I hope you all had a nice summer break! Here in Sweden it has been pretty rainy unfortunately… but good news is the Aerial ROS meetings will be starting again
We will be talking about Message Standards for UAVs as voted with the last meeting. I’ve started a discussion thread about this here so we can look at the current state, so if you would share your thoughts on that, that would help greatly
Hi I am an engineering student from Turkey. I want to learn and improve my self in this area but I am just a beginner and dont have much knowlage or the experience. I started my journey with making an SITL simulation with using Px4 firmware JMAVsim and Qgroundcountrol. I wanna learn more while continue this journy and join your group. I will be attanding to meeting in August 16 to meet the group and learn.
Anyway, if you happen to have an questions about how to get started with PX4 I also recommend the PX4 community Q&A that is right after the ROS-aerial meeting: Dronecode Calendar — Dronecode Foundation
Im so happy for your welcome. I will look at the tutorials thank you for sharing. As a part od the group how can ı participate in the reserch as beginner. I want to take task and be a part of the group as much as possible.Is there any small tasks or reserches that you could give me too.
Our next meeting is planned for the 30th of August and we still have an opening for anyone present their work in aerial robotics! Send me DM before Tuesday 22th.
Yes, I agree with Tully; having the same messages for a mixed fleet of ground and aerial vehicles is highly desireable for mixed fleet work.
What is an example of a minimum set of messages?
The GSoC project I am doing does a few basic things that could be a start
Localization data from Cartographer - Using vision or LiDAR based localization on the ROS computer that’s not possible on the embedded system is very common for GPS denied operation
Position, velocity, and acceleration reporting from the autopilot to the ROS computer. We already have merged this in ArduPilot for both ground and aerial vehicles.
Velocity control from a navigation stack like NAV2 - yes, there are update rate problems doing low level control with out-of-the box Ubuntu systems like a Pi running Ubuntu 22.04, however there are plenty of platforms that can run ROS 2 in realtime. I’m currently working with Steve on this: Switch from Twist to TwistStamped for cmd_vel · Issue #1594 · ros-planning/navigation2 · GitHub
Agree with Jaeyoung - It would be great to define an alternative to trajectory control for how a high level planner controls the drone, but I don’t know if there’s any that are already adopted
Starting from scratch or re-using REP-147?
I’m happy to just revise REP-147. I think a lot of really great info and context with respect to MavLink is there, and since MavLink is still used, it’s relevant. That said, some of the information is incorrect. ArduPilot’s controls are done in the inertial frame, not body frame, so it makes sense to allow ROS users to do control in the inertial frame or body frame depending on what they set for the frame ID. If the autopilot supports a ROS interface for both, the computation to the desired control frame can happen in the autopilot at real-time, and it’s just a transform, so I don’t see there being concerns. The ROS computer can then just control in whatever frame their controller prefers to operate in.
So we don’t have an external speaker for now due to little time, so I’ll actually be giving an presentation about an education and tutorial current state overview! We actually had a discussion meeting about education and tutorials here some in May and after that I’ve done some research and gotten some great replies!
Also take a look at this discourse thread:
(ps @RFRIEDM thanks for the reply! don’t forget to also take a look at this thread too for further discussions about message standards)
So next time we will have a discussion meeting about simulation! The next meeting is planned for the 13th of September so I’ll start a new thread for that.
