Abstract
This paper introduces an innovative framework for Learn from Demonstration applied to collaborative robots, leveraging the advantages of Mediapipe, a computer vision tool in conjunction with a camera system. Given the collaborative nature, robots of this type are well-suited for human-robot interaction scenarios, and our proposed approach enhances their adaptability and learning capabilities through visual demonstration. Within the Robot Operating System environment, the collaborative robot equipped with a camera system serves as a standardized platform for control and communication. The integration of Mediapipe enhances the system’s visual perception and comprehension, empowering it with robust capabilities for Learning from Demonstration. This research extends the realm of collaborative robotics by introducing a unified framework that capitalizes on the synergies among Mediapipe, Robot Operating System, camera systems, and cobots. The proposed method shows potential for applications demanding flexible and intuitive human-robot cooperation in real-world scenarios.
