Generalizing the notion of parallel to curved lines in $\mathbb{R}^2$ and $\mathbb{R}^3$ for an arc-length parameterized line

49 Views Asked by At

Let $a(s)$ be an arclength parameterized plane curve. We can create a "parallel" curve $b(s)$ by setting $b(s)=a(s) + eN$ for some small $e>0$.

I think that this defintion makes sense, because given any $s \in D$, $b'(s)$ and $a'(s)$ will be parallel lines.

Now, let us consider $a(s)$ as an arc-length parameterized curve in space. I want to show that we can obtain a "parallel" curve $b(s)$ by taking $b(s)=a(s) + e((cos(\theta)N + sin(\theta)B$) for an appropriate function $\theta$.

So, i'm a bit confused here. How does the idea of two lines being parallel generalize to three dimensions? By showing that their tangent lines are parallel somehow? Or could we get a cross product involved and show that their binormal vectors are always parallel or something?

Furthermore, what is this function $\theta(s)$ that i'm looking for? My first thought was to use the $\theta(s)$ s.t. $T=cos(\theta(s))i+sin(\theta(s))j$ which exists because $\alpha$ is smooth I believe. But i'm just not sure.

Can anyone offer some general or specific insights into the situation?

Also, I'm looking for an easy way to express the curvature of $b(s)$ in terms of $a(s)$ in the first case where both $a(s)$ and $b(s)$ are plane curves.

Thanks a ton!