Table of Contents
Fetching ...

Replicating Human Anatomy with Vision Controlled Jetting -- A Pneumatic Musculoskeletal Hand and Forearm

Thomas Buchner, Stefan Weirich, Alexander M. Kübler, Wojciech Matusik, Robert K. Katzschmann

TL;DR

The paper addresses achieving human-like dexterity in a robotic hand by combining a multi-material 3D-printed, soft–rigid hybrid hand and forearm with 22 pneumatic artificial muscles actuating a tendon-driven mechanism. It leverages vision-controlled jetting to print an integrated system, including soft joint capsules and tactile sensors, and demonstrates grasping and manipulation of objects up to $272$ g with independent finger movement. Key results include range of motion metrics (Kapandji score of 6/10), a variety of grasps (power, precision, intermediate), and fingertip/grasps forces of about $1.95\,\text{N}$ and $2.97\,\text{N}$ under specific lead pressures, with maximum PAM strain of $30.1\%$ and force around $38$ N at $0.5$ MPa. The work provides a cost-effective, rapid fabrication pathway for biomimetic hands and forearms, offering practical implications for prosthetics and robot manipulation, while identifying durability and mobility challenges to be addressed in future work.

Abstract

The functional replication and actuation of complex structures inspired by nature is a longstanding goal for humanity. Creating such complex structures combining soft and rigid features and actuating them with artificial muscles would further our understanding of natural kinematic structures. We printed a biomimetic hand in a single print process comprised of a rigid skeleton, soft joint capsules, tendons, and printed touch sensors. We showed it's actuation using electric motors. In this work, we expand on this work by adding a forearm that is also closely modeled after the human anatomy and replacing the hand's motors with 22 independently controlled pneumatic artificial muscles (PAMs). Our thin, high-strain (up to 30.1%) PAMs match the performance of state-of-the-art artificial muscles at a lower cost. The system showcases human-like dexterity with independent finger movements, demonstrating successful grasping of various objects, ranging from a small, lightweight coin to a large can of 272g in weight. The performance evaluation, based on fingertip and grasping forces along with finger joint range of motion, highlights the system's potential.

Replicating Human Anatomy with Vision Controlled Jetting -- A Pneumatic Musculoskeletal Hand and Forearm

TL;DR

The paper addresses achieving human-like dexterity in a robotic hand by combining a multi-material 3D-printed, soft–rigid hybrid hand and forearm with 22 pneumatic artificial muscles actuating a tendon-driven mechanism. It leverages vision-controlled jetting to print an integrated system, including soft joint capsules and tactile sensors, and demonstrates grasping and manipulation of objects up to g with independent finger movement. Key results include range of motion metrics (Kapandji score of 6/10), a variety of grasps (power, precision, intermediate), and fingertip/grasps forces of about and under specific lead pressures, with maximum PAM strain of and force around N at MPa. The work provides a cost-effective, rapid fabrication pathway for biomimetic hands and forearms, offering practical implications for prosthetics and robot manipulation, while identifying durability and mobility challenges to be addressed in future work.

Abstract

The functional replication and actuation of complex structures inspired by nature is a longstanding goal for humanity. Creating such complex structures combining soft and rigid features and actuating them with artificial muscles would further our understanding of natural kinematic structures. We printed a biomimetic hand in a single print process comprised of a rigid skeleton, soft joint capsules, tendons, and printed touch sensors. We showed it's actuation using electric motors. In this work, we expand on this work by adding a forearm that is also closely modeled after the human anatomy and replacing the hand's motors with 22 independently controlled pneumatic artificial muscles (PAMs). Our thin, high-strain (up to 30.1%) PAMs match the performance of state-of-the-art artificial muscles at a lower cost. The system showcases human-like dexterity with independent finger movements, demonstrating successful grasping of various objects, ranging from a small, lightweight coin to a large can of 272g in weight. The performance evaluation, based on fingertip and grasping forces along with finger joint range of motion, highlights the system's potential.
Paper Structure (16 sections, 8 figures, 3 tables)

This paper contains 16 sections, 8 figures, 3 tables.

Figures (8)

  • Figure 1: Pneumatic musculoskeletal hand and forearm. Inspired by the human anatomy, we printed a musculoskeletal hand and forearm. The soft-rigid hybrid hand is printed in a single process and does not need assembly apart from attaching thin McKibben muscles for actuation. We replicated the anchor points and muscle lengths from the biological inspiration.
  • Figure 2: Inspiration from human hand and forearm replicated in a 3D print.A Front view of a 3D bone model of a human arm and hand with accurate muscle placement. B Front view of a 3D model of the printed hand and forearm. The bone shapes and positions were extracted from MRI data. The tendons and actuators replacing the muscles are placed according to their biological counterparts' anchor points and positions. C Back view of model shown in sub-figure A. D Back view of model shown in sub-figure B. The sperical volumes shown around the joints are our modeled joint capsules/ligaments.
  • Figure 3: Materials and step-by-step illustration of the manual fabrication process for the PAM.A Overview of the used materials and the insertion of the tube into the braided sleeving. B Manufacturing steps for the end terminal. C Manufacturing steps for the end part with the air supply inlet.
  • Figure 4: Experimental results for the SRL PAM.A The SRL PAMs used in the printed hand and forearm are tested in a force strain test setup. A weight is applied to simulate forces. Displacements can be recorded for different pressures applied to the PAM. B Force strain curves for different pressures applied to the PAMs. The error indicators represent an approximated measurement inaccuracy corresponding to 2 stroke.
  • Figure 5: Pneumatic touch sensor.A Schematic of a pressure sensor pad with tube connection. B Photo of the printed functional pressure sensor pad with tubing. C Sensor pressure over time showing several contacts with the sensor pad.
  • ...and 3 more figures