Can a catmull rom take the place of a dubins curve?

Started by
1 comment, last by Shannon Barber 7 years, 7 months ago

Hello,

I have been spending weeks on rewriting the code of this

https://github.com/unr-arl/DubinsAirplane

Today, I suddenly came up with this,

If it's just too much of an effort, why not just use a catmull rom to replace the dubins curves instead.

BTW, I am doing pathfinding with turns here.

It's just been a huge undertaking, and I really don't want to do this, because the D3DX library is already quite robust.

And I am just interested in the length of it and whether it hits something along the path.

I also look at that code, it seems to be a very very expensive calculation....

A catmull rom spline is fast and clean....

So what do you think about it?

Thanks

Jack

Advertisement
no then
consider a lsl turn
i shall need 6 control points for the third one i need to know the turning radius which is impossible to find with catmull rom

In context, Catmull-Rom has stability issues. It could easily produce a very unnatural curve.

Whereas the a Dubins curve is better behaved because of the constraints it follows (you guide it with the tangents-constraints).

- The trade-off between price and quality does not exist in Japan. Rather, the idea that high quality brings on cost reduction is widely accepted.-- Tajima & Matsubara

This topic is closed to new replies.

Advertisement