If the tangents are parallel at each point for two curves, then so do their principal normal and binormal vectors

151 Views Asked by At

In the book of Differential Geometry by Kreyszig, at page 103, it is asked that

Problem 13.1: Given two twisted curves which are in a one-to-one correspondence so that at corresponding points the tangents are parallel. Prove that at corresponding points the principal normals, and the binormals are parallel.

Let we have two curves satisfying the conditions s.t one is parametrised by $\vec x(t)$ and the other is $\vec y(t)$ s.t $c(t) \vec y(t) = x(t) c(t)$ for some real-valued scalar function, then the calculations blows up, I get nothing; but we just assume $c(t) = a$ is constant, then everything is almost trivial, so how do show this when $c(t)$ is not a constant function ?

Note that, c(t) doesn't even need to be a continuous function.

2

There are 2 best solutions below

2
On

A beginning: You have two curves $t\mapsto{\bf x}(t)$ and $t\mapsto{\bf y}(t)$, and it is assumed that $\langle\dot{\bf y}(t)\rangle= \langle\dot{\bf x}(t)\rangle$ for all $t$. Here $\langle\ldots\rangle$ denotes the span of the listed vectors. We therefore may write $$\dot{\bf y}(t)=\lambda(t)\,\dot{\bf x}(t)\qquad\forall\,t\tag{1}$$ for some scalar function $t\mapsto\lambda(t)$. Differentiating $(1)$ we see that $$\ddot{\bf y}(t)=\dot\lambda(t)\,\dot{\bf x}(t)+\lambda(t)\,\ddot{\bf x}(t)\qquad\forall\,t\ .\tag{2}$$ From $(1)$and $(2)$ we can infer that not only $\langle\dot{\bf y}(t)\rangle= \langle\dot{\bf x}(t)\rangle$, but we have also $$\langle\dot{\bf y}(t),\ddot{\bf y}(t)\rangle= \langle\dot{\bf x}(t),\ddot{\bf y}(t)\rangle\qquad\forall\, t\ .\tag{3}$$ From $(3)$ we now can conclude that the principal normals ${\bf n}_{\bf y}$ and ${\bf n}_{\bf x}$ are parallel, because they are lying in the same plane, and are orthogonal to the same vector $(1)$in this plane.

The binormals, being cross products $\dot{\bf y}\times{\bf n}_{\bf y}$, resp., $\dot{\bf x}\times{\bf n}_{\bf x}$ , then are parallel as well.

0
On

What the statement assumes is that there exists some bijection $\phi:t\to s$ between the domains of the curves such that $y'(t)=c(t)x'(s)$ for some function $c(t)$, that is, $y'(t)=c(t)x'(\phi(t))$. Then $$ y''(t)=c'(t)x'(\phi(t))+c(t)\phi'(t)x''(\phi(t)) $$ and, therefore, $y''(t)$ is parallel to the osculating plane defined by $x'(\phi(t))$ and $x''(\phi(t))$. Therefore, both osculating planes are parallel and directions on them, perpendicular to parallel directions (of $y'(t)$ and $x'(\phi(t))$ coincide. The coincidence of the binormal directions follows from the fact that the osculating planes are parallel. The statement actually asserts that not only the directions, but the orientations of the principal normals and binormals coincide. I think it is safe to assume here that $c$ and $\phi$ are twice differentiable functions and that $c>0$. Since the tangent vectors are rotating in the same direction, and the osculating planes also, both the principal normals and the binormal are actually parallel.