Aligning an Optical Interferometer with Beam Divergence Control and Continuous Action Space

Abstract

Reinforcement learning is finding its way to real-world problem application, transferring from simulated environments to physical setups. In this work, we implement vision-based alignment of an optical Mach-Zehnder interferometer with a confocal telescope in one arm, which controls the diameter and divergence of the corresponding beam. We use a continuous action space; exponential scaling enables us to handle actions within a range of over two orders of magnitude. Our agent trains only in a simulated environment with domain randomizations. In an experimental evaluation, the agent significantly outperforms an existing solution and a human expert.

Cite

Text

Makarenko et al. "Aligning an Optical Interferometer with Beam Divergence Control and Continuous Action Space." Conference on Robot Learning, 2021.

Markdown

[Makarenko et al. "Aligning an Optical Interferometer with Beam Divergence Control and Continuous Action Space." Conference on Robot Learning, 2021.](https://mlanthology.org/corl/2021/makarenko2021corl-aligning/)

BibTeX

@inproceedings{makarenko2021corl-aligning,
  title     = {{Aligning an Optical Interferometer with Beam Divergence Control and Continuous Action Space}},
  author    = {Makarenko, Stepan and Sorokin, Dmitry Igorevich and Ulanov, Alexander and Lvovsky, Alexander},
  booktitle = {Conference on Robot Learning},
  year      = {2021},
  pages     = {918-927},
  volume    = {164},
  url       = {https://mlanthology.org/corl/2021/makarenko2021corl-aligning/}
}