@article{1fd7cb9306fd48419c171922d7ddd3fa,
title = "Model-Based Reinforcement Learning for Closed-Loop Dynamic Control of Soft Robotic Manipulators",
abstract = "Dynamic control of soft robotic manipulators is an open problem yet to be well explored and analyzed. Most of the current applications of soft robotic manipulators utilize static or quasi-dynamic controllers based on kinematic models or linearity in the joint space. However, such approaches are not truly exploiting the rich dynamics of a soft-bodied system. In this paper, we present a model-based policy learning algorithm for closed-loop predictive control of a soft robotic manipulator. The forward dynamic model is represented using a recurrent neural network. The closed-loop policy is derived using trajectory optimization and supervised learning. The approach is verified first on a simulated piecewise constant strain model of a cable driven under-Actuated soft manipulator. Furthermore, we experimentally demonstrate on a soft pneumatically actuated manipulator how closed-loop control policies can be derived that can accommodate variable frequency control and unmodeled external loads.",
keywords = "Dynamic control, machine learning, manipulation, reinforcement learning, soft robotics",
author = "Thuruthel, {Thomas George} and Egidio Falotico and Federico Renda and Cecilia Laschi",
note = "Funding Information: European Unions Horizon 2020 Research and Innovation Program under Grant 785907 Funding Information: Manuscript received July 4, 2018; accepted October 7, 2018. Date of publication November 12, 2018; date of current version February 4, 2019. This paper was recommended for publication by Associate Editor J. Paik and Editor A. Billard upon evaluation of the reviewers{\textquoteright} comments. This work was support in part by the European Commission through the I-SUPPORT project (HORIZON 2020 PHC-19, #643666) and in part by the European Unions Horizon 2020 Research and Innovation Program under Grant 785907 (HBP SGA2). (Corresponding author: Thomas George Thuruthel.) T. G. Thuruthel, E. Falotico, and C. Laschi are with The BioRobotics Institute, Scuola Superiore Sant{\textquoteright}Anna, 56025 Pisa, Italy (e-mail:, t.thuruthel@ santannapisa.it;
[email protected];
[email protected]). Publisher Copyright: {\textcopyright} 2004-2012 IEEE.",
year = "2019",
month = feb,
doi = "10.1109/TRO.2018.2878318",
language = "British English",
volume = "35",
pages = "127--134",
journal = "IEEE Transactions on Robotics",
issn = "1552-3098",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "1",
}