GuiRitter / OpenBase

An omnidirectional mobile platform with a 3 omnidirectional wheels layout
MIT License
142 stars 45 forks source link

Problems controlling the model #19

Open migranram opened 3 years ago

migranram commented 3 years ago

Hello I have been trying to use your mode for an university project, but unfortunately we haven't been able to control it. The robot doesn't seem to respond to any rostopic pub we make for the controllers. We realized that there was an error while running the .launch:

GazeboRosControlPlugin missing <legacyModeNS> while using DefaultRobotHWSim, defaults to true

But it Gazebo runs anyway. We have tried in Kinectic and Lunar.

Any idea of what could be going wrong?

Cumprimentos de Sevilha, Espanha.

GuiRitter commented 3 years ago

Hello and thanks for taking an interest in my project!

This is not an error that I remember coming across. First, I would like to know which *.launch file you're trying to launch. This is because not all *.launch files are meant to be ran. Some are only meant to be included in others.

Second, I searched a little and I came across this. In short, edit xacro/main.urdf.xacro and add <legacyModeNS>true</legacyModeNS> below <robotNamespace>/open_base</robotNamespace> (almost at the end of the file).

See if it works. I suspect it happened to you and not me because maybe you have more than one ROS distro installed and are sourcing setups from different distros or maybe you have a Gazebo distro installed manually and ROS is referencing that instead of its own Gazebo distro.

Try that file change and let me know if it works. Anything else, just ask. Best of luck!

migranram commented 3 years ago

Thanks for your answer!!

First of all the problem about the controllers is already solved. We isolated the urdf model and added the actuators just the way you did. It is the same.

But we noticed a couple more problems. The vehicule seems to move by itself, it is kind of drifting randomly. We tried to add some friction to the joints in case that was the reason and it seemed to solved it partially. Then we added the kinematics posted in your repo to move the robot with Vx and Vy, but for whatever reason, the X component is not working properly.

Vy works just fine, since it is only using 2 motors, but the Vx seems to drift too and we don't understand why. Our project is to control this robot and make some path finding and computer vision work with it and ROS, so we don't really know much about gazebo, we are just discovering everything now.

Any ideas on our latest problem? We cannot make a reliable controller if the velocities don't work properly and we don't know if it's a known problem of Gazebo or something.

Muito obrigado!!