The current, least square (LS) based, method for displacement estimation was introduced 40 years ago. Today, computational power is much more readily available, which has opened up for the possibly employ novel and more advanced methods such as particle filtering. These methods may offer estimates with better statistical properties than the existing estimation method based on LS. These possible improvements are of great interest as displacement interferometry is routinely applied to detect gravitational waves and for high precision manufacturing in both research and industry. The presentation will being with short introductions to both interferometry and particle filters. We then present extensive benchmarks between the LS approach and other inference methods based on the particle filter. Initial studies indicate that some of these methods perform as good as LS, but has the potential to become even better with some additional research.
Petter Ersbo is currently a master student at the Engineering Physics program at Uppsala University, Sweden. Since February, he has been working on new methods of displacement estimation for homodyne Michelson interferometers at the University of Newcastle, Australia. He has earlier as part of his studies been involved in a research project on radio channel estimation for massive MIMO transmission. In August, he starts working with R&D at Ericsson in Stockholm, Sweden.