This question is related to the issue Dynamical footprint in Nav2.

I am currently trying to get a wheel-loader to drive autonomously indoor. Unfortunately , it differs from the supported kinematic model in Nav2, DiffDrive and Ackermann. As of my knowledge and please correct me if am wrong, is that the current smac-planners (or generally all planners) contains information about x,y and theta (pose). In this case, the wheel-loader have a pivot angle in the middle, which based on its kinematics provide different heading angle for both front and back.

Based on the answer from the related issue, Mike Ferguson came with an solution to add the central pivot angle as one of the lattice state variables. In addition to updating the collision checking to use the pivot angle to define the footprint, and also have to update the motion primitives.


Does someone have experience in doing this procedure? Yes, this will probably be a lot to do, so any guidelines or information is appreciated. I've The idea is also to then be able to create a pipeline to add custom models and still be able use Nav2 in the future instead of creating custom navigation software.

  • $\begingroup$ Yes, this has definitely been done before and your best bet is to read some literature on it. I don't think you're going to run into someone that has been there done that with a simple answer; its not a simple question. $\endgroup$ Sep 26 at 4:59
  • $\begingroup$ With that said, if you got to a point that you were able generate a lattice set with your pivot angle in it, I would be more than happy to chat with you about how we can handle collision checking in a lattice node surrounding that, since I think that would be obviously valuable. I can't (or rather don't have the time to learn then teach) do the lattice generation step for you, but if you can make that jump, I can help with the Smac Planner bits to make it work without creation your own custom software. That part I can very much help with. $\endgroup$ Sep 26 at 5:01
  • $\begingroup$ I can't say that I vetted the math or anything in detail, but a quick skim the following approach more broadly seems reasonable: arxiv.org/pdf/1912.06264.pdf $\endgroup$ Sep 26 at 5:07


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Browse other questions tagged or ask your own question.