Vision-Based Robot Hand Using Open Source Software

摘要

The robot hand can plan grasp movements based on the position of the target object and the motion of the robot hand. The position of the target object is recognized from the image captured by a camera mounted on the robot arm, and the motion of the robot hand is estimated from the inertial measurement unit (IMU). We also adopt a variety of mechanisms to determine the target object among the objects detected in the camera scene. Experiments are conducted to verify the validity of control system. The experiments proved that the developed system can support the user to grasp the target object.