%0 Conference Paper %1 9981951 %A Josifovski, Josip %A Malmir, Mohammadhossein %A Klarmann, Noah %A Žagar, Bare Luka %A Navarro-Guerrero, Nicolás %A Knoll, Alois %B 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) %D 2022 %K myown %P 10193-10200 %R 10.1109/IROS47612.2022.9981951 %T Analysis of Randomization Effects on Sim2Real Transfer in Reinforcement Learning for Robotic Manipulation Tasks %U https://ieeexplore.ieee.org/document/9981951/ %X 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.

@inproceedings{9981951, abstract = {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.}, added-at = {2023-02-09T12:34:59.000+0100}, author = {Josifovski, Josip and Malmir, Mohammadhossein and Klarmann, Noah and Žagar, Bare Luka and Navarro-Guerrero, Nicolás and Knoll, Alois}, biburl = {https://www.bibsonomy.org/bibtex/2a1668b3ec66dbaa25cd09314addd648c/nng}, booktitle = {2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)}, doi = {10.1109/IROS47612.2022.9981951}, interhash = {95e8a8bc1821a3100780be0f6b88bf7d}, intrahash = {a1668b3ec66dbaa25cd09314addd648c}, issn = {2153-0866}, keywords = {myown}, month = oct, pages = {10193-10200}, timestamp = {2023-03-15T14:15:03.000+0100}, title = {Analysis of Randomization Effects on Sim2Real Transfer in Reinforcement Learning for Robotic Manipulation Tasks}, url = {https://ieeexplore.ieee.org/document/9981951/}, year = 2022 }