Best practice for tf on an arm with fixed base
Hi
I'm running a program on two very similar robot arms, with a fixed base. One of the issues is the fact that one robot is mounted at an angle compared to the other one, so obviously I need to transform my task to a different frame. I've set it up using tf and tf_broadcaster, as this is the only way I know how, but the solution is pretty inelegant (and still somewhat buggy). Thing is that I only really need to perform transformations when I program my plan (or once for every waypoint), while all the code I can find assumes that I would like to continuously publish the transforms, which seems a little overkill, and constantly running a node that is only really needed for the first 0.5s of the application.
I've tried looking at tf::Transformer to make my own "service" solution, but I haven't made that work yet, and I want to know if it's a bad idea, before I continue. Is it really necessary to use a broadcaster, when all I need is a transformation from a fixed point?
