Hi all, I came across polynomial spirals as generalizations of the Cornu spiral: http://www.2dcurves.com/spiral/spiralps.html At the bottom of this page are 4 examples where the curvature is a polynomial function of the arc length. How can one generate a pair of parameterized equations for x and y, given the fact that the curvature is a polynomial function of the arc length? Thanks, Kerry -- lkmitch@gmail.com www.kerrymitchellart.com http://spacefilling.blogspot.com/
I haven't tested this, but it looks straightforward enough ... Using standard schoolroom stuff, curvature = dt / ds , where t = tangent angle, s = arc length; and tan t = sin t / cos t = dy / dx . Since your curvature is a polynomial function f(s) of arc length, integrating t = g(s) , where dg/ds = f ; integrating again (and quietly losing an irrelevant sign), x = sin g(s), y = cos g(s) . WFL On 1/11/10, Kerry Mitchell <lkmitch@gmail.com> wrote:
Hi all,
I came across polynomial spirals as generalizations of the Cornu spiral:
http://www.2dcurves.com/spiral/spiralps.html
At the bottom of this page are 4 examples where the curvature is a polynomial function of the arc length. How can one generate a pair of parameterized equations for x and y, given the fact that the curvature is a polynomial function of the arc length?
Thanks, Kerry
-- lkmitch@gmail.com www.kerrymitchellart.com http://spacefilling.blogspot.com/ _______________________________________________ math-fun mailing list math-fun@mailman.xmission.com http://mailman.xmission.com/cgi-bin/mailman/listinfo/math-fun
On 1/12/10, Fred lunnon <fred.lunnon@gmail.com> wrote:
I haven't tested this, but it looks straightforward enough ...
Using standard schoolroom stuff,
And heading rapidly towards the bottom of the class ...
curvature = dt / ds , where t = tangent angle, s = arc length; and tan t = sin t / cos t = dy / dx .
Missed out a step there: the point is, tan t = dy/dx, and ds^2 = dx^2 + dy^2 ; so dx/ds = cos t, dy/ds = sin t .
Since your curvature is a polynomial function f(s) of arc length, integrating t = g(s) , where dg/ds = f ;
where g(s) is also a polynomial, OK so far ...
integrating again (and quietly losing an irrelevant sign), x = sin g(s), y = cos g(s) .
Nooo! In fact, all we can say is x = \int cos g(s) ds, y = \int sin g(s) ds, which in general won't be integrable in terms of elementary functions. At this point, I'd suggest reaching for a numerical integrator. But if you want to press on, there's things called Fresnel and Lommel functions ... WFL
On Tue, Jan 12, 2010 at 8:56 AM, Fred lunnon <fred.lunnon@gmail.com> wrote:
Nooo! In fact, all we can say is x = \int cos g(s) ds, y = \int sin g(s) ds, which in general won't be integrable in terms of elementary functions.
At this point, I'd suggest reaching for a numerical integrator. But if you want to press on, there's things called Fresnel and Lommel functions ... WFL
Thank you very much for your explanation. Integrating numerically is what I was looking to do--I have found some interesting artifacts integrating the Fresnel functions numerically and wanted to try with other curves. Thanks, Kerry -- lkmitch@gmail.com www.kerrymitchellart.com http://spacefilling.blogspot.com/
participants (2)
-
Fred lunnon -
Kerry Mitchell