Dear NaveGo-Team, thanks for your very incredible work.
I noticed that in the function "kf_analysis.m" line 47, when you calculate the chi from the innovation v, the S matrix is used instead of the inverse of S.
multiple references state that the formula should use inverse of S. (since for scalar innovation as an example, this would be equivalent to say that the normalized innovation is v/sqrt(S), and the normalized innovation squared is s^2/S)
Apart from this, it would be great also to add the Normalized Estimation Error Squared (NEES) metric for the synthetic-data example by applying similar approach for the Normalized-Innovation-Squared mentioned above. please refer to the following link for more details
https://kalman-filter.com/normalized-estimation-error-squared/
Dear NaveGo-Team, thanks for your very incredible work. I noticed that in the function "kf_analysis.m" line 47, when you calculate the chi from the innovation v, the S matrix is used instead of the inverse of S. multiple references state that the formula should use inverse of S. (since for scalar innovation as an example, this would be equivalent to say that the normalized innovation is v/sqrt(S), and the normalized innovation squared is s^2/S)
kindly check the following link https://kalman-filter.com/normalized-innovation-squared/ and the open source paper https://web.stanford.edu/group/arl/sites/default/files/public/publications/Robust_TRN_Framework.pdf
Apart from this, it would be great also to add the Normalized Estimation Error Squared (NEES) metric for the synthetic-data example by applying similar approach for the Normalized-Innovation-Squared mentioned above. please refer to the following link for more details https://kalman-filter.com/normalized-estimation-error-squared/