Real experiment for Humanoids 2015 paper

naveenoid commented 9 years ago

@jeljaik @prashanthr05 I will try to list out some objectives for the real measurement that we need to perform on 30/06/2015.

[x] Fit IMU in a reasonably similar way to foot, as firm as possible, using double side tape or whatever means
[x] Fit skin under foot
[x] Place robot vertically, perform macumba, try to obtain computed offset values.
[x] Perform toppling on hard ground with skin underneath, with foot IMU plugged in inertial port.
[x] datadump all leg and torso joint angles, F/T sensors in both legs and skin.
[x] Repeat trials with soft mat placed below robot foot (try to first find new toplling angle and velocity ranges
[x] Fold mat once, repeat trial. (:point_up: Remember: Each time redo Macumba on flat surface first)
[x] Fold mat twice, repeat trial
[x] Fold mat third time, repeat trial

jeljaik commented 9 years ago

1. Setting up Datadumpers

Use the script https://github.com/jeljaik/extended-kalman-filter/blob/dynamics-tests/EKF_EulerAngle_IROS2015/robotData/dataDumperAppGenerator.py in order to generate the dataDumper application. Once you launch it, when you run the application all datadumpers (one per port) wait till you connect to actually write to file, and they will stop writing once you stop the application.

Launch as:

python dataDumperAppGenerator.py --ports /icub/left_leg/state:o /icub/right_leg/state:o /icub/torso/state:o  /icub/left_foot/analog:o /icub/right_foot/analog:o /icub/left_leg/analog:o /icub/right_leg/analog:o /icub/skin/left_foot /icub/skin/right_foot /icub/inertial --host localhost --name backwardTippingTest3006201501 > application.xml

Note for Naveen: In the folder I sent to you there's this file 'application.xml' which you can launch with yarpmanager as: yarpmanager --application application.xml, but it was missing the legs' FT sensors, that I just added to the previous lines, so you better run it again.

2. Skin Calibration!

The skin needs to be calibrated before the robot is put on the ground. To do so, from console launch the application skinManager + Feet V2.0 or something alike, can't remember the exact name, run it and don't forget to connect all ports. Then hit the calibrate button before the robot is on the ground. Once you do this, try to be as quick as possible with running the experiment and do it again for every other trial.

3. Sequence

Get the robot in the starting position by leaving the arms in the original home position, and torso and legs as specified next.

Backward Tipping Scenario

torso from 0 0 20 to 0 0 -18.1 in 1.0 sec right_leg fixed at -10 0 0 0 -2 -2 left_leg fixed at -10 0 0 0 0 0

Note for Naveen Also in the folder I sent you inside every backwardTippingCompliant300620150XX theres a sequence directory with these transitions loaded. You need only to load it with yarpmotorgui-gtk (preferably) by going to the tab all, followed by load sequence and to play when you're ready: run sequence (time).

4. Macumba

The old "macumba" is now under the name twoFeetStandingIdleAndCalib.sh and it's installed so you don't need to look for it anywhere in particular.

5. Offsets

Before playing back the tipping sequence you wanna record the FT offsets given by wholeBodyDynamicsTree by simply launching it (the head IMU does not need to be attached for it to run) and simply copying the last lines. Something like:

Printed output of wholeBodyDynamicsTree Experiment backwardTipping3006201501

[INFO]wholeBodyDynamicsThread: complete calibration at system time :  30-06-2015 03:16:32 
[INFO]wholeBodyDynamicsThread: new calibration for FT  l_arm_ft_sensor  is   70.361964   5.230722    245.676737 -3.080854   -5.566984 0.917822 
[INFO]wholeBodyDynamicsThread: new calibration for FT  r_arm_ft_sensor  is   129.303830  2.334646    122.035770 -0.914977   -3.123182 0.284610 
[INFO]wholeBodyDynamicsThread: new calibration for FT  l_leg_ft_sensor  is   113.821837  104.494539 -214.493500 -10.198060   1.850435-0.482308 
[INFO]wholeBodyDynamicsThread: new calibration for FT  l_foot_ft_sensor  is  -23.237569 -38.295595   332.422470  0.706069   -6.038355-0.497729 
[INFO]wholeBodyDynamicsThread: new calibration for FT  r_leg_ft_sensor  is  -51.867380   50.283953  -60.974108  -10.459150   4.045945 0.237656 
[INFO]wholeBodyDynamicsThread: new calibration for FT  r_foot_ft_sensor  is   19.771900  22.325223  -117.840856  0.996938    4.261731-0.213692

jeljaik commented 9 years ago

NOTE TO SELF I have a local copy of the datasets in extended-kalman-filter/EKF_EulerAngle_IROS2015/robotData. A copy was sent to Naveen's personal email.

jeljaik / extended-kalman-filter