馃殌 Feature
The joint spline, is so far only defined between its bounds $[q_{min}, q_{max}]$. In other software, such as scipy or opensim, linear extrapolation is done to avoid undefined behaviour outside of the bounds.
Here is a proposition for linear extrapolation for the joint spline :
- Check if $q_{min} < q < q_{max}$
- If not, evaluate the spline at the nearest endpoint, either $q_{min}$ or $q_{max}$, e.g compute S and M
- Extrapolate the transform using the endpoint twist
$$
G(q) = M_{min \ or \ max} exp((q - q_{min \ or \ max})S_{min \ or \ max})
$$
To speed up the process, when the joint is build, we can precompute $M_{min}$, $M_{max}$, $S_{min}$ and $S_{max}$
@LucasJoseph @jorisv @lpuch
Motivation
To match opensim SimmSpline behaviour.
Alternatives
No response
Additional context
It will have to wait for #2784 to be merged, before it is investigated.
Checklist
馃殌 Feature
The joint spline, is so far only defined between its bounds$[q_{min}, q_{max}]$ . In other software, such as scipy or opensim, linear extrapolation is done to avoid undefined behaviour outside of the bounds.
Here is a proposition for linear extrapolation for the joint spline :
To speed up the process, when the joint is build, we can precompute$M_{min}$ , $M_{max}$ , $S_{min}$ and $S_{max}$
@LucasJoseph @jorisv @lpuch
Motivation
To match opensim
SimmSplinebehaviour.Alternatives
No response
Additional context
It will have to wait for #2784 to be merged, before it is investigated.
Checklist