Rotations and the Euler angles

1 Rotations
Consider two right-handed systems of coordi-
nates, XY Z and x
1
x
2
x
3
, rotated arbitrarily
with respect to one another (see Fig. ). We
would like to be able to link easily the coor-
dinates of any vector

A in the two frames of
reference. Let e
X
, e
Y
, e
Z
be the unit vectors
for the axes of the rst system, and e
1
, e
2
, e
3
the unit vectors for the axes of the second
system. Then, by denition:

A = A
X
e
X
+A
Y
e
Y
+A
Z
e
Z
and

A = A
1
e
1
+A
2
e
2
+A
3
e
3
X
Y
Z
x
x
x
A
A
3
1
2
Fig 1. Projection of the same vector

A onto two dierent right-
handed systems of coordinates.
Then, we can express one set of projections in terms of the other one:
A
1
= e
1


A = (e
1
e
X
)A
X
+ (e
1
e
Y
)A
Y
+ (e
1
e
Z
)A
Z
A
2
= e
1


A = (e
2
e
X
)A
X
+ (e
2
e
Y
)A
Y
+ (e
2
e
Z
)A
Z
A
3
= e
1


A = (e
3
e
X
)A
X
+ (e
3
e
Y
)A
Y
+ (e
3
e
Z
)A
Z
or, in matrix form:
_
_
_
A
1
A
2
A
3
_
_
_ =
_
_
_
e
1
e
X
e
1
e
Y
e
1
e
Z
e
2
e
X
e
2
e
Y
e
2
e
Z
e
3
e
X
e
3
e
Y
e
3
e
Z
_
_
_
_
_
_
A
X
A
Y
A
Z
_
_
_ (1)
Let us analyze the elements of the 3 3 matrix. By denition e
1
e
X
= cos
1X
, where
1X
is the
angle between the two unit vectors e
1
and e
X
. Similarly, all other elements of this matrix depend
only on the various angles between various sets of axes, but are independent of the projected vector

A. It follows that for any other vector

B, we will have automatically:
_
_
_
B
1
B
2
B
3
_
_
_ =
_
_
_
e
1
e
X
e
1
e
Y
e
1
e
Z
e
2
e
X
e
2
e
Y
e
2
e
Z
e
3
e
X
e
3
e
Y
e
3
e
Z
_
_
_
_
_
_
B
X
B
Y
B
Z
_
_
_
In other words, if we know the 3 3 matrix, then we can nd the components of any vector in one
of the systems, if we know them in the other.
Before continuing, let us introduce some simpler notation. We will denote
_
_
_
A
1
A
2
A
3
_
_
_ =

A
body
,
_
_
_
A
X
A
Y
A
Z
_
_
_ =

A
XY Z
and

R =
_
_
_
e
1
e
X
e
1
e
Y
e
1
e
Z
e
2
e
X
e
2
e
Y
e
2
e
Z
e
3
e
X
e
3
e
Y
e
3
e
Z
_
_
_
and therefore we have

A
body
=

R

A
XY Z
. It then follows that:

A
XY Z
=

R
1

A
body
1
where

R
1
is the inverse of matrix

R, and it should be clear that its matrix elements are:

R
1
=
_
_
_
e
X
e
1
e
X
e
2
e
X
e
3
e
Y
e
1
e
Y
e
2
e
Y
e
3
e
Z
e
1
e
Z
e
3
e
Z
e
3
_
_
_
If its not clear, then derive them and check!
We can see that the matrix

R
1
is just the transpose of matrix

R ( by denition, M is the transpose
of N, i.e. M = N
T
, if m
ij
= n
ji
for all i, j). This property is a consequence of the invariance of the
length of any vector under rotations. If we denote:

A
T
body
=
_
A
1
A
2
A
3
_
;

A
T
XY Z
=
_
A
X
A
Y
A
Z
_
then
|

A|
2
=

A

A = A
X
A
X
+A
Y
A
Y
+A
Z
A
Z
=

A
T
XY Z


A
XY Z
=

A
T
body


A
body
(the vector has the same length in any system of coordinates). But

A
body
=

R

A
XY Z


A
T
body
=

A
T
XY Z


R
T
(this last property can be checked easily using the denition of the transposed matrix),
and therefore:

A
T
body


A
body
=
_

A
T
XY Z


R
T
_ _

R

A
XY Z
_
=

A
T
XY Z
_

R
T

R
_

A
XY Z
which implies that

R
T

R = 1

R
T
=

R
1
.
This property is extremely useful, since it allows us to easily nd the inverse of any rotation
matrix, by just taking its transpose.
2 Rotation about one axis
Let us derive the expression of

R
3
for the case where the axes
0Z and 0x
3
are parallel, and the sets of axes XY and x
1
x
2
are
rotated by an angle with respect to one another (see Fig).
In this case, we know (see, for instance, discussion of polar
coordinates) that the relationship between the unit vectors is:
_

_
e
1
= cos e
X
+ sin e
Y
e
2
= sin e
X
+ cos eY
e
3
= e
Z
We can now compute the various dot products; for instance
e
1
e
X
= cos , etc, and we nd

R
3
() =
_
_
_
cos sin 0
sin cos 0
0 0 1
_
_
_
1
2
3
X
Y
x
x
x Z

Fig 2. Rotation by an angle about the

axis Oz = Ox3.
You should check that

R
3
(
1
)

R
3
(
2
) =

R
3
(
1
+
2
) meaning that if I rotate rst by angle
2
followed by a rotation by angle
1
(about the same axis!) its as if I did a single rotation by angle

1
+
2
. Which is true.
2
The inverse matrix is then:

R
1
3
() =

R
T
3
() =
_
_
_
cos sin 0
sin cos 0
0 0 1
_
_
_ =

R
3
()
This makes perfect sense as well; if system 123 is rotated with + with respect to system XY Z,
then system XY Z is rotated with with respect to 123. As a result, the rotation matrices should
have the same form with , and that is precisely what we found.
In the same way, we can write down the matrices for rotations about any other axis. For instance,
if OX and 0x
1
are kept parallel and we perform a rotation by an angle about them, we nd

R
1
() =
_
_
_
1 0 0
0 cos sin
0 sin cos
_
_
_
We can now use the fact that any general 3D rotation can be decomposed into a product of 3
rotations about 3 dierent axes, to nd the form of a general rotation matrix.
3 Eulers angles
We characterize a general orientation of the body system x
1
x
2
x
3
with respect to the inertial system
XY Z in terms of the following 3 rotations:
1. rotation by angle about the Zaxis;
2. rotation by angle about the new x

1
axis, which we will call the line of nodes ;
3. rotation by angle about the new x
3
axis.
These rotations are illustrated in the following gure:
X
Y
Z
1
2
1
2
3
2"
3 3
1
2"
1
2

We can now write the general rotation matrix that links

A
body
with

A
XY Z
as the product of the
3 rotations about the corresponding axes:

R(, , ) =

R
3
()

R
1
()

R
3
() =
_
_
_
cos sin 0
sin cos 0
0 0 1
_
_
_
_
_
_
1 0 0
0 cos sin
0 sin cos
_
_
_
_
_
_
cos sin 0
sin cos 0
0 0 1
_
_
_
3
leading to the rather ugly general formula:

R(, , ) =
_
_
_
cos cos cos sin sin sin cos cos sin cos sin sin
cos sin + cos cos sin sin sin + cos cos cos sin cos
sin sin sin cos cos
_
_
_
Fortunately, we will never need to use this matrix. All we really need is to be able to write the
components of the angular velocity

in both systems of coordinates. Since

describes precisely
how fast the angles vary in time, we have:

=
d

dt
+
d

dt
+
d

dt
=

e
Z
+

e
1
+

e
3
since the three rotations are about these particular axes.
Let us analyze each contribution to

.
1.

= e
Z

(with respect to XY Z system). Following the rotations, we nd that with respect to
123 system, we have:
e
Z
= cos e
3
+ sin e
2
= cos e
3
+ sin (sin e
1
+ cos e
2
)
and therefore:

= sin sin

e
1
+ sin cos

e
2
+ cos

e
3
2.

= e
1

=

(cos e
X
+ sin e
Y
) (with respect to XY Z), whereas
e
1
= cos e
1
sin e
2

= cos

e
1
sin

e
2
with respect to 123.
3.

= e
3

(with respect to 123), whereas
e
3
= cos e
3
sin e
2
= cos e
Z
sin (sin e
X
+ cos e
Y
)

= sin sin

e
X
sin cos

e
Y
+ cos

e
Z
Adding all three components together, we nd that, with respect to the body reference system,

= (sin sin

+ cos

)e
1
+ (sin cos

sin

)e
2
+ (cos

+

)e
3
(2)
while with respect to the inertial reference system:

= (cos

+ sin sin

)e
X
+ (sin

sin cos

)e
Y
+ (

+ cos

)e
Z
So if we can solve the EL equations and nd how these angles vary in time, we can gure out
whats the angular speed in either of the two reference systems.
4 Kinetic energy in terms of Eulers angles
Let us choose the CM as the reference point O, and we will choose the principal axes of inertia as
the body reference frame. The total kinetic energy of the object will be:
T =
1
2
M

V
2
CM
+
1
2

I
CM

=
1
2
M

V
2
CM
+
1
2
_
I
1

2
1
+I
2

3
2
+I
3

2
3
_
4
where
1
= sin sin

+ cos

, etc [see Eq. (2)].

For an asymmetric top, the general formula is rather complicated, and we will not use it. For a
symmetric top with I
1
= I
2
= I
3
, if you put the expressions for
1
,
2
and
3
in and simplify a bit,
you nd:
L =
1
2
M

V
2
CM
+
I
1
2
_

2
+

2
sin
2

_
+
I
3
2
_

+

cos
_
2
U(

R
CM
, , , )
This is our Lagrangian in terms of our 6 generalized coordinates, namely

R
CM
, , , .
5