Machine learning generated solitons for distributed acoustic sensing

Optical solitons are self-sustaining wave packets that maintain their shape and speed as they propagate through optical media, demonstrating a balance between nonlinearity and dispersion effects. Traditionally, solitons have been studied and applied primarily in optical communication systems, where the focus has been on shorter pulses that preserve their shape during transmission. However, these conventional solitons are not well-suited for distributed acoustic sensing (DAS) applications, where preserving phase information is crucial for accurate sensing and detection. In this paper, we introduce a novel soliton design for DAS, featuring an extended pulse duration of 50 nanoseconds. We leverage machine learning techniques to design solitons that effectively maintain their phase integrity during propagation.