Please use this identifier to cite or link to this item: http://dx.doi.org/10.14279/depositonce-16066
For citation please use:
Main Title: Bridging Reinforcement Learning and Iterative Learning Control: Autonomous Motion Learning for Unknown, Nonlinear Dynamics
Author(s): Meindl, Michael
Lehmann, Dustin
Seel, Thomas
Type: Article
URI: https://depositonce.tu-berlin.de/handle/11303/17286
http://dx.doi.org/10.14279/depositonce-16066
License: https://creativecommons.org/licenses/by/4.0/
Abstract: This work addresses the problem of reference tracking in autonomously learning robots with unknown, nonlinear dynamics. Existing solutions require model information or extensive parameter tuning, and have rarely been validated in real-world experiments. We propose a learning control scheme that learns to approximate the unknown dynamics by a Gaussian Process (GP), which is used to optimize and apply a feedforward control input on each trial. Unlike existing approaches, the proposed method neither requires knowledge of the system states and their dynamics nor knowledge of an effective feedback control structure. All algorithm parameters are chosen automatically, i.e. the learning method works plug and play. The proposed method is validated in extensive simulations and real-world experiments. In contrast to most existing work, we study learning dynamics for more than one motion task as well as the robustness of performance across a large range of learning parameters. The method’s plug and play applicability is demonstrated by experiments with a balancing robot, in which the proposed method rapidly learns to track the desired output. Due to its model-agnostic and plug and play properties, the proposed method is expected to have high potential for application to a large class of reference tracking problems in systems with unknown, nonlinear dynamics.
Subject(s): autonomous systems
Gaussian processes
iterative learning control
nonlinear systems
reinforcement learning
robot learning
Issue Date: 12-Jul-2022
Date Available: 1-Aug-2022
Language Code: en
DDC Class: 004 Datenverarbeitung; Informatik
Sponsor/Funder: DFG, 390523135, EXC 2002: Science of Intelligence (SCIoI)
Journal Title: Frontiers in Robotics and AI
Publisher: Frontiers
Volume: 9
Article Number: 793512
Publisher DOI: 10.3389/frobt.2022.793512
EISSN: 2296-9144
TU Affiliation(s): Fak. 4 Elektrotechnik und Informatik » Inst. Energie- und Automatisierungstechnik » FG Regelungssysteme
Appears in Collections:Technische Universität Berlin » Publications

Files in This Item:
DataSheet1.pdf
Format: Adobe PDF | Size: 142 kB
DownloadShow Preview
Thumbnail
frobt-09-793512.pdf
Format: Adobe PDF | Size: 2.41 MB
DownloadShow Preview
Thumbnail
frobt-09-793512-g001.tif
Format: TIFF | Size: 276.26 kB
Download
Thumbnail
frobt-09-793512-g002.tif
Format: TIFF | Size: 265.32 kB
Download
Thumbnail
frobt-09-793512-g003.tif
Format: TIFF | Size: 167.33 kB
Download
Thumbnail
frobt-09-793512-g004.tif
Format: TIFF | Size: 143.16 kB
Download
Thumbnail
frobt-09-793512-g005.tif
Format: TIFF | Size: 450.89 kB
Download
Thumbnail
frobt-09-793512-g006.tif
Format: TIFF | Size: 175.6 kB
Download
Thumbnail
frobt-09-793512-g007.tif
Format: TIFF | Size: 115.89 kB
Download
Thumbnail
frobt-09-793512-g008.tif
Format: TIFF | Size: 325.66 kB
Download
Thumbnail
frobt-09-793512-g009.tif
Format: TIFF | Size: 312.93 kB
Download
Thumbnail
frobt-09-793512-g010.tif
Format: TIFF | Size: 131.88 kB
Download
Thumbnail
frobt-09-793512-g011.tif
Format: TIFF | Size: 289.33 kB
Download
Thumbnail

Item Export Bar

This item is licensed under a Creative Commons License Creative Commons