No robot hardware, How to study ROS?
I have not robot hardware, how to study ROS? how to debug appliction?
ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
asked 2012-07-12 16:12:59 -0600
This post is a wiki. Anyone with karma >75 is welcome to improve it.
I have not robot hardware, how to study ROS? how to debug appliction?
answered 2012-07-12 17:55:28 -0600
This post is a wiki. Anyone with karma >75 is welcome to improve it.
There are various robot simulators, based on stage or gazebo.
EDIT: Most of the officially listed robots have some kind of simulation procedure. The quality and detail no doubt vary depending on the needs of those projects.
You should also consider using Morse (http://www.openrobots.org/wiki/morse/).
First of all, you should complete the tutorials. They are excellent and don't require any hardware. My experience is that people who do the tutorials first save a lot of time later.
As a next step, I agree that a simulator is the best way to go. Personally, I'd recommend the turtlebot Gazebo simulation over the PR2 one, since the Turtlebot things tend to be simpler and easier to grasp.
As a third step, there are several cheap pieces of hardware that work with ROS and that you might already have access to:
Do you have an Android phone? You can visualize your phone's accelerometer/magnetometer/gyroscope data in RViz using android_sensors_driver, or you can teleop your (simulated) robot around using android teleop. Both projects don't quite work out of the box (you'll need some additional parts except from the two I linked to), so you'll have a little project to learn ROS on. Don't hesitate to open up another question if you need help.
Do you have a Microsoft Kinect? You can do some of the Point Cloud Library (PCL) tutorials, and then think up an own little project. The PCL is now a separate project from ROS, but basically all the point cloud perception in ROS is done using the PCL.
answered 2012-07-15 18:58:16 -0600
This post is a wiki. Anyone with karma >75 is welcome to improve it.
At ROS' core, there is a transport framework that allows the different ROS packages that have been developed to communicate. You could easily install ROS on a number of networked computers and watch different nodes communicate. Adding, say, a cheap kinect sensor to one of the computers (with the associated ROS packages and openCV) and developing some control software on another and you have yourself pseudo-robot on one computer and an operator control unit on another (just an example).
There would be a large number of research opportunities with something like that.
Asked: 2012-07-12 16:12:59 -0600
Seen: 2,122 times
Last updated: Jul 15 '12