World models are becoming more and more widely available, typically in the form of maps. The same applies to sensors, the number of which is currently skyrocketing.
The process of merging the information from several different sensors is referred to as sensor fusion. At the core of the sensor fusion problem lies a state inference problem. However, the complexity of today's world models severely limits the applicability of the Kalman filter. On the other hand, it is by now quite well known that the particle filter offers a good solution to the inference problem in this case.
An application where sensor fusion problems on this form commonly arise is localization, where the task is to compute the position of a moving object. In this seminar I will in particular focus on how to make use of a Rao-Blackwellized particle filter (RBPF) in solving various localization problems. Perhaps most importantly, I will show results from real world experiments using objects moving under water, on land, indoors and in the air.