Added files to add osqp, OsqpEigen, UnicyclePlanner and walking-controllers #87

S-Dafarra · 2018-06-13T11:43:24Z

Right now it is not compiling since https://github.com/robotology/capture-point-walking does not have a CMakeLists.txt in the root folder.

TODO:

Rename dcmWalking
Add ICUB as walking dependency
Modify the README to point to the walking tutorial.
Modify Buildosqp.cmake (See Added files to add osqp, OsqpEigen, UnicyclePlanner and walking-controllers #87 (comment))
~~- [ ] Add walking-teleoperation~~ (separate PR)
Fix Compilation fails with clang due to use of "auto" in method declaration walking-controllers#11
Fix Merge the dcmTrajectoryGenerator branch unicycle-footstep-planner#14
Merge the devel into master of the unicyclePlanner
Merge the devel into master of walking-controllers

cc @GiulioRomualdi

S-Dafarra · 2018-06-13T11:47:20Z

For what concerns the tutorial, we may modify and link the README of the walking https://github.com/robotology/capture-point-walking/blob/master/code/cpp/README.md

cmake/BuildUnicyclePlanner.cmake

traversaro · 2018-06-13T11:48:19Z

CMakeLists.txt

+  find_or_build_package(osqp)
+  find_or_build_package(OsqpEigen)
+  find_or_build_package(UnicyclePlanner)
+  find_or_build_package(icubWalkingDCM)


I think we can just leave this line (or at limit the UnicyclePlanner). All other projects will be automatically downloaded and compiled as dependency listed.

I see, even if I prefer to have the full list available in the CMakeLists to make sure at a first sight what is compiled.

I see, but I would at least avoid the external (osqp and qpOASES). Furthermore, for consistency I would put all the non-conditional find_or_build_package's call before the first if(ROBOTOLOGY_USES_GAZEBO).

traversaro

Comments.

cmake/BuildUnicyclePlanner.cmake

cmake/BuildicubWalkingDCM.cmake

traversaro · 2018-06-20T22:17:36Z

New iDynTree released, we can proceed with this PR.

traversaro · 2018-06-20T22:19:24Z

We should update the table of the ROBOTOLOGY_ENABLE_DYNAMICS repositories in https://github.com/robotology/robotology-superbuild#profile-cmake-options . No need to list all of them, just the important ones, like walking-controllers.

traversaro · 2018-07-11T22:06:35Z

README modified (at least for the link to walking-controllers repo) in S-Dafarra@f0a8f7b .

traversaro · 2018-07-11T22:09:06Z

Wow, I was not aware that now you can resolve conflicts in the web UI. I think this PR is ready to be merged, what do you think @S-Dafarra @GiulioRomualdi @hu-yue ? Do you want to wait humanoids deadline to avoid changes in the robot setup?

S-Dafarra · 2018-07-13T12:30:41Z

I think we have to address the compilation issues first. Also the unicyclePlanner needs to be in a specific branch for now. Do you think this is an issue?

traversaro · 2018-07-13T13:44:20Z

Also the unicyclePlanner needs to be in a specific branch for now. Do you think this is an issue?

Yes. Until master of everything compiles with master of everything (except for the temporary YARP 2.3.72.1 situation) I would avoid putting stuff on the superbuild.

S-Dafarra · 2018-07-13T13:52:21Z

Ok then let's merge robotology/unicycle-footstep-planner#15 so then it should be easier to implement the modifications for the walking. When it is ready we can merge devel in master (of the unicyclePlanner repo) and merge also this. Do you agree?

traversaro · 2018-07-13T16:38:53Z

Ok, re-WIPPING for now.

traversaro · 2018-08-23T08:09:29Z

I think after the latest modification in OSQP v0.4.0, the Buildosqp.cmake module needs to be updated. See osqp/osqp#82 and robotology/osqp-eigen#10 .

S-Dafarra · 2018-08-23T09:57:42Z

How the --recursive option can be used from within YCM, i.e., what am I supposed to add/modify? 😅

traversaro · 2018-08-23T10:11:48Z

If this suggestion osqp/osqp#85 get accepted, I guess the proper way is to add also qdldl to the superbuild. That said, I checked the ExternalProject docs and the option is:

GIT_SUBMODULES <module>...
    Specific git submodules that should also be updated. If this option is not provided, all git submodules will be updated.

so perhaps everything already works, we just need to try.

traversaro

Still a few issues pending.

traversaro · 2018-12-07T09:39:13Z

We need to try, but I do not think that the submodules in osqp will be a problem, if anyone was considering that a blocking issue @GiulioRomualdi @S-Dafarra .

S-Dafarra · 2018-12-19T11:21:46Z

Related PRs:

traversaro · 2019-02-01T00:44:32Z

I removed in a fork the .gitattributes file of osqp, and apparently this fixed the problem on osqp. Given that we are going to drop support for Ubuntu 16.04 soonish, I am inclined to just merge this with osqp pointing to our fork, and then we can switch to osqp's upstream when we drop support for 16.04 .

Given the existing limitations of the CI, I think the last step is to test this branch on Windows and macOS at least once.

traversaro · 2019-02-04T09:20:56Z

@DanielePucci I am afraid that until we have someone to test this once on macOS and Windows we are blocked here.

GiulioRomualdi · 2019-02-04T09:24:15Z

For Windows, we can test the PR in the virtualizer-machine. What do you think @kouroshD?

traversaro · 2019-02-04T09:30:00Z

It would be great @GiulioRomualdi and @kouroshD !
For macOS I hope to be able to test on the shared machine.

lrapetti · 2019-02-04T10:00:32Z

@traversaro I can test it on MacOS

lrapetti · 2019-02-04T10:21:36Z

I was able to compile the walking-controllers.

I am trying to run the controller tutorial as explained in https://github.com/robotology/walking-controllers/#computer-how-to-run-the-simulation. The preparation seems to work but it fails when running startWalking:

[INFO][updateModule] The robot is prepared. 
[INFO]IK: 0.480500 ms MPC: 0.985300 ms Total: 2.122200 ms  
[INFO]IK: 0.395200 ms MPC: 0.576200 ms Total: 1.531800 ms  
[INFO]IK: 0.377800 ms MPC: 0.608600 ms Total: 1.407900 ms  
[INFO]IK: 0.452900 ms MPC: 0.660700 ms Total: 1.975400 ms  
[INFO]IK: 0.415400 ms MPC: 0.604500 ms Total: 1.998700 ms  
[INFO]IK: 0.341000 ms MPC: 0.616200 ms Total: 1.913000 ms  
[INFO]IK: 0.462700 ms MPC: 0.673000 ms Total: 1.965600 ms  
[INFO]IK: 0.415600 ms MPC: 0.602600 ms Total: 1.879800 ms  
[INFO]IK: 0.455100 ms MPC: 0.640000 ms Total: 1.879000 ms  
[INFO]IK: 0.387200 ms MPC: 0.606400 ms Total: 2.141300 ms  
[INFO]IK: 0.507900 ms MPC: 0.636700 ms Total: 2.193500 ms  
[INFO]IK: 0.452700 ms MPC: 0.638700 ms Total: 2.284500 ms  
[INFO]IK: 0.406800 ms MPC: 0.567100 ms Total: 2.223800 ms  
[INFO]IK: 0.424200 ms MPC: 0.554700 ms Total: 2.228600 ms  
[INFO]IK: 0.441600 ms MPC: 0.622400 ms Total: 2.149900 ms  
[INFO]IK: 0.491400 ms MPC: 0.595800 ms Total: 2.310500 ms  
[INFO]IK: 0.369900 ms MPC: 0.591100 ms Total: 2.203100 ms  
[INFO]IK: 0.452900 ms MPC: 0.607800 ms Total: 2.227200 ms  
[INFO]IK: 0.405900 ms MPC: 0.632400 ms Total: 2.148500 ms  
[INFO]IK: 0.389400 ms MPC: 0.622600 ms Total: 2.152400 ms  
[INFO]IK: 0.390700 ms MPC: 0.662400 ms Total: 2.247600 ms  
[ERROR][getFeedbacks] The following readings failed: 
[ERROR]	 - Left wrench 
[ERROR]	 - Right wrench 
[ERROR][RobotHelper::getFeedbacks] Unable to get the feedback from the robot 
[ERROR][updateModule] Unable to get the feedback. 
[INFO]RFModule closing.
yarp: Removing input from /wholeBodyDynamics/right_foot/cartesianEndEffectorWrench:o to /walking-coordinator/rightFootWrench:i
yarp: Removing input from /wholeBodyDynamics/left_foot/cartesianEndEffectorWrench:o to /walking-coordinator/leftFootWrench:i
yarp: Removing output from /walking-coordinator/remoteControlBoard/icubSim/torso/rpc:o to /icubSim/torso/rpc:i
yarp: Removing output from /walking-coordinator/remoteControlBoard/icubSim/torso/command:o to /icubSim/torso/command:i
yarp: Removing input from /icubSim/torso/stateExt:o to /walking-coordinator/remoteControlBoard/icubSim/torso/stateExt:i
yarp: Removing output from /walking-coordinator/remoteControlBoard/icubSim/left_arm/rpc:o to /icubSim/left_arm/rpc:i
yarp: Removing output from /walking-coordinator/remoteControlBoard/icubSim/left_arm/command:o to /icubSim/left_arm/command:i
yarp: Removing input from /icubSim/left_arm/stateExt:o to /walking-coordinator/remoteControlBoard/icubSim/left_arm/stateExt:i
yarp: Removing output from /walking-coordinator/remoteControlBoard/icubSim/right_arm/rpc:o to /icubSim/right_arm/rpc:i
yarp: Removing output from /walking-coordinator/remoteControlBoard/icubSim/right_arm/command:o to /icubSim/right_arm/command:i
yarp: Removing input from /icubSim/right_arm/stateExt:o to /walking-coordinator/remoteControlBoard/icubSim/right_arm/stateExt:i
yarp: Removing output from /walking-coordinator/remoteControlBoard/icubSim/left_leg/rpc:o to /icubSim/left_leg/rpc:i
yarp: Removing output from /walking-coordinator/remoteControlBoard/icubSim/left_leg/command:o to /icubSim/left_leg/command:i
yarp: Removing input from /icubSim/left_leg/stateExt:o to /walking-coordinator/remoteControlBoard/icubSim/left_leg/stateExt:i
yarp: Removing output from /walking-coordinator/remoteControlBoard/icubSim/right_leg/rpc:o to /icubSim/right_leg/rpc:i
yarp: Removing output from /walking-coordinator/remoteControlBoard/icubSim/right_leg/command:o to /icubSim/right_leg/command:i
yarp: Removing input from /icubSim/right_leg/stateExt:o to /walking-coordinator/remoteControlBoard/icubSim/right_leg/stateExt:i
[INFO]RFModule finished.

S-Dafarra · 2019-02-04T10:24:25Z

Thanks @lrapetti! Once I have encountered a similar issue when launching the WalkingModule without specifying the YARP_CLOCK. Can you confirm this is not the case?

lrapetti · 2019-02-04T10:34:17Z

Thanks @lrapetti! Once I have encountered a similar issue when launching the WalkingModule without specifying the YARP_CLOCK. Can you confirm this is not the case?

My bad, I probably forgot the YARP_CLOCK. I tested it again and the walking controller is working correctly!

traversaro · 2019-02-04T22:48:04Z

Great, we just need Windows to go (just if everything compiles correctly).

lrapetti · 2019-02-05T08:08:21Z

Great, we just need Windows to go (just if everything compiles correctly).

I may test Windows as well, but, if not urgent, I would prefer doing it tomorrow after demo deadline

traversaro · 2019-02-05T08:21:40Z

Sure!

GiulioRomualdi · 2019-02-05T08:25:11Z

I'm going to test on Windows today.

GiulioRomualdi · 2019-02-05T09:28:15Z

When I tried to compile osqp-eigen on Windows I retrieved the following error.

9>CMake Error at CMakeLists.txt:57 (find_package):
9>  By not providing "FindEigen3.cmake" in CMAKE_MODULE_PATH this project has
9>  asked CMake to find a package configuration file provided by "Eigen3", but
9>  CMake did not find one.
9>
9>  Could not find a package configuration file provided by "Eigen3" with any
9>  of the following names:
9>
9>    Eigen3Config.cmake
9>    eigen3-config.cmake
9>
9>  Add the installation prefix of "Eigen3" to CMAKE_PREFIX_PATH or set
9>  "Eigen3_DIR" to a directory containing one of the above files.  If "Eigen3"
9>  provides a separate development package or SDK, be sure it has been
9>  installed.

In the environment variable, there is EIGEN3_ROOT.
Probably the problem can be easily solved by coping FindEigen3.cmake inside the cmake folder of osqp-eigen.

See robotology/osqp-eigen#18

GiulioRomualdi · 2019-02-05T09:30:18Z

By the way isn't the problem solved here? @traversaro

GiulioRomualdi · 2019-02-05T09:44:53Z

Adding FindEigen3.cmake fixes the problem

GiulioRomualdi · 2019-02-05T09:49:02Z

Now I had another issue. Seems that there is a problem in the linking phase in WBToolbox. Actually I don't know why it's trying to compile it. Matlab is not installed in this machine and the flag ROBOTOLOGY_USES_MATLAB is false.

17>(Link target) ->
17>SimulatorSynchronizer.obj : error LNK2019: unresolved external symbol "class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const blockfactory::core::BlockOptionPrioritizeOrder" (?BlockOptionPrioritizeOrder@core@blockfactory@@3V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@B) referenced in function "public: virtual class std::vector<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >,class std::allocator<class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > > > __cdecl wbt::block::SimulatorSynchronizer::additionalBlockOptions(void)" (?additionalBlockOptions@SimulatorSynchronizer@block@wbt@@UEAA?AV?$vector@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@V?$allocator@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@2@@std@@XZ) [C:\Users\icub\Documents\icub_ws\robotology-superbuild\build\robotology\WBToolbox\toolbox\library\WBToolbox.vcxproj]
17>C:\Users\icub\Documents\icub_ws\robotology-superbuild\build\robotology\WBToolbox\bin\Release\WBToolbox.dll : fatal error LNK1120: 1 unresolved externals [C:\Users\icub\Documents\icub_ws\robotology-superbuild\build\robotology\WBToolbox\toolbox\library\WBToolbox.vcxproj]

@diegoferigo any hint?

diegoferigo · 2019-02-05T10:01:26Z

Actually I don't know why it's trying to compile it.

The blocks provided by WBT are not anymore Matlab-dependent. They are pure C++ classes. For this reason the toolbox is installed also if Matlab is not present.

@diegoferigo any hint?

Currently the WBT version installed from the superbuild is the former v4.1. We never had any windows user and the tests were not running in Windows at that time. If you don't need it, for the time being I would recommend to pass -DWBToolbox_TAG=master to the superbuild configuration.

S-Dafarra · 2019-02-05T10:54:57Z

Currently the WBT version installed from the superbuild is the former v4.1. We never had any windows user and the tests were not running in Windows at that time. If you don't need it, for the time being I would recommend to pass -DWBToolbox_TAG=master to the superbuild configuration.

Actually WBToolbox is already in master. See

robotology-superbuild/cmake/BuildWBToolbox.cmake

Line 17 in 1d1c0bb

TAG master

diegoferigo · 2019-02-05T11:00:29Z

robotology-superbuild/cmake/ProjectsTagsMaster.cmake

Line 7 in 54b978c

set_tag(WBToolbox_TAG 368119e32955e8eed5626ddc871bba76f4cd28c1)

;)

S-Dafarra · 2019-02-05T11:02:13Z

Oh dammit, that was hidden 😂

What about removing it?

diegoferigo · 2019-02-05T11:06:26Z

If you mean uninstalling the toolbox, yes, you can proceed. But in this case the next time you make the superbuild it will be installed again. Instead, editing the tag is a persistent fix, as long as you do not remove the cmake cache.

S-Dafarra · 2019-02-05T11:06:38Z

Actually that tag is in master, thus we downloading already the actual master of WBToolbox. I did not understand whether is necessary or not. If so, I have to rebase.

If you mean uninstalling the toolbox, yes, you can proceed.

I was referring to the tag.

traversaro · 2019-02-05T12:48:25Z

By the way isn't the problem solved here? @traversaro

That FindEigen3.cmake file is not supposed to be installed, the core problem is robotology/osqp-eigen#19 .

traversaro · 2019-02-05T12:49:59Z

The WBToolbox problem on Windows is indeed true, but independent from merging this PR actually. Feel free to disable it and verify that the rest of the superbuild works, thanks!

GiulioRomualdi · 2019-02-05T13:57:40Z

Commenting out

robotology-superbuild/CMakeLists.txt

Lines 130 to 131 in 54b978c

    
           find_or_build_package(BlockFactory) 
        
           find_or_build_package(WBToolbox)

and using robotology/walking-controllers#31 and robotology/osqp-eigen#19 I'm finally able to compile the superbuild on Windows10 Visual Studio 2017

GiulioRomualdi · 2019-02-05T14:50:52Z

@diegoferigo

robotology-superbuild/cmake/ProjectsTagsMaster.cmake

Line 7 in 54b978c

set_tag(WBToolbox_TAG 368119e32955e8eed5626ddc871bba76f4cd28c1)

The issue is still there even if WBToolbox is in master.

traversaro · 2019-02-05T14:52:52Z

As the compilation failure is not related to this PR, and robotology/walking-controllers#31 and robotology/osqp-eigen#19 are going to be merged soon, I will merge the PR so Linux and macOS user can start to benefit from the amazing walking software.

S-Dafarra mentioned this pull request Jun 13, 2018

Add unicycle-footstep-planner, walking-controllers and their dependencies #55

Closed

traversaro reviewed Jun 13, 2018

View reviewed changes

cmake/BuildUnicyclePlanner.cmake Outdated Show resolved Hide resolved

traversaro reviewed Jun 13, 2018

View reviewed changes

cmake/BuildUnicyclePlanner.cmake Show resolved Hide resolved

traversaro reviewed Jun 13, 2018

View reviewed changes

cmake/BuildicubWalkingDCM.cmake Outdated Show resolved Hide resolved

traversaro mentioned this pull request Jun 13, 2018

Release iDynTree 0.10 robotology/idyntree#453

Closed

GiulioRomualdi mentioned this pull request Jun 13, 2018

General folder refactoring robotology/walking-controllers#6

Closed

traversaro changed the title ~~[WIP] Added files to add osqp, OsqpEigen, UnicyclePlanner and dcmWalking~~ Added files to add osqp, OsqpEigen, UnicyclePlanner and walking-controllers Jul 11, 2018

traversaro approved these changes Jul 11, 2018

View reviewed changes

S-Dafarra mentioned this pull request Jul 13, 2018

Compilation fails with clang due to use of "auto" in method declaration robotology/walking-controllers#11

Closed

traversaro changed the title ~~Added files to add osqp, OsqpEigen, UnicyclePlanner and walking-controllers~~ [WIP] Added files to add osqp, OsqpEigen, UnicyclePlanner and walking-controllers Jul 13, 2018

bstellato mentioned this pull request Aug 24, 2018

CMake option to use external qdldl osqp/osqp#85

Open

traversaro mentioned this pull request Oct 21, 2018

[travis] Compile Dynamics profile based on env option #129

Merged

traversaro requested changes Dec 7, 2018

View reviewed changes

traversaro mentioned this pull request Jan 25, 2019

Enable YARP_USE_SDL and ENABLE_yarpmod_SDLJoypad YARP options by default? #167

Closed

traversaro approved these changes Feb 5, 2019

View reviewed changes

traversaro merged commit fe2f176 into robotology:master Feb 5, 2019

traversaro mentioned this pull request Sep 1, 2019

Use latest tagged version of upstream osqp and osqp-eigen instead of local fork #242

Closed

Added files to add osqp, OsqpEigen, UnicyclePlanner and walking-controllers #87

Added files to add osqp, OsqpEigen, UnicyclePlanner and walking-controllers #87

Conversation

S-Dafarra commented Jun 13, 2018 • edited Loading

S-Dafarra commented Jun 13, 2018 • edited Loading

traversaro Jun 13, 2018

Choose a reason for hiding this comment

S-Dafarra Jun 13, 2018

Choose a reason for hiding this comment

traversaro Jun 17, 2018

Choose a reason for hiding this comment

traversaro left a comment

Choose a reason for hiding this comment

traversaro commented Jun 20, 2018

traversaro commented Jun 20, 2018

traversaro commented Jul 11, 2018

traversaro commented Jul 11, 2018

S-Dafarra commented Jul 13, 2018 • edited Loading

traversaro commented Jul 13, 2018

S-Dafarra commented Jul 13, 2018 • edited Loading

traversaro commented Jul 13, 2018

traversaro commented Aug 23, 2018

S-Dafarra commented Aug 23, 2018

traversaro commented Aug 23, 2018

traversaro left a comment

Choose a reason for hiding this comment

traversaro commented Dec 7, 2018

S-Dafarra commented Dec 19, 2018

traversaro commented Feb 1, 2019

traversaro commented Feb 4, 2019

GiulioRomualdi commented Feb 4, 2019

traversaro commented Feb 4, 2019

lrapetti commented Feb 4, 2019

lrapetti commented Feb 4, 2019

S-Dafarra commented Feb 4, 2019

lrapetti commented Feb 4, 2019

traversaro commented Feb 4, 2019

lrapetti commented Feb 5, 2019

traversaro commented Feb 5, 2019

GiulioRomualdi commented Feb 5, 2019

GiulioRomualdi commented Feb 5, 2019 • edited Loading

GiulioRomualdi commented Feb 5, 2019

GiulioRomualdi commented Feb 5, 2019

GiulioRomualdi commented Feb 5, 2019

diegoferigo commented Feb 5, 2019

S-Dafarra commented Feb 5, 2019

diegoferigo commented Feb 5, 2019

S-Dafarra commented Feb 5, 2019 • edited Loading

diegoferigo commented Feb 5, 2019

S-Dafarra commented Feb 5, 2019 • edited Loading

traversaro commented Feb 5, 2019

traversaro commented Feb 5, 2019

GiulioRomualdi commented Feb 5, 2019

GiulioRomualdi commented Feb 5, 2019

traversaro commented Feb 5, 2019

S-Dafarra commented Jun 13, 2018 •

edited

Loading

S-Dafarra commented Jun 13, 2018 •

edited

Loading

S-Dafarra commented Jul 13, 2018 •

edited

Loading

S-Dafarra commented Jul 13, 2018 •

edited

Loading

GiulioRomualdi commented Feb 5, 2019 •

edited

Loading

S-Dafarra commented Feb 5, 2019 •

edited

Loading

S-Dafarra commented Feb 5, 2019 •

edited

Loading