You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I was comparing the decomposition of sigma points of MerweUnscentedTransform.java with my own implementation using Matlab, and the sigma points 7 and 13 were exactly the same for the Hipparchus implementation. I have tracked it down to the cholesky descomposition in the unscentedTransform method of AbstractUnscentedTransform.java, where the last column is all zeros.
I leave here the example of matrix that I am decomposing:
I can work with my implementation for the UT, but I open this issue just in case there is a bug. It could also be a numerical limitation for this particular matrix.
The text was updated successfully, but these errors were encountered:
I was comparing the decomposition of sigma points of
MerweUnscentedTransform.java
with my own implementation using Matlab, and the sigma points 7 and 13 were exactly the same for the Hipparchus implementation. I have tracked it down to the cholesky descomposition in theunscentedTransform
method ofAbstractUnscentedTransform.java
, where the last column is all zeros.I leave here the example of matrix that I am decomposing:
And the code in matlab to test Hipparchus decomposition against Matlab's:
These are my outputs:
I can work with my implementation for the UT, but I open this issue just in case there is a bug. It could also be a numerical limitation for this particular matrix.
The text was updated successfully, but these errors were encountered: