TY - JOUR
T1 - Data-Driven Disturbance Observers for Estimating External Forces on Soft Robots
AU - Santina, Cosimo Della
AU - Truby, Ryan Landon
AU - Rus, Daniela
N1 - Funding Information:
Manuscript received February 24, 2020; accepted June 29, 2020. Date of publication July 21, 2020; date of current version July 28, 2020. This letter was recommended for publication by Associate Editor D. Shin and Editor C. Laschi upon evaluation of the reviewers’ comments. This work was supported by the NSF EFRI Program under Grant 1830901. The work of Ryan Landon Truby was supported by the Schmidt Science Fellows program, in partnership with the Rhodes Trust. (Cosimo Della Santina and Ryan Landon Truby contributed equally to this work.) (Corresponding author: Cosimo Della Santina.) The authors are with the MIT Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139 USA (e-mail: cosimodellasantina@gmail.com; rltruby@mit.edu; rus@csail.mit.edu).
Publisher Copyright:
© 2016 IEEE.
PY - 2020/10
Y1 - 2020/10
N2 - Unlike traditional robots, soft robots can intrinsically interact with their environment in a continuous, robust, and safe manner. These abilities - and the new opportunities they open - motivate the development of algorithms that provide reliable information on the nature of environmental interactions and, thereby, enable soft robots to reason on and properly react to external contact events. However, directly extracting such information with integrated sensors remains an arduous task that is further complicated by also needing to sense the soft robot's configuration. As an alternative to direct sensing, this paper addresses the challenge of estimating contact forces directly from the robot's posture. We propose a new technique that merges a nominal disturbance observer, a model-based component, with corrections learned from data. The result is an algorithm that is accurate yet sample efficient, and one that can reliably estimate external contact events with the environment. We prove the convergence of our proposed method analytically, and we demonstrate its performance with simulations and physical experiments.
AB - Unlike traditional robots, soft robots can intrinsically interact with their environment in a continuous, robust, and safe manner. These abilities - and the new opportunities they open - motivate the development of algorithms that provide reliable information on the nature of environmental interactions and, thereby, enable soft robots to reason on and properly react to external contact events. However, directly extracting such information with integrated sensors remains an arduous task that is further complicated by also needing to sense the soft robot's configuration. As an alternative to direct sensing, this paper addresses the challenge of estimating contact forces directly from the robot's posture. We propose a new technique that merges a nominal disturbance observer, a model-based component, with corrections learned from data. The result is an algorithm that is accurate yet sample efficient, and one that can reliably estimate external contact events with the environment. We prove the convergence of our proposed method analytically, and we demonstrate its performance with simulations and physical experiments.
KW - Modeling
KW - and learning for soft robots
KW - contact modeling
KW - control
KW - model learning for control
UR - http://www.scopus.com/inward/record.url?scp=85089903184&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85089903184&partnerID=8YFLogxK
U2 - 10.1109/LRA.2020.3010738
DO - 10.1109/LRA.2020.3010738
M3 - Article
AN - SCOPUS:85089903184
SN - 2377-3766
VL - 5
SP - 5717
EP - 5724
JO - IEEE Robotics and Automation Letters
JF - IEEE Robotics and Automation Letters
IS - 4
M1 - 9145617
ER -