I have a curve that looks like this (it's cyclical):
I can get a partial fit by fitting a 3rd degree polynomial, but I have a feeling there must be a better fit (something that involves sin & cos).
The fitted $a*t^3 + b*t$ curve looks like this:
Any ideas? thanks.
You could try to fit higher-order polynomials, which should give you the flexibility needed to ensure differentiability or some such constraint. Or if you're willing to try parametric curves, you can get a fit relatively easily, like $$x(t)=\sin(t)+t, y(t)=\sin(t)$$
I assume that you explicitly want the skewed nature of the wave in your picture. Tuning the amount of skew, height, etc. in the parametric formula is just a matter of tweaking a few parameters.