Robotic gripper design to handle an arbitrarily shaped object by emulating human finger motion

Controlling a robotic gripper to handle objects in different sizes and shapes in real-time is a complex task. The challenge is to adapt and position the gripper according to a target object. In this work, by considering a three fingered robotic gripper with a finger structure similar to that of human finger, we study the positioning of fingers in a gripper in the two dimensional plane. Using an inverse kinematic model for a pulley based underactuated mechanical finger, we derive a region of reach in the plane of the finger. Simulations demonstrate the limits of reach under different conditions. This can be used to derive the positioning of each finger, as well as in path planning to control the reach of the gripper.