In 3-space geometry we have curvatures when a point is proceeding along a curved arc. Similarly when particle motion occurs with respect to time we have accelerations.

Is there a one to one correspondence or combination relation among curvatures/ tangent rotations on the one hand and accelerations/velocities on the other?


I hope to be able to see a table of dynamic/geometric analogues.

e.g., for plane curves

Centripetal acceleration ~ normal curvature Coreolis acceleration ~ ? tangential acceleration ~ ? Angular momentum ~ ? ~ geodesic torsion ?

Most importantly a query remains with me longtime:

What are the characteristic accelerations or velocities of a particle moving along a curved asymptotic line of a ( negative curved ) surface?


added tags: curves and differential geometry

  • $\begingroup$ Incidentally, if my answer does address what you were asking, it's probably a good idea to change the tags to "curves" and "differential-geometry". :) $\endgroup$ – Andrew D. Hwang Feb 27 '15 at 16:38

$\newcommand{\Reals}{\mathbf{R}}\DeclareMathOperator{\Tgt}{\mathbf{T}}\DeclareMathOperator{\Nml}{\mathbf{N}}$Your question is qualitative and open-ended, but perhaps the following observations get at what you're asking:

Let $I$ be an open interval of real numbers, and suppose $\gamma:I \to \Reals^{n}$ is a twice continuously-differentiable curve with non-vanishing velocity. For each $t$ in $I$, the velocity $\gamma'(t)$ is non-zero, so we may write $$ \gamma'(t) = \|\gamma'(t)\|\, \frac{\gamma'(t)}{\|\gamma'(t)\|} = v(t) \Tgt(t), $$ with $v(t) = \|\gamma'(t)\|$ the speed at $t$ and $\Tgt(t)$ the unit tangent vector.

By the product rule, the acceleration decomposes as $$ \gamma''(t) = v'(t) \Tgt(t) + v(t) \Tgt'(t). \tag{1} $$ Remarkably, this is the decomposition of the acceleration $\gamma''(t)$ into parallel and orthogonal components with respect to $\Tgt(t)$, i.e., into tangential and normal components: Since $\Tgt(t) \cdot \Tgt(t) = 1$ for all $t$, we have $2\Tgt(t) \cdot \Tgt'(t) = 0$ for all $t$.

If $\Tgt'(t) \neq 0$, we may write $$ \Tgt'(t) = \|\Tgt'(t)\| \frac{\Tgt'(t)}{\|\Tgt'(t)\|} = \frac{\kappa(t)}{v(t)} \Nml(t), $$ with $\kappa(t)/v(t) = \|\Tgt'(t)\|$. Qualitatively, (1) says acceleration can be resolved into a term $v'(t) \Tgt(t)$ measuring how the speed changes at each instant, and a term $v(t) \Tgt'(t) = \kappa(t) \Nml(t)$ measuring how the direction changes at each instant.

The function $\kappa(t)$, the "rate of rotation of $\Tgt(t)$, normalized according to the speed $v(t)$", has a pleasant geometric interpretation as the reciprocal radius of curvature of $\gamma$ at $t_{0}$. Let $s:I \to \Reals$ be the arc length measured along the curve starting at $\gamma(t_{0})$, i.e., $$ s(t) = \int_{t_{0}}^{t} \|\gamma'(\tau)\|\, d\tau,\quad \text{so that $s'(t) = v(t)$ and $s(t_{0}) = 0$.} $$ It's straightforward to check that the curve $$ C_{t_{0}}(t) = \left[\gamma(t_{0}) + \frac{1}{\kappa(t_{0})} \Nml(t_{0})\right] - \frac{1}{\kappa(t_{0})} \cos \bigl(\kappa(t_{0}) s(t)\bigr) \Nml(t_{0}) + \frac{1}{\kappa(t_{0})} \sin \bigl(\kappa(t_{0}) s(t)\bigr) \Tgt(t_{0}), $$ which parametrizes the osculating circle, of radius $r(t_{0}) = 1/\kappa(t_{0})$ lying in the plane through $\gamma(t_{0})$ and spanned by $\Tgt(t_{0})$ and $\Nml(t_{0})$, agrees with $\gamma$ to second order at $t_{0}$, in the sense that $$ C_{t_{0}}(t_{0}) = \gamma(t_{0}),\qquad C_{t_{0}}'(t_{0}) = \gamma'(t_{0}), \qquad C_{t_{0}}''(t_{0}) = \gamma''(t_{0}). $$

In words, if $\gamma'(t_{0})$ and $\gamma''(t_{0})$ are non-zero, then so far as velocity and acceleration (at that instant) are concerned, a particle may as well be moving along a circle of radius $1/\kappa(t_{0})$.

Particularly, if a particle moves at unit speed, i.e., $\gamma$ is "parametrized by arc length", then $\Tgt'(t) = \kappa(t) \Nml(t)$, and the rate of rotation of $\Tgt$ at $t_{0}$ is precisely the reciprocal radius of the osculating circle to $\gamma$ at $t_{0}$.


As user86418 pointed out, your question is too vast. Here is a table with analogous concepts in mechanics and geometry that I can remember now. You can see more details about each one by a simple search on google.

Geometry | Mechanics

tangent vector$\quad\quad$ velocity vector

curvature of a curve$\quad\quad$ acceleration intensity

geodesic curvature$\quad\quad$ tangential acceleration intensity

normal curvature$\quad\quad$ normal acceleration intensity

geodesics$\quad\quad$ path taken by a free particle

Clairaut's relation$\quad\quad$ angular momentum


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.