Coherent beam combination can be used to overcome limitations associated with the power handling capability of a single fibre laser. However, due to interference effects, the spatial intensity profile of the combined beam is directly affected by the phase of each fibre. Therefore, monitoring and control of the fibre phases is required for practical application. Here, we show that a neural network can extract this phase information from a far-field intensity profile, in a single step, hence unlocking the potential for real-time beam shaping. Further investigation shows that the neural network encoded fundamental rules associated with interference theory.

© (2023) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.

Ben Mills, James A. Grant-Jacob, and Michalis N. Zervas, "Optimisation of coherent beam combination using deep learning," Proc. SPIE 12400, Fiber Lasers XX: Technology and Systems, 124000G (Presented at SPIE LASE: January 31, 2023; Published: 8 March 2023); https://doi.org/10.1117/12.2659491.