ビーム発散制御と連続行動空間による光干渉計のアライメント
Aligning an optical interferometer with beam divergence control and continuous action space
マッハ・ツェンダー干渉計の視覚ベースのアライメントを強化学習で行い、シミュレーションのみで訓練したエージェントが実機で既存手法と人間の専門家を上回った。
著者: Stepan Makarenko, Dmitry Sorokin, Alexander Ulanov, A. I. Lvovsky
分類: cs.RO, cs.AI, cs.LG, physics.optics
原文アブストラクト
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.
