Compared to the past, robots are now widely used in industrial settings as well as in daily life. However, adapting all work environments to accommodate robots introduces challenges in terms of time, space, and cost. Consequently, research on Human-Ro...
Compared to the past, robots are now widely used in industrial settings as well as in daily life. However, adapting all work environments to accommodate robots introduces challenges in terms of time, space, and cost. Consequently, research on Human-Robot Collaboration (HRC), where humans and robots coexist and perform the same tasks simultaneously, has been actively pursued. Among these studies, Co-transportation has received significant attention. Key considerations in Co-transportation include understanding human intent what task the human wants to perform and how ensuring the safety of both humans and the environment, maintaining work efficiency, and enabling obstacle avoidance to build trust during collaboration. Most Co-transportation studies have focused on rigid objects, for which existing haptic-based methods can easily infer human intent. In contrast, deformable objects disperse forces due to their deformation, making it difficult to determine human intent using traditional haptic approaches. Therefore, alternative methods are required to transport deformable objects effectively. In this Thesis, introduces a study in which humans and robots collaborate to transport deformable objects while avoiding obstacles and performing tasks. Unlike traditional approaches, the proposed method uses a vision sensor to infer human intent and employs Quadratic Programming (QP) as an optimization framework to enable the robot to reach desired target positions while simultaneously avoiding obstacles. This allows the execution of HRC in which both the human and the robot can cooperatively transport deformable objects.