Add the iCub 3 model - Githubissues

S-Dafarra commented 4 years ago

In order to start performing experiments with the iCub 3 robot, it is fundamental to have its urdf and the possibility of using it in Gazebo.

With this issue, I would like to track the generation and the issues arising when trying to use this model for the first time.

cc @pattacini, @traversaro, @fiorisi

pattacini commented 4 years ago

Are we maybe discussing about 3 different conventions?

the "yellow-arrow";
the "iCub 2.5 legs";
the one proposed in https://github.com/robotology/icub-models/issues/41#issuecomment-694780076.

That one @traversaro seems to support is not the so-called "yellow-arrow" but rather that one already implemented in the iCub 2.5 legs. On the other hand, @salvi-mattia, yours is more "traditional".

Did I get this correct?

traversaro commented 4 years ago

Did I get this correct?

Exactly. The main difference between "yellow-arrow" and "iCub 2.5 legs" is at which level you try to have link frame all with the same orientation. For "yellow-arrow" it is at the robot level (that given iCub shape leads to peculiar behavior) while "iCub 2.5 legs" is just to keep the same orientation for all the links in a given limb.

X axis on the link main axis [pointing next joint]

Small note: with this convention you still have have a small ambiguity (the same that you have in DH) when assigning frames for a joint whose axis coincide (such as the ankle or hip in iCub).

kouroshD commented 4 years ago

Humanoid models that follow the convention "all frame have the same orientation in a given configuration):

I agree with @traversaro . Back in time, when I was using the different robots urdf for the retargeting purposes, I had this problem with iCub2.5 urdf model. Instead, other robots were straight forward to do retargeting.

I do not agree with this comment https://github.com/robotology/icub-models/issues/41#issuecomment-694780076 , since the robot motor axis can be different, and we will end up with urdf model that every link has different frame orientations. But instead, putting all the joint to N-pose or T-pose (or zero values for all the motor joint angles) we will have a much cleaner urdf of the robot.

traversaro commented 4 years ago

Hi @S-Dafarra @kouroshD , if you could discuss internally with other DIC people and reach a consensus on this it would be great, otherwise the risk is that the discussion become a bit too dispersive.

S-Dafarra commented 4 years ago

Hi @S-Dafarra @kouroshD , if you could discuss internally with other DIC people and reach a consensus on this it would be great, otherwise the risk is that the discussion become a bit too dispersive.

Sorry to show up just now. We discussed internally and we agreed on the following. The convention we prefer is defined as follows:

When all the joints are in the zero position, all the links belonging to a limb share the same orientation. E.g., all the links of the torso would be aligned with frame A, all the links of the left arm with frame B, the right arm with frame C,....
These "common" frames (A, B, C, ...) have the z-axis pointing upward, the x-axis pointing forward (clearly, upward and forward depend on the limb)

Given the particular design of the shoulders, in the zero position, a frame on an arm link is slightly rotated around the z-axis compared to a frame in the torso.

This convention is basically the one defined as "iCub 2.5 legs" in https://github.com/robotology/icub-models/issues/41#issuecomment-694884201, which corresponds to the blue arrow of https://github.com/robotology/icub-models/issues/41#issuecomment-694738366.

salvi-mattia commented 4 years ago

after a meeting with @S-Dafarra @FabioBergonti and @pattacini we agreed on the following:

every link has an attached frame with:
- X axis pointing forward
- Z axis pointing up
- Y axis to complete the right-handed triplet
one of the axes has to be coincident with the joint axis

when it's not applicable (in plain words, we may derive a better formalism later):

one axis is coincident with the joint axis
Z axis is the one pointing upwards at the most
X axis pointing forward at the most
Y axis to complete the right-handed triplet

salvi-mattia commented 4 years ago

immagine

salvi-mattia commented 4 years ago

I added the frames in the linke and reassembled the mechanism. Everything is pushed in cad-mechanics @Nicogene u can now make your magic-button-stuff

after that it could be useful to check if we are getting what we want, so that we can proceed with the legs in the correct way

Nicogene commented 4 years ago

@salvi-mattia added and checked the link frames but 4 eyes are better than two 😉

@S-Dafarra could you please double-check if the frames are following the right convention?

In icub3-upperbody.zip you should find all you need for opening the urdf in gazebo. Please let me know if you experience any issue.

S-Dafarra commented 4 years ago

@salvi-mattia added and checked the link frames but 4 eyes are better than two

@S-Dafarra could you please double-check if the frames are following the right convention?

In icub3-upperbody.zip you should find all you need for opening the urdf in gazebo. Please let me know if you experience any issue.

Hi @Nicogene, I just tried, but I am not able to load the model. I get the following error

[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [sim_icub3_root_link_prt.stl]
[Err] [Visual.cc:2951] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [sim_icub3_torso_1_prt.stl]
[Err] [Visual.cc:2951] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [sim_icub3_torso_2_prt.stl]
[Err] [Visual.cc:2951] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [sim_icub3_chest_prt.stl]
[Err] [Visual.cc:2951] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [sim_icub3_l_shoulder_1_prt.stl]
[Err] [Visual.cc:2951] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [sim_icub3_l_shoulder_2_prt.stl]
[Err] [Visual.cc:2951] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [sim_icub3_l_shoulder_3_prt.stl]
[Err] [Visual.cc:2951] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [sim_icub3_l_upperarm_prt.stl]
[Err] [Visual.cc:2951] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [sim_icub3_r_shoulder_1_prt.stl]
[Err] [Visual.cc:2951] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [sim_icub3_r_shoulder_2_prt.stl]
[Err] [Visual.cc:2951] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [sim_icub3_r_shoulder_3_prt.stl]
[Err] [Visual.cc:2951] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [sim_icub3_r_upperarm_prt.stl]
[Err] [Visual.cc:2951] No mesh specified

I just specified the GAZEBO_MODEL_PATH, should I add something else?

Nicogene commented 4 years ago

For simplicity run gazebo in the folder where the stl files are located.

This issue should not occur once CMake will install them in the right location.

S-Dafarra commented 4 years ago

For simplicity run gazebo in the folder where the stl files are located.

This issue should not occur once CMake will install them in the right location.

Thanks, that worked

@S-Dafarra could you please double-check if the frames are following the right convention?

Actually, there seems to be a problem. It seems that the shoulders motor axis are not aligned to any axis

It seems it is not following what was defined in https://github.com/robotology/icub-models/issues/41#issuecomment-697561933 but rather the "yellow arrow" convention of https://github.com/robotology/icub-models/issues/41#issuecomment-694738366

Also, the y axis (green one) is forward, not the x as defined in https://github.com/robotology/icub-models/issues/41#issuecomment-697561933

Nicogene commented 4 years ago

Sorry probably I used the wrong yaml in the generation 🙏 I am still learning the procedure Stay tuned !

salvi-mattia commented 4 years ago

oh, well, my fault too I named all the frame CSYS_<child_link_name>. instead of SCSYS_<child_link_name>.

lets have a check with ths wrong naming convention and then I'll fix it [if we think it's worth it]

Nicogene commented 4 years ago

Now seems ok to me

Please @S-Dafarra could you check it now?

icub3-upperbody-final.zip

immagine

S-Dafarra commented 4 years ago

Now seems ok to me

Please @S-Dafarra could you check it now?

icub3-upperbody-final.zip

Looks good to me, thanks a lot!

cc @kouroshD

salvi-mattia commented 4 years ago

Plot twist, the feet mounted on iCub3 are the same that in the cad: onf single FTsensor and passvie fingers with spring. I'm aware that exist a new project with different feet [not mounted] Which is the state of this enhancement? @fiorisi @Fabrizio69 @Mick3Lozzo

which kind of feet do we want in the URDF? @S-Dafarra

S-Dafarra commented 4 years ago

which kind of feet do we want in the URDF? @S-Dafarra

Preferably the one of the new project, which should be rectangular without the passive toe, holding two FTs.

Fabrizio69 commented 4 years ago

Which is the state of this enhancement? @fiorisi @Fabrizio69 @Mick3Lozzo

The state is that the new feet are ready to be mounted on the robot (see https://github.com/icub-tech-iit/tickets/issues/224#issuecomment-645229863). There are three possible sizes for each foot plate: 100 mm 110 mm 120 mm The simplified feet replacing kit is cod.14658 - IG_022_A_001. At the moment the kit is not mounted on the iCub3 CAD.

Nicogene commented 4 years ago

🔔 Just a quick update, we managed to create the urdfs and import in gazebo the left and right legs, the corresponding files are respectively in this and in this commit.

RIGHT LEG	LEFT LEG

In the near future, we will work on the assemble of them to the torso

cc @maggia80 @DanielePucci @pattacini @salvi-mattia @S-Dafarra

Nicogene commented 4 years ago

The https://github.com/robotology/icub-model-generator/pull/133 has been merged.

The model should be available

This is a first draft of the model, it has some open points/issues that we worked-around but I think it is something usable.

Here it is the iCub 3 - Venere di Milo edition

immagine

We added also the frames l_sole and r_sole as requested.

🔘 Click here to see the output of `idyntree-model-info`

``` Model: Links: [0] root_link [1] l_hip_1 [2] r_hip_1 [3] torso_1 [4] torso_2 [5] chest [6] l_shoulder_1 [7] r_shoulder_1 [8] neck_1 [9] neck_2 [10] neck_3 [11] head [12] r_shoulder_2 [13] r_shoulder_3 [14] r_upper_arm [15] l_shoulder_2 [16] l_shoulder_3 [17] l_upper_arm [18] r_hip_2 [19] r_hip_3 [20] r_upper_leg [21] r_lower_leg [22] r_ankle_1 [23] r_ankle_2 [24] r_foot_front [25] r_foot_rear [26] l_hip_2 [27] l_hip_3 [28] l_upper_leg [29] l_lower_leg [30] l_ankle_1 [31] l_ankle_2 [32] l_foot_rear [33] l_foot_front Frames: [34] base_link --> root_link [35] r_sole --> r_foot_rear [36] l_sole --> l_foot_rear [37] l_arm_ft_sensor --> l_upper_arm [38] r_arm_ft_sensor --> r_upper_arm [39] l_leg_ft_sensor --> l_hip_3 [40] l_foot_front_ft_sensor --> l_foot_front [41] l_foot_rear_ft_sensor --> l_foot_rear [42] r_leg_ft_sensor --> r_hip_3 [43] r_foot_front_ft_sensor --> r_foot_front [44] r_foot_rear_ft_sensor --> r_foot_rear Joints: [0] l_hip_pitch (dofs: 1) : root_link<-->l_hip_1 [1] r_hip_pitch (dofs: 1) : root_link<-->r_hip_1 [2] torso_roll (dofs: 1) : root_link<-->torso_1 [3] torso_pitch (dofs: 1) : torso_1<-->torso_2 [4] torso_yaw (dofs: 1) : torso_2<-->chest [5] l_shoulder_pitch (dofs: 1) : chest<-->l_shoulder_1 [6] r_shoulder_pitch (dofs: 1) : chest<-->r_shoulder_1 [7] neck_pitch (dofs: 1) : neck_1<-->neck_2 [8] neck_roll (dofs: 1) : neck_2<-->neck_3 [9] neck_yaw (dofs: 1) : neck_3<-->head [10] r_shoulder_roll (dofs: 1) : r_shoulder_1<-->r_shoulder_2 [11] r_arm_ft_sensor (dofs: 1) : r_shoulder_3<-->r_upper_arm [12] l_shoulder_roll (dofs: 1) : l_shoulder_1<-->l_shoulder_2 [13] l_arm_ft_sensor (dofs: 1) : l_shoulder_3<-->l_upper_arm [14] r_hip_roll (dofs: 1) : r_hip_1<-->r_hip_2 [15] r_hip_yaw (dofs: 1) : r_hip_3<-->r_upper_leg [16] r_knee (dofs: 1) : r_upper_leg<-->r_lower_leg [17] r_ankle_pitch (dofs: 1) : r_lower_leg<-->r_ankle_1 [18] r_ankle_roll (dofs: 1) : r_ankle_1<-->r_ankle_2 [19] l_hip_roll (dofs: 1) : l_hip_1<-->l_hip_2 [20] l_hip_yaw (dofs: 1) : l_hip_3<-->l_upper_leg [21] l_knee (dofs: 1) : l_upper_leg<-->l_lower_leg [22] l_ankle_pitch (dofs: 1) : l_lower_leg<-->l_ankle_1 [23] l_ankle_roll (dofs: 1) : l_ankle_1<-->l_ankle_2 [24] neck_fixed_joint (dofs: 0) : chest<-->neck_1 [25] r_shoulder_yaw (dofs: 0) : r_shoulder_2<-->r_shoulder_3 [26] l_shoulder_yaw (dofs: 0) : l_shoulder_2<-->l_shoulder_3 [27] r_leg_ft_sensor (dofs: 0) : r_hip_2<-->r_hip_3 [28] r_foot_front_ft_sensor (dofs: 0) : r_ankle_2<-->r_foot_front [29] r_foot_rear_ft_sensor (dofs: 0) : r_ankle_2<-->r_foot_rear [30] l_leg_ft_sensor (dofs: 0) : l_hip_2<-->l_hip_3 [31] l_foot_rear_ft_sensor (dofs: 0) : l_ankle_2<-->l_foot_rear [32] l_foot_front_ft_sensor (dofs: 0) : l_ankle_2<-->l_foot_front ```

You should find it on devel branch.

I would keep this issue open until we will have the complete model.

cc @S-Dafarra @DanielePucci @maggia80 @salvi-mattia @pattacini @fiorisi @traversaro @FabioBergonti

S-Dafarra commented 4 years ago

Thanks a lot @Nicogene @salvi-mattia and every one, this is something from which we can start!

S-Dafarra commented 4 years ago

I have tried to change the branch and install all robots. I can see iCubGenova09 in the list of models, but it fails to load it. These are the errors I get

[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [urdf file/urdf file/sim_icub3_root_link_prt.stl]
[Err] [Visual.cc:2953] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [urdf file/urdf file/sim_icub3_torso_1_prt.stl]
[Err] [Visual.cc:2953] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [urdf file/urdf file/sim_icub3_torso_2_prt.stl]
[Err] [Visual.cc:2953] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [urdf file/urdf file/sim_icub3_chest_prt.stl]
[Err] [Visual.cc:2953] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [urdf file/urdf file/sim_icub3_l_shoulder_1_prt.stl]
[Err] [Visual.cc:2953] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [urdf file/urdf file/sim_icub3_l_shoulder_2_prt.stl]
[Err] [Visual.cc:2953] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [urdf file/urdf file/sim_icub3_l_shoulder_3_prt.stl]
[Err] [Visual.cc:2953] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [urdf file/urdf file/sim_icub3_l_upperarm_prt.stl]
[Err] [Visual.cc:2953] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [urdf file/urdf file/sim_icub3_r_shoulder_1_prt.stl]
[Err] [Visual.cc:2953] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [urdf file/urdf file/sim_icub3_r_shoulder_2_prt.stl]
[Err] [Visual.cc:2953] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [urdf file/urdf file/sim_icub3_r_shoulder_3_prt.stl]
[Err] [Visual.cc:2953] No mesh specified
[Wrn] [SystemPaths.cc:459] File or path does not exist [""] [urdf file/urdf file/sim_icub3_r_upperarm_prt.stl]
[Err] [Visual.cc:2953] No mesh specified

traversaro commented 4 years ago

I have tried to change the branch and install all robots. I can see iCubGenova09 in the list of models, but it fails to load it.

I guess in the yaml of simmechanics-to-urdf the filenameformatchangeext used to specify where the meshes are installing is missing, so the mesh elements in the URDF have <mesh filename="sim_icub3_torso_1_prt.stl" scale="0.001 0.001 0.001"/> instead of the correct path as in the iCub 2 URDFs.

Nicogene commented 4 years ago

The stl problem should be fixed by https://github.com/robotology/icub-model-generator/pull/145.

(Tested it successfully)

traversaro commented 4 years ago

The stl problem should be fixed by robotology/icub-model-generator#145.

(Tested it successfully)

Merged! The deployment time from icub-model-generator to icub-models can be probably be improved by migrating from Travis to GitHub Actions and using caching, but that's another story.

traversaro commented 4 years ago

The stl problem should be fixed by robotology/icub-model-generator#145. (Tested it successfully)

Merged! The deployment time from icub-model-generator to icub-models can be probably be improved by migrating from Travis to GitHub Actions and using caching, but that's another story.

Model deployed after ~30 minutes, ~13 minutes of queue and 17 minutes of CI: https://github.com/robotology/icub-models/commit/29f5fddad7f9da4dbebc67e2fac13f9dcb2aaceb .

S-Dafarra commented 4 years ago

I am still not able to load the model in gazebo. iCubV3 loads only the upperbody.

Edit: It loads only the upper-body.

traversaro commented 4 years ago

Just to check, does iCubGenova09 works well instead? If you run gazebo --verbose reveals something?

traversaro commented 4 years ago

Note that there still present the regression https://github.com/robotology/icub-models/issues/50 that could create problems.

S-Dafarra commented 4 years ago

Just to check, does iCubGenova09 works well instead? If you run gazebo --verbose reveals something?

After a couple of reinstallation of the robotology-superbuild I see this

The same for iCubV3. Here the output of the terminal

sdafarra@iiticublap104:~$ gazebo -slibgazebo_yarp_clock.so --verbose
Gazebo multi-robot simulator, version 11.2.0
Copyright (C) 2012 Open Source Robotics Foundation.
Released under the Apache 2 License.
http://gazebosim.org

[Msg] Waiting for master.
Gazebo multi-robot simulator, version 11.2.0
Copyright (C) 2012 Open Source Robotics Foundation.
Released under the Apache 2 License.
http://gazebosim.org

[DEBUG] GazeboYarpClock loaded. Clock port will be  /clock
[Msg] Waiting for master.
[Msg] Connected to gazebo master @ http://127.0.0.1:11345
[Msg] Publicized address: 192.168.1.84
[INFO] |yarp.os.Port| Port /clock active at tcp://192.168.1.84:10002/
[INFO] |yarp.os.Port| Port /clock/rpc active at tcp://192.168.1.84:10003/
[Msg] Connected to gazebo master @ http://127.0.0.1:11345
[Msg] Publicized address: 192.168.1.84
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/isaac/isaac-wp1/EVAmodelFiles/generatedUrdfs/robots/isaacEvaYarp_all"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/isaac/isaac-wp1/EVAmodelFiles/generatedUrdfs"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/isaac/isaac-wp1/EVAmodelFiles/leo_description"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/robotology-superbuild/build/install/share/gazebo/models/.git"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/robotology-superbuild/build/install/share/iCub/robots/icubGazeboSim"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/robotology-superbuild/build/install/share/BlockFactory"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/robotology-superbuild/build/install/share/ICUBcontrib"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/robotology-superbuild/build/install/share/WBToolbox"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/robotology-superbuild/build/install/share/YCM"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/robotology-superbuild/build/install/share/bash-completion"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/robotology-superbuild/build/install/share/gazebo"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/robotology-superbuild/build/install/share/iCub"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/robotology-superbuild/build/install/share/robotology-superbuild"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/robotology-superbuild/build/install/share/wbc"
[Err] [InsertModelWidget.cc:402] Missing model.config for model "/home/sdafarra/Software/robotology-superbuild/build/install/share/yarp"
[INFO] |yarp.devices.AnalogWrapper| No ROS group found in config file ... skipping ROS initialization.
[INFO] |yarp.os.Port| Port /icubSim/left_arm/analog:o/rpc:i active at tcp://192.168.1.84:10004/
[INFO] |yarp.devices.AnalogWrapper| AnalogServer : no ROS initialization required
[INFO] |yarp.dev.PolyDriver| Created wrapper <analogServer>. See C++ class AnalogWrapper for documentation.
[INFO] |yarp.dev.PolyDriver| Created device <gazebo_forcetorque>. See C++ class GazeboYarpForceTorqueDriver for documentation.
[INFO] |yarp.os.Port| Port /icubSim/left_arm/analog:o active at tcp://192.168.1.84:10005/
[Wrn] [msgs.cc:1842] Conversion of sensor type[force_torque] not supported.
[Wrn] [msgs.cc:1842] Conversion of sensor type[force_torque] not supported.
[INFO] |yarp.devices.AnalogWrapper| No ROS group found in config file ... skipping ROS initialization.
[INFO] |yarp.os.Port| Port /icubSim/right_arm/analog:o/rpc:i active at tcp://192.168.1.84:10006/
[INFO] |yarp.devices.AnalogWrapper| AnalogServer : no ROS initialization required
[INFO] |yarp.dev.PolyDriver| Created wrapper <analogServer>. See C++ class AnalogWrapper for documentation.
[INFO] |yarp.dev.PolyDriver| Created device <gazebo_forcetorque>. See C++ class GazeboYarpForceTorqueDriver for documentation.
[INFO] |yarp.os.Port| Port /icubSim/right_arm/analog:o active at tcp://192.168.1.84:10007/
[Wrn] [msgs.cc:1842] Conversion of sensor type[force_torque] not supported.
[Wrn] [msgs.cc:1842] Conversion of sensor type[force_torque] not supported.
[ERROR] |yarp.devices.ControlBoardWrapper| Parameter networks use deprecated syntax
[INFO] |yarp.devices.ControlBoardWrapper| /icubSim/torso : no ROS initialization required
[INFO] |yarp.devices.ControlBoardWrapper| /icubSim/torso  initting YARP initialization
[INFO] |yarp.os.Port| Port /icubSim/torso/rpc:i active at tcp://192.168.1.84:10008/
[INFO] |yarp.os.Port| Port /icubSim/torso/command:i active at tcp://192.168.1.84:10009/
[INFO] |yarp.os.Port| Port /icubSim/torso/state:o active at tcp://192.168.1.84:10010/
[INFO] |yarp.os.Port| Port /icubSim/torso/stateExt:o active at tcp://192.168.1.84:10011/
[INFO] |yarp.dev.PolyDriver| Created wrapper <controlboardwrapper2>. See C++ class ControlBoardWrapper for documentation.
[DEBUG] found:  torso_yaw torso_yaw
[DEBUG] found:  torso_pitch torso_pitch
[DEBUG] found:  torso_roll torso_roll
[DEBUG] done
[WARNING] No positionToleranceLinear value found in ini file, default one will be used!
[WARNING] No positionToleranceRevolute value found in ini file, default one will be used!
[WARNING] No max torques value found in ini file, default one will be used!
[DEBUG] max_torques: [   2000.000000     2000.000000     2000.000000  ]
[INFO] min_stiffness param found!
[INFO] max_stiffness param found!
[INFO] min_damping param found!
[INFO] max_damping param found!
[DEBUG] min_stiffness: [   0.000000  0.000000    0.000000  ]
[DEBUG] max_stiffness: [   1000.000000   1000.000000     1000.000000  ]
[DEBUG] min_damping: [   0.000000    0.000000    0.000000  ]
[DEBUG] max_damping: [   100.000000  100.000000  100.000000  ]
[WARNING] VELOCITY_CONTROL: 'velocityControlImplementationType' param missing.  Using the default value of 'direct_velocity_pid', but the parameter will be compulsory gazebo-yarp-plugins 4.
[DEBUG] Initializing Trajectory Generator with current values
[INFO] |yarp.dev.PolyDriver| Created device <gazebo_controlboard>. See C++ class GazeboYarpControlBoardDriver for documentation.
[Wrn] [Publisher.cc:136] Queue limit reached for topic /gazebo/default/user_camera/pose, deleting message. This warning is printed only once.

traversaro commented 4 years ago

Indeed, by directly inspecting the iCubGazebo9 and iCubV3 models it is possible to see that those models only contain the upper body links, no sign of the lower bodies.

traversaro commented 4 years ago

It seems that the iCubV3 and iCubGazebo9 models are both generated from the upperbody models, see the files used in: https://github.com/robotology/icub-model-generator/blob/master/simmechanics/CMakeLists.txt#L285 . Probably we should switch that to use the all body option, using the files provided in https://github.com/robotology/icub-model-generator/tree/master/simmechanics/data/icub3 ?

Nicogene commented 4 years ago

Sorry, I didn't expect that the deployment gave so many troubles, I couldn't check these issues since I always worked inside icub-model-generators, I will work with @traversaro in order to fix the model inside icub-models, in the meanwhile here is the model if you need to try it. Probably you would need to copy by hand the conf files and meshes in the installation folders.

icub3-all.zip

S-Dafarra commented 4 years ago

I can load the iCubV3 model now :tada: :tada:

It seems that the arms are not controlled though

Nicogene commented 4 years ago

It seems that the arms are not controlled though

Yes because in the asm we have only the shoulders, we thought to create yet another conf file for controlling only the shoulders like right/left_arm_no_forearm.ini but in practice it may be not so useful since we are going to add the forearms relatively soon.

If this is problematic to you in any manner, then we can add those files ore make the shoulders fixed

S-Dafarra commented 4 years ago

Yes because in the asm we have only the shoulders, we thought to create yet another conf file for controlling only the shoulders like right/left_arm_no_forearm.ini but in practice it may be not so useful since we are going to add the forearms relatively soon.

Ah ok, I see. I think we are going to use it extensively starting from Wednesday, do you think that adding the forearms may require more time?

Nicogene commented 4 years ago

We decided to exploit the momentum and we added also the arms 💪

immagine

The PR is https://github.com/robotology/icub-model-generator/pull/147

The model changed the name from iCubV3 to iCubGazeboV3 since we had to fake the arms inertias for this issue https://github.com/robotology/icub-models/issues/53.

The model should be soon available on icub-models@devel.

cc @maggia80 @pattacini @S-Dafarra @DanielePucci @salvi-mattia @traversaro

pattacini commented 4 years ago

For the record, @GiulioRomualdi and @S-Dafarra managed to get iCub 3 (preliminarily) walking w/ the latest URDF ✨

icub3-walking

GiulioRomualdi commented 4 years ago

Here I drop some issues related to the generation of the model: https://github.com/robotology/icub-models/issues/60 https://github.com/robotology/icub-models/issues/61 https://github.com/robotology/icub-models/issues/57

Nicogene commented 4 years ago

This PR should(https://github.com/robotology/icub-model-generator/pull/163) reduce the weight of the feet, wrongly exported by the magic button

traversaro commented 4 years ago

I think the improvements on the iCub3 model will be constant for the next few years, so what do you think if we close this issue and open separate ones to track the particular improvements?

S-Dafarra commented 4 years ago

I think the improvements on the iCub3 model will be constant for the next few years, so what do you think if we close this issue and open separate ones to track the particular improvements?

Ok for me

traversaro commented 4 years ago

As agreed, let's close the issue, to check all the issues related to iCub3 it is possible to check out the icub3 label: https://github.com/robotology/icub-models/labels/icub3 .

robotology / icub-models

Add the iCub 3 model #41