[ros-users] Moving from Player to ROS

[Jack O'Quin]

On Sat, Dec 11, 2010 at 11:40 AM, Amir H. Soltanzadeh <amirhst at gmail.com> wrote:

> Since 2007 we use Player in our laboratory to develop accurate real-time 2D
> SLAM to be used in our mobile robots in rough terrain. At the time being we
> have a fine odometry-free 2D mapping in very harsh environments and we want
> to extent our system to be capable of autonomous navigation in such areas.
> Since this will be a relatively long-term project and ROS has potentially
> much more advantages in this area, I'm wondering whether we should move from
> Player to ROS (at the cost of loosing time) or not!
>
> Any advice or comment will be highly appreciated!

It depends on your project, of course. For us, it was worthwhile
converting a fairly large project from Player to ROS. The conversion
was not difficult, and the ROS tools are much more productive.

I suggest doing a feasibility study. Pick a low-level driver than does
not depend on much else and convert it to ROS. Do some testing and get
comfortable with the tools. The first port is the hardest, because you
need to learn the ROS equivalents of Player interfaces that are
already familiar. As mentioned already, there are some guidelines
here:

  http://www.ros.org/wiki/ROS/PlayerIntegration

If you have questions, ask the mailing list. Good luck!
-- 
 joq