kth-sml / motion_capture_system Goto Github PK

Error messages that does not print any new line appears in several places. These should have end with a newline or even better, be replaced with ROS logging macros.

The subject lost and subject initialized messages are throttled to avoid spam. However, when the system is turned on only one visible subject is currently announced as initialized.

ROS only supports names consisting of A-Z, a-z, 0-9, _ and /. QTM will however accept names containing other characters, which causes the qualisys_mocap ROS node to crash.
Should the names be sanitized with a warning instead?

several new settings were added when the QTM API was updated. The readme should be updated to describe the new settings, and any other changes.

It should be possible to use the automatic server detection to find the QTM server when no IP is specified.

The comparison node is currently barely documented and several parameters are not in the example launch file.

The CPU usage of the node keeps increasing. After a few minutes the processing will be too slow to keep up with the data stream, causing increasing delays.

There are several pragmas in the Vicon SDK that generates error during the compilation. These pragma macros appears to be intended to suppress compilation warnings.
Example:

/src/svea_research/src/motion_capture_system/mocap_vicon/src/vicon_sdk/Vicon/CrossMarket/StreamCommon/Buffer.h:41:0: warning: ignoring #pragma warning  [-Wunknown-pragmas]
 #pragma warning( disable : 4355 )

Hello. Appologies if this is the wrong place to post this. I am trying to build the mocap_qualisys package.

I have setup ROS as described here: http://wiki.ros.org/Installation/Windows
I created a catkin workspace as described here: http://wiki.ros.org/catkin/Tutorials/create_a_workspace
I have cloned the git repository here into my workspace and added the workspace path to ROS_PACKAGE_PATH.
When I run catkin_make --pkg mocap_qualisys --cmake-args -DCMAKE_BUILD_TYPE=Release, I get the following output:

Base path: C:\opt\motion_capture_system
Source space: C:\opt\motion_capture_system\src
Build space: C:\opt\motion_capture_system\build
Devel space: C:\opt\motion_capture_system\devel
Install space: C:\opt\motion_capture_system\install
Packages "mocap_qualisys" not found in the workspace

It seems like the ROS_PACKAGE_PATH is not being recognized.

I have tried moving the mocap_qualisys and mocap_base folders into my workspace's src folder. In this case, I get the following output when running catkin_make --pkg mocap_qualisys --cmake-args -DCMAKE_BUILD_TYPE=Release

Base path: C:\opt\motion_capture_system
Source space: C:\opt\motion_capture_system\src
Build space: C:\opt\motion_capture_system\build
Devel space: C:\opt\motion_capture_system\devel
Install space: C:\opt\motion_capture_system\install
####
#### Running command: "cmake C:\opt\motion_capture_system\src -DCMAKE_BUILD_TYPE=Release -DCATKIN_DEVEL_PREFIX=C:\opt\motion_capture_system\devel -DCMAKE_INSTALL_PREFIX=C:\opt\motion_capture_system\install -G Ninja" in "C:\opt\motion_capture_system\build"
####
-- The C compiler identification is MSVC 19.29.30137.0
-- The CXX compiler identification is MSVC 19.29.30137.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/Enterprise/VC/Tools/MSVC/14.29.30133/bin/Hostx64/x64/cl.exe - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/Enterprise/VC/Tools/MSVC/14.29.30133/bin/Hostx64/x64/cl.exe - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Using CATKIN_DEVEL_PREFIX: C:/opt/motion_capture_system/devel
-- Using CMAKE_PREFIX_PATH: C:\opt\ros\noetic\x64\tools\vcpkg\installed\x64-windows;C:\opt\ros\noetic\x64
-- This workspace overlays: C:/opt/ros/noetic/x64
-- Found PythonInterp: C:/opt/ros/noetic/x64/python.exe (found suitable version "3.8.3", minimum required is "3")
-- Using PYTHON_EXECUTABLE: C:/opt/ros/noetic/x64/python.exe
-- Using default Python package layout
-- Found PY_em: C:\opt\ros\noetic\x64\lib\site-packages\em.py
-- Using empy: C:/opt/ros/noetic/x64/lib/site-packages/em.py
-- Using CATKIN_ENABLE_TESTING: ON
-- Call enable_testing()
-- Using CATKIN_TEST_RESULTS_DIR: C:/opt/motion_capture_system/build/test_results
-- Found gtest: gtests will be built
-- Using Python nosetests: C:/opt/ros/noetic/x64/Scripts/nosetests.exe
-- catkin 0.8.9
-- BUILD_SHARED_LIBS is on
-- BUILD_SHARED_LIBS is on
-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-- ~~ traversing 2 packages in topological order:
-- ~~ - mocap_base
-- ~~ - mocap_qualisys
-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-- +++ processing catkin package: 'mocap_base'
-- ==> add_subdirectory(mocap_base)
-- Found Boost: C:/opt/ros/noetic/x64/include (found version "1.73.0")
-- +++ processing catkin package: 'mocap_qualisys'
-- ==> add_subdirectory(mocap_qualisys)
-- Using these message generators: gencpp;geneus;genlisp;gennodejs;genpy
-- Configuring done
-- Generating done
-- Build files have been written to: C:/opt/motion_capture_system/build
####
#### Running command: "ninja -j20 -l20" in "C:\opt\motion_capture_system\build\mocap_qualisys"
####
ninja: error: loading 'build.ninja': The system cannot find the file specified.Invoking "ninja -j20 -l20" failed.

In this case, it seems as though the ninja build files for the packages are not being generated. Any help would be appreciated. Thank you.

I am using the mocap qualisys node to stream data from qualisys computer to raspi (ubuntu 20.04) connected to pixhawk via uart. The connections are verified .The data is remapped to /mavros/vision_pose/pose and same data is shown on /mavros/local_position/pose. but in the mavlink inspector the LOCAL_POSITION_NED , the x and y values are flipped. The altitude hold mode is working but the position hold is not working.
My doubt is whether the qualisys node is converting the data in to ENU frame or is it streaming the same frame as the corrdinate system in the mocap software and if somebody have solved the issue ,please help.

Current version of this method can only request 6DOF body streams. The previous version was straight up broken. Both could theoretically cause a segfault due to use of snprintf on a small buffer (that is still large enough that stack overflows could be a worry).
The method should be rewritten with C++ strings and able to send any valid request.

When the data stream is requested a boolean is returned to indicate if the stream is successfully started or not.
The value of the bool is never checked, so if the request to start a stream fails, the node will still go ahead and report that the initialization was successful.
I'm not aware what could cause this situation to occur or if it is even possible.
Several (properly checked) interactions with the server have to succeed before the stream is requested, but the check should still be there.

Suggestion: Divide the udp_port argument into two arguments. The transport protocol would be set by transport_protocol,
a string that should be either UDP or TCP, and the port should be set by stream_port, an integer.

Motivation: Currently, the udp_port parameter decides both if TCP or UDP is used, and which port that is used. The port is however only selectable for UDP. If TCP is used, the port is always random. Specifying which TCP port that should be used could be useful.
The current parameter name is also misleading.

The qualysis mocap node delivers an odometry message and the twist part of the message (velocities) does not follow the convention.

Odometry message definition (https://docs.ros.org/melodic/api/nav_msgs/html/msg/Odometry.html) states:

The twist in this message should be specified in the coordinate frame given by the child_frame_id

The "child_frame_id" is correctly set to <tracked_object/base_frame> but the velocities are expressed in the mocap world frame and not the object local frame.

Hi,

I would like to get (labelled) marker position from QTM to ROS, in order to be able to use AIM.
Your implementation of the protocol seems quite complete, so I will probably try to write PR to do what I want.

Any suggestions/recommendations before I start ?
What do you think of that ?

Best,
Etienne.