Oh, thank you! I changed it to an ECI frame?
However, I am still encountering the same issue with nan and 0.0 as values for the collision probability.
Oh i did not mean it as if it was a mistake but rather to confirm that you know in which frame these covariances are expressed.
I still need to investigate your issue. i’ll try to do it asap 
Cheers,
Vincent
Hello @Pluto ,
After taking a look at your code, i noticed that the first diagonal term of your covariance is negative which is not possible and is the reason why you are getting NaN afterwards.
Covariances matrices are supposed to be positive semi-definite as explained here:
How do you get this initial state covariance ?
Cheers,
Vincent