Skip to content

NavMesh-based random sampling + shortest-path for arbitrary maps (data collection / training) #84

Description

@DiTenix

Hi SimWorld team,

Thanks for the release of this great work.

I’m currently trying to use SimWorld to collect training data. My goal is:

  • load an arbitrary UE map (including non-city / indoor / natural scenes),
  • randomly sample reachable start/goal points,
  • compute the shortest path on the navigation mesh,
  • then generate trajectories + labels for training.

Questions:

  1. Do you already provide any NavMesh-based APIs for:

    • sampling random reachable points, and
    • querying shortest paths (or path waypoints) between two points
      in an arbitrary map?
  2. If this is not currently supported, what would be the recommended way to extend SimWorld to add it?

    • Is it feasible to add a Blueprint actor (packaged in a .pak) and call it via existing UnrealCV+/vbp style commands to return path info (JSON or chunked results)?
    • Or does this require modifying/rebuilding the UE backend / UnrealCV+ plugin to expose dedicated commands for NavMesh queries?
    • Any guidance on where such functionality would best live in your architecture (UE side vs Python side) would be appreciated.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions