Add locations (with labels) on the map to allow quick navigation. Labels need to be stored in RTAB-Map.

Change the opentera_webrtc_demos::robot_status.py node to be generic and extensible and work on all robots, then change the robot_status.py in T-Top and Beam repos to use the new facilities.

Possible implementations: subclasses, dependency injection

Right now, when the map grows out of bounds of the generated map image, the map stops being able to be drawn. Ideally, the map image generator should be able to crop the parts of the map hat are outside of the image.
Changes will probably have to be made in OccupancyGridImageDrawer.cpp

To make the map more intuitive when navigating, the center of the map should coincide with the position of the robot. The map should also rotate with the robot, similar to how a car GPS displays the map.

We should have clean launch files with a updated README and examples.

Hi,

I am trying to build this package in my catkin workspace and I'm getting the following error:

CMake Error at opentera-webrtc-ros/CMakeLists.txt:9 (add_compile_definitions): Unknown CMake command "add_compile_definitions".

I am using cmake version 3.10.2

The user should be able to mute the robot directly on the robot's screen. The teleoperator can also control mute at a distance.

This would enable to understand where the sounds are coming around the robot in relation to the global map.

See http://wiki.ros.org/image_transport

Hi,

I want to play with webrtc demo first. But encountered lots of issues during the setup process. Here's a list of my findings.
catkin_make has failed for me due to a number of missing deps: fftw3-dev / libconfig-dev and gfortran of a specific version.

Had to spent some time on the latter, as the trickiest part here was not the missing gfortran (it was installed in my system), but rather an inconsistency in gcc / g++ deps. These 3 must match. But I had gcc/g++ 8.4.0 and gfortran 9.4.0 installed. And the main problem here is that for my Ubuntu 20.04 distribution there was neither gcc/g++-9.4.0 nor gfortran-8.4.0 available. So I have to install the 10th version of these 3 and run update-alternatives to continue the build process.

Then it failed on the python dependencies installation. It was a weird one as it didn't have permission to install packages linked to python3.7. However, I have a 3.8 version installed and available via python3 -V and update-alternatives. I had python3.7 at some point in the past but couldn't find any direct system reference except the availability of a python3.7 folder in the /usr/local/lib. So I have to remove it first. And then the build failed again due to the missing python interpreter error. The root cause was here: ./opentera-webrtc-ros/opentera_webrtc_ros/opentera-webrtc/CMakeLists.txt. I'm not sure why, but if we try to find python 3 package, it can't see an installed 3.8 version. When I replaced 3 with 3.8 the build finally passed. So it'd probably make sense to make python version configurable via parameter or env variable.

Ok, so setup is done. But when I tried to run the following demo:

roslaunch opentera_webrtc_demos demo.launch

it failed with a bunch of errors related to missing client_config.json:

Traceback (most recent call last):
  File "/home/user/ros1_workspaces/teleop_ws/src/opentera-webrtc-ros/opentera_client_ros/scripts/opentera_client_ros.py", line 256, in <module>
    with config_file_name.open() as json_file:
  File "/usr/lib/python3.8/pathlib.py", line 1222, in open
    return io.open(self, mode, buffering, encoding, errors, newline,
  File "/usr/lib/python3.8/pathlib.py", line 1078, in _opener
    return self._accessor.open(self, flags, mode)
FileNotFoundError: [Errno 2] No such file or directory: '/home/user/.ros/opentera/client_config.json'

I don't see any notes about this config in the README. So wondering where I can find it or what should be the content?

Which fails, because it isn't a Python module .so, just a standard .so.
It should be excluded from the autosummary generation.

One way to exclude it is to add an underscore prefix to it, but then we might need to make sure that it is still correctly loaded dynamically.

Custom JSON and WebSockets could benefit from being replaced by a standardized rosbridge + roslibjs to facilitate the communication with ROS for any kind of message and any topic.
We need to evaluate if that helps the frontend code be more easy to understand and flexible.

On T-Top and Sammy, it happens semi-regularly. When it happens, it lasts for some time, and doesn't seem to get fixed by restarting the robot nor the router.
It happened with the simulated beam too, but less often.

Sometimes, there is no problem at all and everything just magically works fine.

Manage the local camera's default size based on the deviceProperties.json file and the incoming video's resolution, so that it adjust to the aspect ratio and so that the user has control on the default size.

VideoSink should be replaced by callback :

void StreamClient::setOnVideoFrameReceived(
const std::function<void(const Client&, const cv::Mat& bgrImg, uint64_t timestampUs)>& callback)

https://github.com/introlab/opentera-webrtc-teleop/blob/93afd6187b9d54aa5306aa78789de5e7192cac82/opentera-webrtc-native-client/OpenteraWebrtcNativeClient/include/OpenteraWebrtcNativeClient/StreamClient.h#L120

Messages are "bridged" with protobuf JSON messages, from opentera-messages package.

Should modify RosDataChannelBridge.cpp to take change into account.