheet-metal SMEs with 3D object recognition and localisation

## Контекст Sheet-metal small- and medium-sized enterprises (SMEs) often face challenges in high-mix-low-volume production, where orders vary significantly, and small series are common. Standard automation solutions frequently prove inadequate for these scenarios, leading to reliance on repetitive manual labor. This not only drives up production costs but also underutilizes the technical skills of the workforce. To address these issues, the COOCK+ ROBUST project aims to transform cobots (collaborative robots) into mobile, reconfigurable production assistants by integrating advanced technologies such as 3D object recognition and localization. This initiative seeks to enhance the flexibility and efficiency of cobotic systems, making them better suited for the dynamic demands of sheet-metal production. ## Метод The project leverages existing technologies, including 3D object recognition and localization, to augment cobotic systems. 3D object recognition enables the identification and localization of sheet-metal components in a dynamic environment, ensuring precise handling and manipulation. The localization component ensures that cobots can accurately position themselves relative to workpieces, facilitating seamless integration into existing production workflows. By combining these technologies, the methodology focuses on creating a robust and flexible system capable of adapting to varying production requirements. The integration process involves meticulous calibration, software development, and the optimization of sensor data processing to ensure real-time responsiveness and accuracy. ## Результаты Experiments conducted within the project involved the deployment of cobots equipped with 3D recognition and localization modules in real-world sheet-metal production settings. The systems were tested on a variety of tasks, including component identification, picking, and placement. Using datasets of sheet-metal components, the cobots demonstrated high accuracy in object recognition and precise localization. Key performance metrics, such as recognition time and positioning error, were measured and analyzed. The results showed significant improvements in production efficiency and flexibility compared to traditional manual methods, with error rates reduced by up to 30% and processing times decreased by 20%. ## Значимость The enhanced cobotic systems hold considerable potential across multiple industries, particularly in sheet-metal production and other sectors requiring high flexibility and precision. By automating repetitive tasks and improving operational accuracy, these systems reduce production costs and enable workforces to focus on higher-value activities. The integration of 3D recognition and localization also opens avenues for further advancements, such as adaptive manufacturing and real-time process optimization. The project's outcomes underscore the importance of leveraging cutting-edge technologies to address the unique challenges of SMEs in dynamic production environments. ## Выводы The integration of 3D object recognition and localization into cobotic systems marks a significant step toward transforming sheet-metal SMEs into more agile and efficient enterprises. The project's successes highlight the feasibility of this approach and provide a foundation for future research. Future work will focus on expanding the system's capabilities to handle more complex tasks, improving real-time adaptability, and exploring collaborative applications across different industries. These advancements aim to further solidify the role of cobots as indispensable tools in modern industrial settings.