Validating ZMP estimator
Experiments were run on the robot in order to validate the ZMP estimator. The results were definitely encouraging, but they also highlighted some work that remains to be done.
In particular, a calibration routine for the force sensors seems necessary (see #53).
Further, some criticality was evidenced in case the bias on one of the sensors lead to negative force readings, in which case, the ZMP estimation might explode. I will quickly fix this.
Once these issues are addressed, we will re-run the validation experiments on the robot. Next, we will be able to test the CoM admittance control scheme with force sensor feedback.