On Tue, May 26, 2015 at 6:35 AM Serge Stinckwich serge.stinckwich@gmail.com wrote:
On Mon, May 25, 2015 at 4:48 PM, Miguel Campusano mcampusa@dcc.uchile.cl wrote:
Rviz is a good visualization, but it focuses on a representing the data
in a
3d environment. Rviz, in my experience, represents the interactions of
the
robot (with its sensors) in the environment, you can draw maps and do a
lot
of things.
rviz is not only for 3D but also for 2D viz. Look at this video for example:
http://wiki.ros.org/navigation/Tutorials/Using%20rviz%20with%20the%20Navigat...
Right, I have seen that a long time ago but I forgot it :( . But still it holds the idea that rviz focuses on representing the data in the environment (drawing the map for example) rather than the data itself (in a point of time, what was the distance between the robot and the nearest wall)
However, rviz does not provide access to the data itself. Long history short, while I can see there is an object with the laser scan at a
certain
distance, I can not see the actual distance of the object. Sometimes we
want
more info about the data to see what values are best suited to use in our programs.
Yes, you are right rviz does not provide access to the data and is quite difficult to customize to specific needs. You have to write plugins in order to add more visualisations.
Yes, you can represent the data in the visualization, but still with rviz I can't see how you can have direct access to the data from the visualization. At least, I can't see how you can do something like we did we the laser scan in rviz -showing information about the area of detection and other things we have in mind ;)-
Having a robotic data viz engine that could be extended with ROASSAL
scripting could be nice. This is what you are trying to do ?
Our main focus is, for every different sensor, understand the data separately and use that data in algorithms and programs in a live programming context. For example, for us, having the distance of an object with the laser scan has a great value because we can determine the best suited distance to take different actions. After the actions are implemented, you can see with the visualization if the actions are being done at the right moment. But I can not denied that having a robotic data viz engine extensible with roassal scripting would be cool.