TODO retest all the trajectories in simulation and in reality

fix bug in the com vertical trajectories

There where auto-collision and singularities

- Implemented BSplines for foot trajectory along Y axis

implement correctly the foot trajectory along the vertical axis using BSpline

- ComputeAllWithPolynom
- ComputeAllWithBSplines

Introduces BSplines for X and Y axes

minor modification off output debug in the rest of the files

Still some discontinuities in theta.

Modify the way the data of the ZMPVeolicityReferencedQP class are treated
continue the implementation of the dynamical filter of Kajita2003icra

Continue the work on Herdt online

Begin the implementation of a multicontact PG based on Hirukawa2007

over a trajectory of com and feet