PlanetMath (more info)
 Math for the people, by the people.
Encyclopedia | Requests | Forums | Docs | Wiki | Random | RSS  
Login
create new user
name:
pass:
forget your password?
Main Menu
Owner confidence rating: High Entry average rating: No information on entry rating
Serret-Frenet equations (Theorem)

Let $ I\subset\mathbb{R}$ be an interval, and let $ \gamma:I\to\mathbb{R}^3$ be an arclength parameterization of an oriented space curve, assumed to be regular, and free of points of inflection. Let $ T(s)$, $ N(s)$, $ B(s)$ denote the corresponding moving trihedron, and $ \kappa(s), \tau(s)$ the corresponding curvature and torsion functions. The following differential relations, called the Serret-Frenet equations, hold between these three vectors.

$\displaystyle T'(s)$ $\displaystyle =$ $\displaystyle \kappa(s) N(s);$ (1)
$\displaystyle N'(s)$ $\displaystyle =$ $\displaystyle -\kappa(s) T(s) + \tau(s)B(s);$ (2)
$\displaystyle B'(s)$ $\displaystyle =$ $\displaystyle -\tau(s) N(s).$ (3)

Equation (1) follows directly from the definition of the normal $ N(s)$ and from the definition of the curvature, $ \kappa(s)$. Taking the derivative of the relation

$\displaystyle N(s)\cdot T(s) = 0,$
gives
$\displaystyle N'(s)\cdot T(s) = - T'(s) \cdot N(s) = -\kappa(s).$
Taking the derivative of the relation
$\displaystyle N(s)\cdot N(s) = 1,$
gives
$\displaystyle N'(s) \cdot N(s) = 0.$
By the definition of torsion, we have
$\displaystyle N'(s)\cdot B(s) = \tau(s).$
This proves equation (2). Finally, taking derivatives of the relations
$\displaystyle T(s)\cdot B(s) = 0,$    
$\displaystyle N(s)\cdot B(s) = 0,$    
$\displaystyle B(s)\cdot B(s) =1,$    

and making use of (1) and (2) gives
$\displaystyle B'(s) \cdot T(s) = -T'(s)\cdot B(s) = 0,$    
$\displaystyle B'(s) \cdot N(s) = -N'(s)\cdot B(s) = -\tau(s),$    
$\displaystyle B'(s)\cdot B(s) = 0.$    

This proves equation (3).

It is also convenient to describe the Serret-Frenet equations by using matrix notation. Let $ F:I \to \operatorname{SO}(3)$ (see - special orthogonal group), the mapping defined by

$\displaystyle F(s) = (T(s),N(s),B(s)),\quad s\in I$
represent the Frenet frame as a $ 3\times 3$ orthonormal matrix. Equations (1) (2) (3) can be succinctly given as
$\displaystyle F(s)^{-1} F'(s) = \begin{pmatrix} 0 & \kappa(s) & 0 \ -\kappa(s) & 0 & \tau(s) \ 0 & -\tau(s) & 0 \end{pmatrix}$
In this formulation, the above relation is also known as the structure equations of an oriented space curve.



Anyone with an account can edit this entry. Please help improve it!

"Serret-Frenet equations" is owned by rmilson. [ full author list (3) | owner history (1) ]
(view preamble)

View style:

See Also: space curve, torsion (space curve), curvature (space curve)

Other names:  Frenet equations, Frenet-Serret equations, Frenet-Serret formulas, Serret-Frenet formulas, Frenet formulas

Attachments:
expressions for curvature and torsion (Theorem) by Simone
Log in to rate this entry.
(view current ratings)

Cross-references: structure, orthonormal, represent, mapping, special orthogonal group, matrix, derivative, equation, vectors, relations, moving trihedron, points of inflection, oriented space curve, arclength parameterization, interval
There are 4 references to this entry.

This is version 16 of Serret-Frenet equations, born on 2002-02-02, modified 2007-06-02.
Object id is 1636, canonical name is SerretFrenetFormulas.
Accessed 16867 times total.

Classification:
AMS MSC53A04 (Differential geometry :: Classical differential geometry :: Curves in Euclidean space)

Pending Errata and Addenda
None.
[ View all 5 ]
Discussion
Style: Expand: Order:
forum policy
A note for Mr. rmilson by perucho on 2005-07-20 11:08:24
Dear rmilson,
In relation to your entries 'space curve' and 'Frenet-Serret equations', although you make clear that \gamma:I \to \mathbb{R}^3 is an 'arclength parameterization', I see disadvantage of using letter 't' to denote space since is practical universal to use it to designate 'time' and such parameter is essential in 'movement'. Let's recall that it is usual to define 'speed' as v=ds/dt reason why the mentioned designation could cause great confusion to the novice. So I allow myself to suggest that you make the suitable change in the referred parameterization in order to avoid confusion.
Best,
Pedro
[ reply | up ]

Interact
post | correct | update request | prove | add result | add corollary | add example | add (any)