TU Darmstadt / ULB / TUbiblio

Robot Learning From Randomized Simulations: A Review

Muratore, Fabio ; Ramos, Fabio ; Turk, Greg ; Yu, Wenhao ; Gienger, Michael ; Peters, Jan (2022)
Robot Learning From Randomized Simulations: A Review.
In: Frontiers in Robotics and AI, 2022, 9
doi: 10.26083/tuprints-00021227
Article, Secondary publication, Publisher's Version

Abstract

The rise of deep learning has caused a paradigm shift in robotics research, favoring methods that require large amounts of data. Unfortunately, it is prohibitively expensive to generate such data sets on a physical platform. Therefore, state-of-the-art approaches learn in simulation where data generation is fast as well as inexpensive and subsequently transfer the knowledge to the real robot (sim-to-real). Despite becoming increasingly realistic, all simulators are by construction based on models, hence inevitably imperfect. This raises the question of how simulators can be modified to facilitate learning robot control policies and overcome the mismatch between simulation and reality, often called the “reality gap.” We provide a comprehensive review of sim-to-real research for robotics, focusing on a technique named “domain randomization” which is a method for learning from randomized simulations.

Item Type: Article
Erschienen: 2022
Creators: Muratore, Fabio ; Ramos, Fabio ; Turk, Greg ; Yu, Wenhao ; Gienger, Michael ; Peters, Jan
Type of entry: Secondary publication
Title: Robot Learning From Randomized Simulations: A Review
Language: English
Date: 2022
Place of Publication: Darmstadt
Year of primary publication: 2022
Publisher: Frontiers
Journal or Publication Title: Frontiers in Robotics and AI
Volume of the journal: 9
Collation: 19 Seiten
DOI: 10.26083/tuprints-00021227
URL / URN: https://tuprints.ulb.tu-darmstadt.de/21227
Corresponding Links:
Origin: Secondary publication via sponsored Golden Open Access
Abstract:

The rise of deep learning has caused a paradigm shift in robotics research, favoring methods that require large amounts of data. Unfortunately, it is prohibitively expensive to generate such data sets on a physical platform. Therefore, state-of-the-art approaches learn in simulation where data generation is fast as well as inexpensive and subsequently transfer the knowledge to the real robot (sim-to-real). Despite becoming increasingly realistic, all simulators are by construction based on models, hence inevitably imperfect. This raises the question of how simulators can be modified to facilitate learning robot control policies and overcome the mismatch between simulation and reality, often called the “reality gap.” We provide a comprehensive review of sim-to-real research for robotics, focusing on a technique named “domain randomization” which is a method for learning from randomized simulations.

Status: Publisher's Version
URN: urn:nbn:de:tuda-tuprints-212275
Additional Information:

Keywords: robotics, simulation, reality gap, simulation optimization bias, reinforcement learning, domain randomization, sim-to-real

Classification DDC: 000 Generalities, computers, information > 004 Computer science
Divisions: 20 Department of Computer Science
20 Department of Computer Science > Intelligent Autonomous Systems
Date Deposited: 04 May 2022 13:47
Last Modified: 27 Oct 2023 10:16
PPN:
Corresponding Links:
Export:
Suche nach Titel in: TUfind oder in Google
Send an inquiry Send an inquiry

Options (only for editors)
Show editorial Details Show editorial Details