ISyHand: A Dexterous Multi-finger Robot Hand with an Articulated Palm

TL;DR

The paper presents ISyHand, a low-cost open-source 18-DoF dexterous robot hand featuring a 2-DoF articulated palm to boost in-hand manipulation while preserving anthropomorphic finger joints. It combines off-the-shelf actuation with 3D-printed, modular components and a careful cable-routing scheme to achieve robust performance at about $1,300 in parts. The authors perform a systematic simulation study, comparing ISyHand to Allegro and LEAP in a cube reorientation task using grid-based RL evaluation, and demonstrate a significant early-learning advantage due to the articulated palm, along with a sim-to-real transfer using DeXtreme-style domain randomization. Real-world experiments with a pose-tracked cube confirm the feasibility of dexterous manipulation on the physical hardware, despite some tracking-related challenges, underscoring the approach’s practical potential and open-source accessibility.

Abstract

The rapid increase in the development of humanoid robots and customized manufacturing solutions has brought dexterous manipulation to the forefront of modern robotics. Over the past decade, several expensive dexterous hands have come to market, but advances in hardware design, particularly in servo motors and 3D printing, have recently facilitated an explosion of cheaper open-source hands. Most hands are anthropomorphic to allow use of standard human tools, and attempts to increase dexterity often sacrifice anthropomorphism. We introduce the open-source ISyHand (pronounced easy-hand), a highly dexterous, low-cost, easy-to-manufacture, on-joint servo-driven robot hand. Our hand uses off-the-shelf Dynamixel motors, fasteners, and 3D-printed parts, can be assembled within four hours, and has a total material cost of about 1,300 USD. The ISyHands's unique articulated-palm design increases overall dexterity with only a modest sacrifice in anthropomorphism. To demonstrate the utility of the articulated palm, we use reinforcement learning in simulation to train the hand to perform a classical in-hand manipulation task: cube reorientation. Our novel, systematic experiments show that the simulated ISyHand outperforms the two most comparable hands in early training phases, that all three perform similarly well after policy convergence, and that the ISyHand significantly outperforms a fixed-palm version of its own design. Additionally, we deploy a policy trained on cube reorientation on the real hand, demonstrating its ability to perform real-world dexterous manipulation.

Figures (9)

• Figure 1: The ISyHand's articulated palm facilitates grasping.
• Figure 2: Size comparison between our ISyHand (left), the Allegro hand (center), and the LEAP hand (right).
• Figure 3: Six motions that display the 18 articulated DoF of the ISyHand. (a) Three DoF in each finger allow for flexion. (b, c) Each finger is connected to the palm via a revolute joint that enables ab- and adduction. (d) Two articulated palm joints create a flexible palm. (e) The joint that connects the thumb to the palm enables thumb rotation. (f) Like the other fingers, the thumb has three DoF for flexion. While the finger flexions depicted in a and f are limited by self-contacts, the other articulations each show the full range of motion.
• Figure 4: The ISyHand performing a subset of grasps from the GRASP taxonomy feix2016grasp.
• Figure 5: Left: The ISyHand cable covers streamline routing of the three-wire motor cables, which are daisy-chained along each finger. Top right: Thumb adapter. Bottom right: Power injection and distribution.