<?xml version="1.0" encoding="utf-8"?>
<export-example>
  <doc>
    <id>2580</id>
    <completedYear>2022</completedYear>
    <publishedYear/>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>10193</pageFirst>
    <pageLast>10200</pageLast>
    <pageNumber>8</pageNumber>
    <edition/>
    <issue/>
    <volume/>
    <type>conferenceobject</type>
    <publisherName/>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>--</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">Analysis of Randomization Effects on Sim2Real Transfer in Reinforcement Learning for Robotic Manipulation Tasks</title>
    <abstract language="eng">Randomization is currently a widely used approach in Sim2Real transfer for data-driven learning algorithms in robotics. Still, most Sim2Real studies report results for a specific randomization technique and often on a highly customized robotic system, making it difficult to evaluate different randomization approaches systematically. To address this problem, we define an easy-to-reproduce experimental setup for a robotic reach-and-balance manipulator task, which can serve as a benchmark for comparison. We compare four randomization strategies with three randomized parameters both in simulation and on a real robot. Our results show that more randomization helps in Sim2Real transfer, yet it can also harm the ability of the algorithm to find a good policy in simulation. Fully randomized simulations and fine-tuning show differentiated results and translate better to the real robot than the other approaches tested</abstract>
    <parentTitle language="eng">2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Kyoto, Japan</parentTitle>
    <identifier type="doi">10.1109/IROS47612.2022.9981951</identifier>
    <enrichment key="PeerReviewed">Ja</enrichment>
    <enrichment key="RS_Correlation">Ja</enrichment>
    <enrichment key="opus.source">publish</enrichment>
    <enrichment key="opus.doi.autoCreate">false</enrichment>
    <enrichment key="opus.urn.autoCreate">true</enrichment>
    <author>Josip Josifovski</author>
    <author>Mohammadhossein Malmir</author>
    <author>Noah Klarmann</author>
    <author>Bare Luka Žagar</author>
    <author>Nicolás Navarro-Guerrero</author>
    <author>Alois Knoll</author>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Sim2Real</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Robotics</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Reinforcment Learning</value>
    </subject>
    <collection role="institutes" number="">Fakultät für Wirtschaftsingenieurwesen</collection>
    <thesisPublisher>Technische Hochschule Rosenheim</thesisPublisher>
  </doc>
</export-example>
