2
$\begingroup$

Suppose you have immersed plane curve $\gamma:\mathbb{R}_+\rightarrow\mathbb{R}^2$ twice continuously differentiable and parametrized by arc-length such that its curvature $\kappa(t)=\|\gamma''(t)\|$ goes to $0$ as $t\rightarrow+\infty$. I was wondering whether it is possible for such curve to be bounded, meaning that it is contained in some ball in $\mathbb{R}^2$.

$\endgroup$
  • $\begingroup$ What is meant by $t$? $\endgroup$ – Oscar Lanzi Sep 22 '17 at 12:51
2
$\begingroup$

Yes, I think your intuition is correct.

HINT:

Recall that if $\gamma \colon I \to \mathbb{R}^2$ is parametrized by arc length ( that is $\|\gamma(t)'\| = 1$ for all $t$) then the curvature satisfies $$|\kappa(t)| = \|\gamma''(t)\|$$

Now the intuition is this: if $|\kappa|$ is small then $\gamma'$ varies little, so the direction of $\gamma$ changes little. This, coupled with the fact that the velocity is constant, implies that $\gamma$ approaches a line. Therefore, in some finite time it should get out of a given disk. Let's formalize that:

We have for any $a$, $b \in I$ $$\gamma(b) = \gamma(a) + \int_a^b \gamma'(t) dt $$

Assume that on the interval $[a,b]$ we have $|\kappa(t)|\le \epsilon$. Then we get $$\|\gamma'(t) - \gamma'(a)\| \le \epsilon(t-a) $$ Therefore $$\|(\gamma(b) - \gamma(a)) - (b-a)\gamma'(a) \|\le \epsilon \frac{(b-a)^2}{2}$$ We conclude: $$\|(\gamma(b) - \gamma(a))\| \ge (b-a)\| ( 1 - \frac{(b-a)\epsilon}{2})$$ Now assume that the interval $[a,b]$ has length $\frac{1}{\epsilon}$. Then we get $$\| \gamma(b) - \gamma(a) \| \ge \frac{1}{\epsilon} \cdot ( 1 - \frac{1}{2}) = \frac{1}{2 \epsilon}$$.

Hence $\gamma(I)$ cannot be contained in a disk of radius $< \frac{1}{4 \epsilon}$

Notice the proof works in any dimension $\ge 2$.

$\endgroup$
  • $\begingroup$ What we showed: If on an interval of length $R$ the curvature is $\le \frac{1}{R}$ then the distance between the end points is $\ge \frac{1}{2} R$. We suspect this is not the best estimate , probably the constant should be $2 \sin \frac{1}{2} > 0.95$ instead of $\frac{1}{2}$. $\endgroup$ – Orest Bucicovschi Sep 22 '17 at 17:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.