Aligning an optical interferometer with beam divergence control and continuous action space

2021/07/09 by Stepan Makarenko, Makarenko, Stepan, Dmitry Sorokin +5
Computer Science · #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #FOS: Physical sciences #Machine Learning (cs.LG) #Metaheuristic Optimization Algorithms Research #Neural Networks and Reservoir Computing #Optics (physics.optics) #Reinforcement Learning in Robotics #Robotics (cs.RO)

paper · pdf · doi:10.48550/arxiv.2107.04457

openalex publication_date 2021/07/09 · openalex created_date 2021/07/19 · openalex updated_date 2026/07/28

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.

Citations

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Playing Atari with Deep Reinforcement Learning
• Continuous control with deep reinforcement learning
• Challenges of Real-World Reinforcement Learning
• Learning dexterous in-hand manipulation
• Dota 2 with Large Scale Deep Reinforcement Learning

Related

• Active Alignments of Lens Systems with Reinforcement Learning
• Precision positioning in free-space optical communication systems via PID control tuned by RL
• An Optical Control Environment for Benchmarking Reinforcement Learning Algorithms
• Action-Attentive Deep Reinforcement Learning for Autonomous Alignment of Beamlines
• Automating Experimental Optics with Sample Efficient Machine Learning Methods