Bayesian inference applied to interferometric calibration of extreme-precision radial-velocity spectrographs
Lead supervisor: David Buscher, Department of Physics
Research proposal
The detection of "Earth twins'' – rocky planets orbiting at radii of order 1au around solar-type stars – will be one of the major stepping-stones in our search for life in the Universe. One of the most promising avenues to make these detections is to use extreme-precision radial-velocity (EPRV) spectrographs to detect the minute (10 cm/s) stellar reflex Doppler signals caused by the Earth twin.
A critical element to obtaining this 10cm/s precision is calibrating the wavelengths of the spectral features used for radial-velocity measurement to better than 1/10,000th of a spectrograph pixel. This project aims to test a new strategy for this extreme precision calibration, involving projection of interferometric fringes on the spectrograph pixels and Bayesian analysis of these fringe patterns to derive an extremely accurate map of pixel-by-pixel wavelength variations in the spectrograph.
The project will involved developing software in Python to implement this Bayesian data-analysis strategy and testing it on simulated fringe data which includes simulated disturbances such as vibrations and temperature fluctuations. The end product will be a prediction of the achievable performance of such a system under realistic conditions.
The lead applicant is an expert in the analysis of interferometric data and is involved in calibration of the ANDES high-resolution spectrograph.