Sounds solid. I’d start with a basic energy balance—potential energy into kinetic, minus friction and air drag—then layer in the turning radius as a constraint on the curvature of the path. The trick is keeping the differential equation stable over the whole run, so the model doesn’t blow up when the slope gets steeper. You’ve got the right idea, just make sure you capture the small-angle turns; they add up faster than you think.

Got it, I’ll keep the curvature small and watch for that singularity. If it does start to blow up, let me know where it skids and we’ll trim it. Thanks for the heads‑up.