1
$\begingroup$

I'm doing a change of variables that involves doing simple rotations on the standard basis vectors in R^n, and I'm wondering what the standard orientations are in n dimensions are and why. For example, in 3d, looking down the +z axis, +y is CCW from +x and down the +y axis, +x is CCW from +z. This seems obvious, but in 4 dimensions, if we assume that x,y and z have the standard orientations when looking down the +u axis, then when looking down the +z+y plane, is +u CCW from +x or the other way round? And in 6d? The simple rotation matrices in 4d here:

http://ken-soft.com/2009/01/08/graph4d-rotation4d-project-to-2d/

don't seem to have any discernable pattern to me...

I suppose, specifically, I need to know how to methodically create simple rotation matrices in n-dimensions, but if anyone knows where I can learn something deeply about the subject I would appreciate a reference.

$\endgroup$

migrated from mathoverflow.net Apr 11 '14 at 5:34

This question came from our site for professional mathematicians.

2
$\begingroup$

Let me suggest one article, which explicitly describes the simple rotation in $\mathbb{R}^n$ as rotation about an $\mathbb{R}^{n-2}$ subspace:

Mortari, Daniele. "On the Rigid Rotation Conept in $n$-Dimensional Spaces." Journal of the Astronautical Sciences 49.3 (2001): 401-420. (PDF download link)

This reaches the following expression as the $n \times n$ orthogonal matrix that represents a rigid rotation in $\mathbb{R}^n$:


      Eq.18
where $J_2 = \left[ \begin{array}{cc} 0&-1\\ 1&0 \end{array} \right] $, and the meaning of the remaining notation can be found in the paper. Here is a neat figure later in the same paper:
      Coning
      By bending the axes, Figure 2 artistically provides a "way to see" the geometry of coning about the subspace identifed by the three Orthogonal axes $a_1$, $a_2$, and $a_3$.

$\endgroup$
  • 1
    $\begingroup$ Awesome. Also, for anyone else reading this, P is the $n\times 2$ matrix describing the plane of rotation, so, for example, rotating in such a way that a positive angle takes the positive $x$ axis to the positive $y$ axis, $P$ would be $[X,Y]$, where $X = [1,0,0,0..]^T$ and $Y = [0,1,0,0...]^T$ $\endgroup$ – Scott Apr 11 '14 at 14:16

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.