I recently defended my PhD thesis with the title "Adaptive Estimation using Gaussian Mixtures". Stay tuned for its publication.
Today, we released a big update of our robust sensor fusion library libRSF.
In conjunction to our recent RA-L paper we released a novel Gaussian Mixture representation for factor graphs and much more.
libRSF on GitHub
I had the pleasure to visit the
Intelligent Positioning and Navigation Laboratory
at the Honk Kong PolyU to spend two weeks on robust sensor fusion in the urban canyons of Hong Kong.
I also gave a seminar talk on "Incrementally learned Mixture Models for GNSS Localization in Urban Areas".
Thanks to Dr. Hsu and Weisong Wen for the invitation and the great time in HK.
In October 2023 I moved to the Siemens AG to develop robust Indoor localization systems. This page will vanish soon, so have a look at my website.
I currently focus on sensor fusion algorithms which are robust against measurements that are subject to a non-Gaussian error distribution. With tools like non-linear optimization, maximum-likelihood methods or Bayesian inference I want to create probabilistic plausible algorithms that are able to adapt themselves to unforeseen conditions and improve the robustness of today's autonomous systems.
More information can be found here:
- Adaptive Gaussian Mixtures and the EM Algorithm
- Robust Sensor fusion with Self-tuning Mixtures
- Robust Satellite-based Vehicle Localization in Urban Environments
Only the possibility to evaluate algorithms on the data of others makes research comparable. Therefore, I want to provide the datasets of my publications below.