Filtering a Two-Dimensional state in presence of One-Dimensional Observations

What is it about?

Showing that a very simple approximate filter can do the job of filtering well enough if we have a low noise chanel. No need for the EKF or other expensive filter

Featured Image

Photo by Adi Goldstein on Unsplash

Photo by Adi Goldstein on Unsplash

Why is it important?

It can be used in robotics or any situation where we need to estimate the state (position, velocity) in real time and the available observations are of one quantity only, for instance 1 sensor. We propose an approximate solution and mathematically prove that it is as good as the exact solution. It can naturally be generalized to multiple dimensions (dimension 2n for the state nad dimension n for the observations or sensors)

Perspectives

“

Explains that good approximations for the estimation/filtering problem exist and there is no need to use computer intensive solution that are unfeasible in real time.

”