rlabbe/filterpy

indexing issue in rts_smoother

mjsobrep opened this issue · 0 comments

It looks like there is an indexing issue on the Fs:

Pp[k] = dot(dot(Fs[k+1], P[k]), Fs[k+1].T) + Qs[k+1]

I'm not sure why there are the k+1 indeces. I would expect this to be the same as the calculation of P in the forward pass:
self.P = self._alpha_sq * dot(dot(F, self.P), F.T) + Q

Comparing to the standalone smoother:

pP[k] = dot(dot(Fs[k], P[k]), Fs[k].T) + Qs[k]

the k+1 terms are not there.