Design, fabrication, modeling and control of a fabric-based spherical robotic arm
Open access
Date
2020-06Type
- Journal Article
ETH Bibliography
yes
Altmetrics
Abstract
In this paper we present a spherical soft robotic arm made from fabric. The inflatable arm has a small mass and is pneumatically actuated. A configuration is employed with only three actuators controlling the two rotational degrees of freedom of a spherical joint. This differs from the commonly employed antagonistic pairs, including four actuators for two degrees of freedom. The fabrication procedure of the lightweight and compliant system is discussed in detail and uses commonly available materials and tools. The capability of the robotic arm to adjust the joint stiffness as a function of the actuator pressures is investigated and characterized for different deflection directions. The static mapping from the actuator pressures to the orientation of the robotic arm is identified from data and the inverse mapping is employed in a position controller. The modeling and controller derivation are performed for three different stiffness levels demonstrating the ability of the spherical robotic arm to change the joint stiffness independently of controlling the position. The position tracking performance is experimentally evaluated by tracking a square trajectory. A comparison of the tracking performance for the different stiffness levels shows that accurate tracking is more challenging for the smallest joint stiffness. A gray-box model capturing the interactions of the two degrees of freedom is used in a learning scheme that is applied for the smallest stiffness level. The learning approach reduces interactions between the two degrees of freedom and demonstrates the control performance achievable with the system developed. Show more
Permanent link
https://doi.org/10.3929/ethz-b-000465615Publication status
publishedExternal links
Journal / series
MechatronicsVolume
Pages / Article No.
Publisher
PergamonSubject
Soft robotics; Pneumatic bellow actuator; Fabrication; Data-Driven modeling; Learning controlOrganisational unit
03758 - D'Andrea, Raffaello / D'Andrea, Raffaello
More
Show all metadata
ETH Bibliography
yes
Altmetrics