Improved displacement estimation in homodyne Michelson interferometers

Abstract

A Michelson interferometer is a victim to lots of displacement noise that changes over time and distorts the measurement making them less accurate. There are some types of estimators for these distortions, but they usually claim a lot of computer power and it is therefore a need for investigating methods using less computationally dependent filters for these estimates. In this paper an estimation method for estimating the unknown parameters of a Homodyne Michelson interferometer system using the Extended Kalman filter is suggested. A good mathematical model of an interferometer based off Jones calculus is implemented into Simulink to design and verify the results of the filter and compared with data obtained with a Michelson interferometer learning kit from ThorLabs. The designed state space model doesn’t estimate the amplitude and the quadrature error but shows good results when taking the final estimated movement of the measured object. The mathematical model used in Simulink is well described as well as the state space model used for in the Extended Kalman Filter. This paper shows only one approach on how to use the EKF for this purpose and there are still improvements to be done on this approach. Parallel with this project a bachelor thesis have been written with a more practical approach on how to use the interferometer (Nesjø & Kvamme, 2021).