Hi,(adsbygoogle = window.adsbygoogle || []).push({});

I have a faced a research problem where I would need to recover a frame field given its connection forms. More precisely, I begin with an orthonormal frame field (given by data) in [itex]\Re^3[/itex] written as

[tex]

\mathbf F=\begin{pmatrix}\vec f_1\\\vec f_2\\\vec f_3\end{pmatrix}

[/tex]

where [itex]\vec f_i:\Re^3\rightarrow\Re^3[/itex] are vectors fields with [itex]\vec f_i\cdot\vec f_j=\delta_{ij}[/itex], and [itex]\delta_{ij}[/itex] is the Kronecker delta. I then obtain the connection forms

$$

\omega_{ij}=\text{d}\vec f_i\cdot\vec f_j,

$$

which yields

$$

\text{d}\mathbf F=\Omega\mathbf F

$$

where [itex]\Omega=[\omega_{ij}]\in\Re^{3\times3}[/itex] is the skew-symmetric matrix of connection forms.

In my application, I then proceed by computing the interior product of the 1-forms [itex]\omega_{ij}[/itex] onto the frame fields themselves (the Christoffel symbols), i.e., I compute

$$

\omega_{ijk}\equiv\omega_{ij}\langle\vec f_k\rangle\in\Re

$$

where [itex]\langle\cdot\rangle[/itex] denotes the standard interior product for forms. I thus obtain 9 different measurements at each point, i.e., [itex]\omega_{121},\omega_{122},\omega_{123},\omega_{131},\omega_{132},\omega_{133},\omega_{231},\omega_{232},\omega_{232}[/itex], each of which has a very precise meaning in the application at hand.

Now, I would like to do the converse, i.e., **I would like to solve for [itex]\mathbf F[/itex] given the list of interior contractions** [itex]\omega_{121},\omega_{122},\omega_{123},\omega_{131},\omega_{132},\omega_{133},\omega_{231},\omega_{232},\omega_{232}[/itex].

The Frobenius theorem states the unique existence of [itex]\mathbf F[/itex] in the neighborhood of [itex]0[/itex] if we set [itex]\mathbf F(0)=I[/itex] and if the following are satisfied:

$$

\Omega=\text{d}(F)F^{-1}

$$

$$

\text{d}\Omega-\Omega^2=0.

$$

but I'm unsure about where to start. Would anybody have a suggestion on how to approach this problem? Is there a formal name to what I'm trying to do?

Perhaps a better way to formulate the problem is to enumerate the frame axis differentials directly:

\begin{align}

\text{d}\vec f_1&=\omega_{12}\vec f_2+\omega_{13}\vec f_3\\

\text{d}\vec f_2&=-\omega_{12}\vec f_1+\omega_{23}\vec f_3\\

\text{d}\vec f_3&=-\omega_{13}\vec f_1-\omega_{23}\vec f_2.

\end{align}

Writing the direction of contraction as

$$\vec v=\sum_i^3(\vec v\cdot\vec f_i)\vec f_i=\sum_i^3v_i\vec f_i$$

we obtain

\begin{aligned}

\text{d}\vec f_1\langle\vec v\rangle&=\left(\sum_i^3v_i\omega_{12i}\right)\vec f_2+\left(\sum_i^3v_i\omega_{13i}\right)\vec f_3\\

\text{d}\vec f_2\langle\vec v\rangle&=-\left(\sum_i^3v_i\omega_{12i}\right)\vec f_1+\left(\sum_i^3v_i\omega_{13i}\right)\vec f_3\\

\text{d}\vec f_3\langle\vec v\rangle&=-\left(\sum_i^3v_i\omega_{13i}\right)\vec f_1-\left(\sum_i^3v_i\omega_{23i}\right)\vec f_2.

\end{aligned}

Which would give us three coupled differential equations to solve...

Thanks in advance for any help!

**Physics Forums - The Fusion of Science and Community**

The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

# Recovering a frame field from its connection forms

Loading...

Similar Threads - Recovering frame field | Date |
---|---|

A Pushforward of Smooth Vector Fields | Jan 12, 2018 |

A The Frenet Serret frame | Apr 4, 2017 |

Null-homologous framing? | Jan 21, 2014 |

Existence of orthonormal frame | Oct 18, 2012 |

Normal coordinate frame | Sep 11, 2012 |

**Physics Forums - The Fusion of Science and Community**