1

I need to prove that for a given a smooth trajectory $\vec{z}(t)$ in $\mathbb{R}^2$ defined on some interval $I$ and such that $\|\vec{z}(t)\|=1$, there exists a unique twice differentiable $\varphi(t)$ defined on $I$ such that $\varphi(0)=0$ and \begin{equation} \vec{z}=\begin{pmatrix} \cos\varphi \\ \sin\varphi \end{pmatrix}. \end{equation}

I suppose I'm free to assume that $\vec{z}(0)=( 1 ~~0)$, otherwise the problem doesn't make sense. I tried constructing the function $\varphi$ (adding or subtracting $2\pi$ whenever the trajectory crosses {$\{ (x,y)\in\mathbb{R}^2~:~x<0\}$} so that $\varphi$ is continuous and then proving it's twice differentiable) but the professor wasn't happy with that solution. Can someone point me in the right direction?

keplerr
  • 13
  • 3

1 Answers1

1

I assume that $z$ is twice differentiable else $\varphi$ can not exist. The quicker answer uses covering spaces theory. Indeed, $z$ is a $\mathcal{C}^2$ function from $\mathbb{R}$ to the circle $\mathbb{S}^1$ which admits for universal covering space, $$ p : \left\{\begin{array}{rcl} \mathbb{R} & \rightarrow & \mathbb{S}^1 \\ t & \mapsto & e^{it} \end{array}\right. $$ Covering spaces satisfy the lifting property that says that onc you fix $\varphi(0) = 0$, $z$ lifts into a unique continuous $\varphi : \mathbb{R} \rightarrow \mathbb{R}$ such that $z = p \circ \varphi : t \mapsto e^{i\varphi(t)}$, which is what you want.

To get the twice differentiability of $\varphi$, use the fact that $p$ is locally a smooth diffeomorphism. Indeed, for all $t \in \mathbb{R}$, there is a small neighborhood $U$ of $\varphi(t)$ such that $p_{|U} : U \rightarrow p(U)$ is a smooth diffeomorphism. Therefore, in $U$, $\varphi = p_{|U}^{-1} \circ z$ is twice differentiable and it is true around any point so $\varphi$ is twice differentiable on $\mathbb{R}$.

However, if you didn't see covering spaces, your handmade proof seems to be a good idea.

Cactus
  • 6,123
  • 1
    All I can find is the path lifting property. How does that work for $z$ on open interval? Is that something related to the fact that $p$ is a universal cover?

    Your answer was very helpful, now I know where to look at least.

    – keplerr May 23 '23 at 10:00
  • You can lift any path from a closed or an open interval even if the path lifting theorem often only use closed interval. But if you can lift $z : [-A,A] \rightarrow \mathbb{S}^1$ uniquely, make $A$ grow toward infinity and you deduce that can can lift $z : \mathbb{R} \rightarrow \mathbb{S}^1$ uniquely. – Cactus May 23 '23 at 10:54