Deep learning based phase-locking despite missing laser beams

Read full text
Authors: Nathanaël Hulard, Bastien Rouzé, Laurent Lombard, S. Barland, Pierre Bourdon
Cited by: 0

Abstract

Far-field interference pattern obtained in the non-focal plane on the experimental setup of CBC of 7 fiber lasers using AI for phase-control. The camera auto-scaling is on. Each beam is powered off one after another until only one is left. The interference patterns remains stable after each powered off beam illustrating the neural network phase-locking capability despite missing beams. 0:02 CBC 7 beams, 0:12 CBC 6 beams,0:14 CBC 5 beams,0:16 CBC 4 beams,0:19 CBC 3 beams,0:21 CBC 2 beams, 0:23 1 beam left.

Publication details

Journal / source
Figshare
Publication date
1 January 2026