A 3D-printable modular robotic gripper

Robotic systems are key in industry 4.0 context. While there are many robot models available in the market, the number of grippers with sensing capabilities at an affordable cost is reduced. Traditional robot grippers are targeted to perform specific tasks, with unchangeable configurations and limited ability to adapt to different working scenarios. In this paper, we present the design and fabrication of a modular and low-cost 3D-printable robotic gripper to grasp and hold different objects. It is a parallel gripper with a two-finger two-motor configuration. The power transmission was conceived to maximize the grasping force while keeping the gripper compactness. The gripper structure is 3D-printed and includes three different types of fingers: (1) hard fingers, (2) flexible fingers, and (3) soft fingers. Objects to grasp are recognized using an embedded camera with integrated computer vision processing. The gripper components are integrated into a common network for monitoring and control. Experimental tests were conducted with objects from the YCB dataset featuring different sizes, weights and shapes. Results indicate that it can grasp different objects, reaching a maximum force of 76 N per finger and a positioning accuracy in the millimetres range.