7

In Spiegel's Outline Of Theoretical Mechanics (more precisely in the Moving Coordinate Systems chapter, § "Derivative Operators") I find (both in the 1968 and the 1977 edition) the following formula:

$D_F \equiv D_M +\omega \times$

where $D_F$ and $D_M$ denote time derivative operators, and $\omega$ stands for the angular velocity of the moving frame $M$ (with respect to a fixed frame $F$).

Is there some explanation for this notation, or is it a misprint?

enter image description here

Sandejo
  • 5,478
  • What exactly is confusing you about this formula? Is it the fact that there is a "cross" but nothing after it? Or do you have a problem even with the Equation (2) which says the same thing, but for an arbitrary vector $\mathbf{A}$? – Philip Oct 16 '21 at 11:56
  • What motivates my question is the fact that there is a cross with nothing after it. Should not the formula also contain some arbitrary vector to be meaningfull? Also, to tell the truth, I undersand the formula above, without being able to derive it . – Vince Vickler Oct 16 '21 at 12:03
  • 2
    The formula is an operator equation: it relates two operators to each other, and operators make no sense on their own, they are always assumed to act on arbitrary functions, it's just notation :) – Philip Oct 16 '21 at 12:05
  • 1
  • https://en.wikipedia.org/wiki/Cross_product#Conversion_to_matrix_multiplication – John Alexiou Oct 17 '21 at 05:41
  • @Philip actually the one-term cross product has meaning on its own, see my answer below. – John Alexiou Oct 17 '21 at 06:14
  • 1
    The passage says equation (2) is derived in problem 6.1, and once that's done one realizes that since the choice of $\mathbf{A}$ is arbitrary in (2), we could insert any $\mathbf{A}$, i.e. we can treat the rest of the stuff as an operator acting on a vector the way $\frac{d}{dt}$ is an operator acting on the vector, that's all there is to that part of it, the derivation in 6.1 is the important thing to understand. As they show, the reason it breaks up into two types of terms is that the basis vectors depend on time ($\hat{i}(t)$ etc...) in one frame so they get differentiated too. – bolbteppa Oct 17 '21 at 11:13

3 Answers3

20

What you're seeing there is an operator equation. What it means is that the two operations on each side of the equation have the same effect on any vector $\mathbf{A}$: $$ D_F \mathbf{A} = D_M \mathbf{A} + \pmb{\omega} \times \mathbf{A}. $$ By analogy, if we have two functions $f$ and $g$, saying that $f = g$ is a more compact way of saying $f(x) = g(x)$ for all $x$.

18

$\newcommand{\e}{\boldsymbol=}$ $\newcommand{\m}{\boldsymbol-}$ $\newcommand{\x}{\boldsymbol\times}$

Given two vectors $\:\boldsymbol\omega\:$ and $\:\mathbf x\:$ represented by their cartesian coordinates \begin{equation} \boldsymbol\omega\e \begin{bmatrix} \omega_1 \vphantom{\dfrac{a}{b}}\\ \omega_2 \vphantom{\dfrac{a}{b}}\\ \omega_3 \vphantom{\dfrac{a}{b}} \end{bmatrix}\qquad \texttt{and} \qquad \mathbf x\e \begin{bmatrix} x_1 \vphantom{\dfrac{a}{b}}\\ x_2 \vphantom{\dfrac{a}{b}}\\ x_3 \vphantom{\dfrac{a}{b}} \end{bmatrix} \tag{01}\label{01} \end{equation} their outer (cross) product is \begin{equation} \boldsymbol\omega\x\mathbf x\e \begin{bmatrix} \mathbf i & \mathbf j & \mathbf k \vphantom{\dfrac{a}{b}}\\ \omega_1 & \omega_2 & \omega_3 \vphantom{\dfrac{a}{b}}\\ x_1 & x_2 & x_3 \vphantom{\dfrac{a}{b}} \end{bmatrix} \e \begin{bmatrix} \omega_2 x_3\m \omega_3 x_2 \vphantom{\dfrac{a}{b}}\\ \omega_3 x_1\m \omega_1 x_3 \vphantom{\dfrac{a}{b}}\\ \omega_1 x_2\m \omega_2 x_1 \vphantom{\dfrac{a}{b}} \end{bmatrix} \e \begin{bmatrix} \hphantom{\m} 0 & \m \omega_3 & \hphantom{\m}\omega_2 \vphantom{\dfrac{a}{b}}\\ \hphantom{\m}\omega_3 & \hphantom{\m} 0 & \m\omega_1 \vphantom{\dfrac{a}{b}}\\ \m\omega_2 & \hphantom{\m}\omega_1 & \hphantom{\m}0 \vphantom{\dfrac{a}{b}} \end{bmatrix} \begin{bmatrix} x_1 \vphantom{\dfrac{a}{b}}\\ x_2 \vphantom{\dfrac{a}{b}}\\ x_3 \vphantom{\dfrac{a}{b}} \end{bmatrix} \tag{02}\label{02} \end{equation} So it's reasonable to use the symbol $\:\boldsymbol\omega\x\:$ for the linear operator (transformation) represented by a $\:3\times 3\:$ antisymmetric matrix \begin{equation} \boldsymbol\omega\x \boldsymbol\equiv \begin{bmatrix} \hphantom{\m} 0 & \m \omega_3 & \hphantom{\m}\omega_2 \vphantom{\dfrac{a}{b}}\\ \hphantom{\m}\omega_3 & \hphantom{\m} 0 & \m\omega_1 \vphantom{\dfrac{a}{b}}\\ \m\omega_2 & \hphantom{\m}\omega_1 & \hphantom{\m}0 \vphantom{\dfrac{a}{b}} \end{bmatrix} \tag{03}\label{03} \end{equation}

$=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!=\!$

Related 1 : Velocity in a turning reference frame.

Related 2 : Vector product in a 4-dimensional Minkowski spacetime.

Frobenius
  • 15,613
4

The answers are already given suffice, so I wanted to offer another use of the $\boldsymbol{r}\times$ term.

Find the mass moment of inertia tensor about the origin of a point mass $m$ located at $\boldsymbol{r}$.

There are two formulas for this, and they both produce the same result

  1. Use $$\mathrm{I}_0 = -m [\boldsymbol{r} \times] [\boldsymbol{r} \times] = -m \begin{bmatrix} 0 & -z & y \\ z & 0 & -x \\ -y & x & 0 \end{bmatrix} \begin{bmatrix} 0 & -z & y \\ z & 0 & -x \\ -y & x & 0 \end{bmatrix}$$

  2. Use $$\mathrm{I}_0 = m \left( \boldsymbol{r} \cdot \boldsymbol{r} - \boldsymbol{r} \odot \boldsymbol{r} \right) = m \left( \begin{bmatrix} x \\ y \\ z \end{bmatrix}^\top \begin{bmatrix} x \\ y \\ z \end{bmatrix} - \begin{bmatrix} x \\ y \\ z \end{bmatrix} \begin{bmatrix} x \\ y \\ z \end{bmatrix}^\top \right) $$

If you work out the math both result in

$$ \mathrm{I}_0 = m \begin{bmatrix} y^2+z^2 & -x y & -x z\\ -x y & x^2+z^2 & - y z \\ -x z & -y z & x^2+y^2 \end{bmatrix} $$

My preference is to use $[\boldsymbol{r}\times]$ notation because it keeps the connection to cross products evident, and it codes easily in programming.

The above produces the parallel axis theorem in 3D, from the center of mass C to the origin 0 as

$$ \mathrm{I}_0 = \mathrm{I}_c - m [ \boldsymbol{c}\times][\boldsymbol{c}\times] $$

The above has a direct connection if the angular momentum about a pivot point which is

$$ \begin{aligned} \boldsymbol{L}_0 & = \boldsymbol{L}_c + \boldsymbol{c} \times m \boldsymbol{v}_c \\ & = \mathrm{I}_c \boldsymbol{\omega} + \boldsymbol{c} \times m ( \boldsymbol{\omega} \times \boldsymbol{c} ) \\ & = \mathrm{I}_c \boldsymbol{\omega} - m\, \boldsymbol{c} \times (\boldsymbol{c} \times \boldsymbol{\omega}) \\ & = \mathrm{I}_0 \boldsymbol{\omega} \end{aligned} $$

Additionally, constructing the 6×6 spatial inertia matrix you also use this term

$$ \boldsymbol{I} = \begin{Bmatrix} m \boldsymbol{1} & -m [ \boldsymbol{c} \times] \\ m [\boldsymbol{c} \times] & \mathrm{I}_c - m [\boldsymbol{c}\times][\boldsymbol{c}\times] \end{Bmatrix} $$

Now back to taking derivatives on rotating frames, this is done with spatial algebra with the following operator

$$ \mathbf{v} \times = \begin{Bmatrix} [\boldsymbol{\omega} \times & [\boldsymbol{v}\times] \\ [0] & [\boldsymbol{\omega}\times] \end{Bmatrix} $$

So the acceleration due to the motion of joint axis $\mathbf{s}$ is calculated with

$$ \mathbf{a} = \mathbf{v} \times \mathbf{s} \,\,\dot{q} $$

As you can see this notation has extensive usage in robotics and all levels of rigid body dynamics.

John Alexiou
  • 38,341