Enhancing Human-Robot Collaboration through 5G-Enabled Transformation and Image-Guided Control with TYOLOV5 Algorithms

Factories that prioritize digital transformation experience heightened production rates and gain a competitive advantage by enhancing control and operational efficiency. In this study, image processing techniques were employed to achieve digital transformation objectives, and the data obtained was transmitted via 5G communication to a collaborative robot arm, enabling remote control. The robot arm was fitted with a 3D-printed humanoid hand for bin picking, where each finger was controlled by five servo motors. To manipulate the fingers, a user donned a glove equipped with flex sensors, which conveyed the finger positions to the servo motors, facilitating precise pick and place operations. Additionally, image processing was utilized to accurately control the robot arm’s position. The research involved the implementation of tuned YOLO V5 (TYOLOV5) algorithms, using Python software, for object detection within the user’s gloves. By comparing this approach with other existing methods, the experiment achieved the highest detection accuracy, demonstrating the effectiveness of the proposed approach in achieving superior recognition and estimation compared to traditional and existing deep learning methods. To validate its efficacy, experiments were conducted in a laboratory setting, where a collaborative robot was deployed for a small part assembly task. The results conclusively highlight the potential of this approach to significantly enhance recognition and estimation capabilities, especially in the realm of human-robot collaboration.