Robotic pick-and-handover maneuvers with camera-based intelligent object detection and impedance control

Abstract

A vision-based impedance control method is applied to a 7-degree-of-freedom Franka Emika (FE) Panda robotic manipulator to complete pick-and-handover tasks with human-like grasping of fruits. The interfaces between different hardware, controllers, and path planners are developed using robot operating systems (ROS). An RGB camera is applied using the YOLOv5 (You Only Look Once v5) object detection algorithm, which is trained for various fruits and human hands and then employed for identifying the location of the targeted objects. A qb-SoftHand robotic hand is used as the end-effector for the grasping tasks. By integrating these components together with ROS, the FE Panda robot successfully achieves autonomous human-like handover tasks, as shown in the experimental results.