Impedance control in cartesian space

[PeterBowman]

As a follow-up to #194, let's add some spring-like behavior to the ftCompensation app now that it knows how to convert FT sensor data into cartesian force commands while performing gravity compensation.

[PeterBowman]

Done at 488028b. There is a lot of friction mainly in the elbow joint that should be somehow compensated in the low-level (roboticslab-uc3m/yarp-devices#261), but for now it seems to work well in orientations (Y and Z axes) with this configuration:

ftCompensation --from ftCompensation-stump.ini --mode wrench \
               --sensorName ch0 --sensorRemote /jr3 \
               --cartesianRemote /teo/leftArm/CartesianControl \
               --linGain 1 --rotGain 1 \
               --linStiffness 250 --rotStiffness 10 \
               --linDamping 0 --rotDamping 1

Note this could have been implemented as an add-on for the GCMP command in BasicCartesianControl. Such attempt was initiated in 6e78a53...b8ed4ca and explained in roboticslab-uc3m/yarp-devices#258, but discarded (for now) to focus on low-level joint impedance control.

[jgvictores]

Just for the historic, and completely low priority, would a video be possible (maybe of this vs gcmp)?