[20p-E214-1~14] 4.3 Ultrafast Optics and Laser Processing

▲ [20p-E214-2] OAM Beams via Deep Reinforcement Learning based Coherent Beam Combining

〇Henrik Tuennermann¹, Akira Shirakawa¹ (1.University of Electro-Communications)

Keywords:Coherent Beam Combining, Artificial Intelligence, Reinforcement Learning

Coherent beam combining (CBC) is an important method to realize laser sources which are not possible to realize with a single emitter laser system. To do this, multiple emitters are phase-stabilized and coherently combined. There are two main methods of CBC: filled aperture combining, which combines the near and the far-field using an optical element like a mirror or a polarizer, and tiled aperture combining. Tiled aperture combining only combines in the far-field, which limits beam quality but allows beam shaping. We theoretically investigated the use of deep reinforcement learning for generating OAM beams via tiled aperture combining.