Hello,
I have a large and mostly flat world (static) in which there are buildings with detailed, multi-layered interiors (instantiated at runtime), and I’m wondering what is the best way to achieve good-looking and fast pathfinding.
What I’m exactly doing now is that I have handmade meshes for those “interiors”, then I merge those meshes with the big one covering outside, then I create a recast graph from that big mesh only. (I’m not using navmesh graph beacuse I need navmeshCuts; my implementation needs to have deterministic movement, which I assume RVO lacks)
The problem is that, to have meaningful movement in buildings, the recast graph has to be a little detailed, but this detail level is absolutely redundant for outside.
It’s my understanding that having a single graph that’s covering the whole world is usually better way to go, but that would require a considerable deal of memory in my case. I’m thinking of multiple graphs, where I will have a less detailed recast for outside, and more detailed ones for interiors. When finding a path from “outside” to “inside”, I will query for two paths: one to source to “entrance” of the building, and one from entrance to target, then merge the paths.
Is this how it is supposed to happen? Am I in a correct mindset?
Thanks.