Update iCub3 with feet covers #210

mfussi66 · 2021-09-22T14:10:07Z

The purpose of this PR is to update the simulation model of iCub 3 with the new feet covers. The model is shown in the picture below. The checkout used to generate the model is cad-mechanics-public master , commit: icub-tech-iit/cad-mechanics-public@a83def6

cc @S-Dafarra @GiulioRomualdi

Nicogene

If these changes have been tested on gazebo, green light from my side 👍🏻

Nicogene · 2021-09-22T14:46:08Z

But check it out that it is failing to generate the models :https://github.com/robotology/icub-models-generator/pull/210/checks?check_run_id=3676209300

mfussi66 · 2021-09-22T14:55:52Z

@Nicogene They were not tested in Gazebo, but I can of course do it as needed.

pattacini · 2021-09-22T15:00:07Z

Here's one error:

2021-09-22T14:21:58.9629855Z   File "/usr/local/lib/python3.8/dist-packages/simmechanics_to_urdf-0.2-py3.8.egg/simmechanics_to_urdf/firstgen.py", line 916, in buildTree
2021-09-22T14:21:58.9630992Z KeyError: ('r_ankle_2', 'SCSYS_R_ANKLE_2')
2021-09-22T14:21:58.9985516Z make[2]: *** [simmechanics/CMakeFiles/generate-models-simmechanics.dir/build.make:180: simmechanics/iCubGazeboV3.urdf] Error 1

cc @Mick3Lozzo

Mick3Lozzo · 2021-09-22T15:03:34Z

Well, I have to verify this with @mfussi66, is seems regarding the ankle, but I don't know if it is a problem of the CAD simulation model.

pattacini · 2021-09-22T16:09:50Z

Quoting @Mick3Lozzo

This coordinate system is inside a suppressed "test" group, so it is not included in the simulation model generation. Now we have SCSYS_R_ANKLE_2_FT_FRONT and SCSYS_R_ANKLE_2_FT_REAR.

I think we could remove it. Is it still used somehow in your application @GiulioRomualdi @S-Dafarra ?

S-Dafarra · 2021-09-22T18:15:06Z

Quoting @Mick3Lozzo

This coordinate system is inside a suppressed "test" group, so it is not included in the simulation model generation. Now we have SCSYS_R_ANKLE_2_FT_FRONT and SCSYS_R_ANKLE_2_FT_REAR.

I think we could remove it. Is it still used somehow in your application @GiulioRomualdi @S-Dafarra ?

I did not understand what you are referring to

pattacini · 2021-09-22T18:18:20Z

I'm referring to the frame SCSYS_R_ANKLE_2, which is causing the problem although seems to belong to the "suppressed test group" as we now have two frames, one for the front and one for the rear soles.

S-Dafarra · 2021-09-22T18:22:52Z

I'm referring to the frame SCSYS_R_ANKLE_2, which is causing the problem although seems to belong to the "suppressed test group" as we now have two frames, one for the front and one for the rear soles.

I don't think I ever seen it

mfussi66 · 2021-09-23T09:22:31Z

Since it is a KeyError, it means that the script cannot find the key r_ankle_2, which of course is a consequence of the reference frame not being present in the model (aka the XML generated file). So I think that it means that I have to remove the instances of r_ankle_2 in the file ICUB_3_all_options.yaml, correct?

Nicogene · 2021-09-23T09:41:16Z

It is quite strange that r_ankle_2 is missing, it has been substituted by a link with a different name?

(and why just r_ankle_2 and not l_ankle_2?)

That link is involved in these joints: r_ankle_roll, r_foot_front_ft_sensor and r_foot_rear_ft_sensor.

If I understood correctly this PR should just update some stls, but seems that something went wrong in the export of the simmechanics xml

Need more investigation

mfussi66 · 2021-09-23T09:51:10Z

It is quite strange that r_ankle_2 is missing, it has been substituted by a link with a different name?

It seems like it belongs to a now suppressed test group of frames: https://github.com/icub-tech-iit/fix/issues/966#issuecomment-925045131

(and why just r_ankle_2 and not l_ankle_2?)

Could it be that the script fails and exits on the first error, which happens to be caused by r_ankle_2 that is parsed first?

Nicogene · 2021-09-23T09:54:07Z

Could it be that the script fails and exits on the first error, which happens to be caused by r_ankle_2 that is parsed first?

Yes probably after removing that frame we will have other issues, I suggest to test it locally is quite easy, just follow:

icub-models-generator/.github/workflows/ci.yml

Lines 37 to 99 in 36ca1b9

    
               - name: Generate models 
        
                 run: | 
        
                   sudo apt-get update 
        
                   sudo apt-get install libeigen3-dev libace-dev libtinyxml-dev libxml2-dev 
        
                   # Save the url of the repository and the user-name of the committ author 
        
                   export CURRENT_REPOSITORY_URL=`git remote get-url origin` 
        
                   # Start in the parent directory of icub-model-generator 
        
                   cd ${GITHUB_WORKSPACE} 
        
                   sudo apt-get install  --assume-yes --force-yes python-lxml python-yaml python-numpy python-setuptools 
        
                   # probably python on the path return a python interpreter and the find_package(PythonInterp) in CMake another, 
        
                   # let's install both debian packages and pip packages to be sure  
        
                   sudo pip install lxml numpy pyyaml catkin_pkg 
        
                   # install urdf_parser_py and save the last commit SHA1 hash 
        
                   git clone $URDF_PARSER_PY_REPOSITORY_URL 
        
                   cd urdf_parser_py 
        
                   # workaround for https://github.com/robotology/simmechanics-to-urdf/issues/36 
        
                   git checkout 31474b9baaf7c3845b40e5a9aa87d5900a2282c3 
        
                   export URDF_PARSER_PY_COMMIT=`git rev-parse HEAD` 
        
                   sudo python setup.py install 
        
                   cd ${GITHUB_WORKSPACE} 
        
                   # install simmechanics-to-urdf and save the last commit SHA1 hash 
        
                   git clone $SIMMECHANICS_TO_URDF_REPOSITORY_URL 
        
                   cd simmechanics-to-urdf 
        
                   export SIMMECHANICS_TO_URDF_COMMIT=`git rev-parse HEAD` 
        
                   sudo python setup.py install 
        
                   cd ${GITHUB_WORKSPACE} 
        
                   # get C++ dependencies and save their last commit SHA1 hash 
        
                   # ycm 
        
                   git clone https://github.com/robotology/ycm.git 
        
                   cd ycm 
        
                   git checkout v0.12.0 
        
                   mkdir build 
        
                   cd build 
        
                   cmake .. 
        
                   sudo cmake --build . --target install 
        
                   cd ${GITHUB_WORKSPACE} 
        
                   ## yarp 
        
                   git clone https://github.com/robotology/yarp.git 
        
                   cd yarp 
        
                   git checkout v3.4.0 
        
                   export YARP_COMMIT=`git rev-parse HEAD` 
        
                   mkdir build 
        
                   cd build 
        
                   cmake -DCREATE_LIB_MATH:BOOL=OFF -DYARP_COMPILE_EXECUTABLES:BOOL=OFF .. 
        
                   sudo cmake --build . --target install 
        
                   cd ${GITHUB_WORKSPACE} 
        
                   ## idyntree 
        
                   git clone https://github.com/robotology/idyntree.git 
        
                   cd idyntree 
        
                   git checkout v1.2.0 
        
                   export IDYNTREE_COMMIT=`git rev-parse HEAD` 
        
                   mkdir build 
        
                   cd build 
        
                   cmake .. 
        
                   sudo cmake --build . --target install 
        
                   cd ${GITHUB_WORKSPACE} 
        
                   # Prepare icub-model-generator build 
        
                   mkdir build 
        
                   cd build 
        
                   cmake -DICUB_MODELS_SOURCE_DIR=$ICUB_MODELS_SOURCE_DIR  .. 
        
                   # Build and run 
        
                   make VERBOSE=1 
        
                   ctest --output-on-failure

If you have any issue you can call me

pattacini · 2021-09-23T10:07:50Z

Hi @Nicogene

It is quite strange that r_ankle_2 is missing, it has been substituted by a link with a different name?

My understanding is that since some time the sole is split in two (a front and a rear part), thus the two distinct frames.

(and why just r_ankle_2 and not l_ankle_2?)

Maybe the CI stopped before processing the left ankle?

If I understood correctly this PR should just update some stls, but seems that something went wrong in the export of the simmechanics XML

It indeed remains unclear why the problem with r_ankle_2 just showed up now.

Nicogene · 2021-09-23T10:35:13Z

My understanding is that since some time the sole is split in two (a front and a rear part), thus the two distinct frames.

This is how the ft sensor was defined in iCub 2.5

icub-models-generator/simmechanics/data/icub2_5/ICUB_2-5_BB_simmechanics_options.yaml.in

Line 75 in 36ca1b9

SIM_SEA_2-5_R_FOOT--SIM_SEA_2-5_R_SOLE: r_foot_ft_sensor

Here instead how they are defined in iCub 3:

icub-models-generator/simmechanics/data/icub3/ICUB_3_all_options.yaml

Lines 76 to 77 in 36ca1b9

    
           SIM_ICUB3_R_ANKLE_2--SIM_ICUB3_R_FOOT_FRONT: r_foot_front_ft_sensor 
        
           SIM_ICUB3_R_ANKLE_2--SIM_ICUB3_R_FOOT_REAR: r_foot_rear_ft_sensor

As you can see the r_ankle_2 has been always involved in iCub 3

pattacini · 2021-09-23T10:53:42Z

So the problem might be that we need the frame, which has been somehow removed then.
@Mick3Lozzo

Mick3Lozzo · 2021-09-23T13:04:11Z

Re-added the frame in both feet @pattacini.

Now we should clarify their placement inside the assembly.

Probably they should be here ad not inside a suppressed group, but I'll wait for @salvi-mattia, maybe there is a reason.

Nicogene

If the model works on gazebo, we can merge it 👍🏻

pattacini · 2021-09-23T14:34:13Z

If the model works on gazebo, we can merge it 👍🏻

@mfussi66
The CI now works ok 👍🏻
Do you have feedback from the test with Gazebo?

mfussi66 · 2021-09-23T14:42:09Z

If the model works on gazebo, we can merge it 👍🏻

@mfussi66
The CI now works ok 👍🏻
Do you have feedback from the test with Gazebo?

Not yet, as I need to check how to generate it from the proposed model.

mfussi66 · 2021-09-23T16:21:09Z

Yep it seems to be displayed correctly in gazebo!

Nicogene · 2021-09-24T07:11:40Z

Merged, thanks!

Updated icub3 with feet covers

58f67f8

pattacini assigned mfussi66 Sep 22, 2021

pattacini requested review from Nicogene, S-Dafarra and GiulioRomualdi September 22, 2021 14:28

Nicogene previously approved these changes Sep 22, 2021

View reviewed changes

Nicogene requested a review from traversaro September 22, 2021 14:46

Nicogene self-requested a review September 23, 2021 09:42

Fix reference frames on ankles

d6b2db3

Nicogene approved these changes Sep 23, 2021

View reviewed changes

Nicogene merged commit 924a1e8 into robotology:master Sep 24, 2021

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Update iCub3 with feet covers #210

Update iCub3 with feet covers #210

mfussi66 commented Sep 22, 2021

Nicogene left a comment

Nicogene commented Sep 22, 2021

mfussi66 commented Sep 22, 2021

pattacini commented Sep 22, 2021

Mick3Lozzo commented Sep 22, 2021

pattacini commented Sep 22, 2021

S-Dafarra commented Sep 22, 2021

pattacini commented Sep 22, 2021 •

edited

Loading

S-Dafarra commented Sep 22, 2021

mfussi66 commented Sep 23, 2021

Nicogene commented Sep 23, 2021 •

edited

Loading

mfussi66 commented Sep 23, 2021 •

edited

Loading

Nicogene commented Sep 23, 2021 •

edited

Loading

pattacini commented Sep 23, 2021

Nicogene commented Sep 23, 2021

pattacini commented Sep 23, 2021

Mick3Lozzo commented Sep 23, 2021

Nicogene left a comment

pattacini commented Sep 23, 2021

mfussi66 commented Sep 23, 2021

mfussi66 commented Sep 23, 2021

Nicogene commented Sep 24, 2021

Update iCub3 with feet covers #210

Update iCub3 with feet covers #210

Conversation

mfussi66 commented Sep 22, 2021

Nicogene left a comment

Choose a reason for hiding this comment

Nicogene commented Sep 22, 2021

mfussi66 commented Sep 22, 2021

pattacini commented Sep 22, 2021

Mick3Lozzo commented Sep 22, 2021

pattacini commented Sep 22, 2021

S-Dafarra commented Sep 22, 2021

pattacini commented Sep 22, 2021 • edited Loading

S-Dafarra commented Sep 22, 2021

mfussi66 commented Sep 23, 2021

Nicogene commented Sep 23, 2021 • edited Loading

mfussi66 commented Sep 23, 2021 • edited Loading

Nicogene commented Sep 23, 2021 • edited Loading

pattacini commented Sep 23, 2021

Nicogene commented Sep 23, 2021

pattacini commented Sep 23, 2021

Mick3Lozzo commented Sep 23, 2021

Nicogene left a comment

Choose a reason for hiding this comment

pattacini commented Sep 23, 2021

mfussi66 commented Sep 23, 2021

mfussi66 commented Sep 23, 2021

Nicogene commented Sep 24, 2021

pattacini commented Sep 22, 2021 •

edited

Loading

Nicogene commented Sep 23, 2021 •

edited

Loading

mfussi66 commented Sep 23, 2021 •

edited

Loading

Nicogene commented Sep 23, 2021 •

edited

Loading