It depends on what you mean by "similar".
The TurtleBot 2's two goals are:
- Be ready to run ROS out of the box without hours of building your own platform. You can plug it in and start doing sophisticated things with ROS like SLAM in minutes.
- Be relatively inexpensive.
If you spend a lot more money, then there are lots of options, but that isn't like Kobuki, that is expensive.
If you build your own platform and write your own ROS interface there are options too, but that isn't like Kobuki either.
The ROS site already has a wide variety of robots with active support:
http://wiki.ros.org/Robots
If you want to build your own "Turtle like" robot I suggest taking a look at the work of Patrick Goebel at http://www.pirobot.org/
as well as Dr. Rainer Hessmer at http://www.hessmer.org/blog/2010/11/2...
Personally I am attempting to adapt ROS to work with an "Arlo" platform from Parallax ( http://www.parallax.com/product/arlo-... ) which uses a controller that is somewhat different from most. I have found that by adapting Dr. Hessmer's code I was able to feed odometry data from the controller to ROS and accept twist commands from ROS on the controller fairly easily.
Again, though, it is a matter of what you mean by "similar". Mine will be round like the Turtlebot, but it won't be a under $2000 ready to run robot. I only bring up the Arlo because with the aluminum wheel option it has a 60 pound payload, which relates to your question about payload.