Automated testing of ROS Robots

asked 2016-02-02 09:30:52 -0600

Sebastian from Proboter gravatar image

Hi, we are on the way to setup Travis CI for a first step into continuous integrating our ROS based robot. How do the rest of this great community test their products? Are there any articles or books explaining strategies on how to test robots? The topic "automated testing" seems so obviously for the robotic stuff but we couldn´t find great information in our research.

edit retag flag offensive close merge delete

Comments

The PR2 had automated regression tests many years ago, but I don't think they run or pass any more, and they're very poorly documented: http://wiki.ros.org/pr2_simulator/Tut...

ahendrix gravatar image ahendrix  ( 2016-02-02 15:20:23 -0600 )edit
1

Hi Ahendrix, thank you very much for your response. We did get also some response on the ROS Google Usergroup which I´d like to link here: https://groups.google.com/forum/#!topic/robot-web-tools/9sMvU4y83Bk

Sebastian from Proboter gravatar image Sebastian from Proboter  ( 2016-02-03 03:31:41 -0600 )edit

Just a short comment in reply to your question in the ROS Google Usergroup: I could not find "articles or books explaining strategies on how to test robots" as well. It seems like testing related information is covered by kind of "fog of mistery" and everyone has to learn it on its own the hard way.

thinwybk gravatar image thinwybk  ( 2017-09-16 06:09:58 -0600 )edit

And another comment: "Running on the real robot will be tricky to do automated testing and I am not sure if there is value." It is tricky but possible and there is value for sure :) Fetch Robotics calls it "Physical Continuous Integration".

thinwybk gravatar image thinwybk  ( 2017-09-16 06:17:50 -0600 )edit

I proposed to open a Testing Working Group (discourse.ros.org) a while ago with the aim to distribute generic knowledge about how to test and to improve generic tooling support in ROS1. However there was little interest...

thinwybk gravatar image thinwybk  ( 2017-09-16 06:31:09 -0600 )edit