Effective Human–Robot Interaction Through Motion Retargeting Approach

While motion retargeting is a task that humans typically find easy, it can be challenging for robots. To address this issue and enhance social well-being, this paper proposes a bioinspired motion retargeting design method for robots. The method involves mapping human motion to a robot with a different structural design. The paper provides simulations of human–robot interaction (HRI) and discusses the challenges and solutions involved in motion retargeting. Various methods such as trajectory optimization, human pose estimation, dynamic simulation, and retargeting method have been proposed in paper for effective human–robot interaction. To estimate human poses more quickly, MediaPipe is employed. The MediaPipe Pose (MPP) framework, which is open-source and cross-platform, has been used to calculate the 2D and 3D joint coordinates of humans in robotic systems.