Estimation smoothing splines TLC
Opened this issue · 1 comments
sigmanotation commented
It seems like the smoothing splines can heavily distort the geometry of certain trajectories, resulting in poor kinematic variable guesses that the least-squares minimizer cannot recover from but the LBFGS minimizer can. Needs some investigation into this.
gwm17 commented
Maybe the best approach here is to revert to using the neighborhood smoothing... at least then the smoothing factor had some physical meaning.