Fix foot force

[EtienneAr]

Currently, the feet force is set to 100 arbitrarily (for pybullet)

Ideally it should:

• Measure the actual force at the feet
• Output an integer that has the same unit has the real robot (e.g, the real robot feet sensors should be calibrated, and that calibration should be propagated here)

The feet force should also be added for simple wrapper

[EtienneAr]

After doing some quick measurements on the robot feet, it seems like there is a simple linear correlation between the foot force and the value given in /lowstate/foot_force. (We put the robot foot on a scale and did several measurements for weights going from 0 to 5kg with 1kg increments, and repeated for every feet).

Data

Weight, Force, FR, FL, RL, RR
0, 0, 15, 14, 15, 16
1, 9.81, 18, 18, 20, 21
2, 19.62, 24, 24, 25, 26
3, 29.43, 29, 30, 31, 31
4, 39.24, 33, 38, 38, 38
5, 49.05, 37, 43, 43, 45

And the regression gives
Slope 0.5621 ± 0.02257
Y-intercept 14.21 ± 0.6703

To be implemented