Smoothed propagator using Unscented Kalman Filter

Orekit usage
5

Hi @markrutten :smiley: . Thanks for your detailed response. I am so sorry for having left this conversation for so long.

Regarding your comment, I thought smoothing could be used to do the following:

  1. First, a pass with the Kalman Filter.
  2. After reaching the last measurement, perform a backward smoothing to correct the information of the first predictions using the future results and the information of the last ones.
  3. Third, a second forward Kalman filtering to get the “best” possible orbit determination

Am I wrong? Is the last estimation provided by the first forward Kalman filtering the best possible outcome and nothing could be done?

Thank you so much in advance :laughing: