7–11 Dec 2026
The University of Sydney
Australia/Sydney timezone
AIP Congress 2026

Optimal design of application-specific photonic lanterns using Bayesian optimisation

Not scheduled
20m
Belinda Hutchinson Building (The University of Sydney )

Belinda Hutchinson Building

The University of Sydney

Abercrombie St & Codrington St NSW 2008
Contributed Oral ANZOS | Photonics and Optics (ANZCOP)

Speaker

Justin Vella (University of Sydney)

Description

We simulate how modulating the properties of the centre core within a photonic lantern (PL) can greatly affect the mode selectivity of LP modes. Such knowledge can allow for the fabrication of a 2-in-1 monolothic device: a phase-preserving focal-plane wavefront sensor and an optimised science fibre. The novel variation, where a single mode is coupled into a chosen output fibre (and is suppressed elsewhere) that rejects all other modes is known as the Hybrid Mode Selective PL (HMS-PL). The preferential coupling of fibre modes that have considerable overlap with that of a corrected wavefront enables the optimisation of starlight injection into a single fibre while interpreting power in other fibres being due to atmospheric turbulence.
We used the finite beam propagation (BeamPROP) algorithm and finite element method (FemSIM) to effectively simulate the propagation of supported LP modes throughout the PL. By incorporating Gaussian Process Regression, we sample a range of values adopted by the central core within the PL and measured the performance against a perfect HMS-PL Loss Function. Using this range of samples and performance we estimate the underlying probability distribution function of this Loss Function using Kernel Density Estimators. This Bayesian inference allowed us to visualise any co-dependencies between parameters, and allows the prediction of the maximum likelihood estimate (MLE) to achieve hybrid-mode selectivity. We checked the desired hybrid mode selectivity of this MLE by calculating the ratio of LP01 power coupling into the central core against the surrounding cores. From this, we conclude with a set of values that will likely result in the first HMS-PL, along with the knowledge of how these parameters vary with respect to each other.

I am the presenting author Yes

Author

Justin Vella (University of Sydney)

Co-authors

Dr Barnaby Norris (University of Sydney) Prof. Sergio Leon-Saval (University of Sydney)

Presentation materials

There are no materials yet.