Check of mbs_data at modules start

See merge request robotran/mbsysc!299

[Invdyn] fix qact calculation + warn for the case 4

See merge request robotran/mbsysc!298

Fix norm h

Closes #180

See merge request robotran/mbsysc!297

use nhu rather than Ncons to decide wether to save lambda

See merge request robotran/mbsysc!296

Saving Lagrange multipliers

See merge request robotran/mbsysc!295

# Conflicts:
#   CHANGELOG.md

Small correction to correct memory leak due to sensors

See merge request robotran/mbsysc!293

: Improve random initial perturbation in coord_part

See merge request robotran/mbsysc!294

[BUFFER][LAMBDA] Lagrange multipliers are saved during Invdyn simulation (was partialy implemented).

Addon test

See merge request robotran/mbsysc!292

An initial random perturbation is done at begining.
- This perturbation must be done around the initial position rather
than 0
- The amplitude is often too large==> we should add a parameter for
that  (usefull for W/R contact and cam/follower contact)