chvmp / champ

MIT Cheetah I Implementation
BSD 3-Clause "New" or "Revised" License
1.47k stars 342 forks source link

Gazebo walking ROS2 #86

Closed AlexKaravaev closed 2 years ago

AlexKaravaev commented 2 years ago

Just a continuation of #85. Managed to get it working in Gazebo. Just as a side note, the main source of bugs was that gazebo was publishing clock with rate 10 and I have increased it to 50.

Also I have an issue of really low rtf (~0.5) in a world with objects. And it's happening only when the contact_sensor is launched. Debugging further I have observed, that even not launching the contact_sensor, but just subscribing in command line to gazebo topic contacts results in the same behavior. So it's probably something with the gazebo itself.

Summing up, this is how the future roadmap for ros2 looks like. SLAM and nav2 should not take too long.

https://user-images.githubusercontent.com/30314738/172949100-4844e2da-fd4d-4e88-b9fe-e8ac837932d8.mp4

grassjelly commented 2 years ago

hey @AlexKaravaev . Thanks for this working PR. Before I merge this, could you send a PR to the libchamp ROS2 banch as well? Thanks

AlexKaravaev commented 2 years ago

@grassjelly thank you! Done https://github.com/chvmp/libchamp/pull/5

sskorol commented 1 year ago

@AlexKaravaev thanks for the hard work! As README hasn't been updated yet, can you confirm if slam, navigation and velocity smoother are still under development? I want to test your changes with a real robot, so at least slam and velocity smoother might be very helpful. Do you have any estimate on these parts?

elpimous commented 1 year ago

@sskorol Working on Champ too on a real robot (with BLDC). Velocity smoother really interest me too

AlexKaravaev commented 1 year ago

@sskorol @elpimous Hi everyone! So I was dealing with the other things recently and haven't had any time to work on the ROS2 branch, but I am happy that more people right now eager to contribute and be involved!

So briefly the state of the ROS2 branch is as follows: Champ can walk in gazebo, meaning that all things related to controlling of joints and state estimation is working well in the simulation

As for the packages you mentioned:

I think I will not be available to jump right on these issues straight away, but probably at least would be happy to review PR related to these changes. Ah and also I will create PR with updated README for ROS2 branch shortly, just so everyone who wants to contribute would at least see where it's best to do.