Euler Angles

1. NASA Standard

1.1 Rotation Matrix

Euler angles are three successive rotation angles that determine the orientation of a reference frame (x, y, z) with respect to another reference frame (xo, yo, zo).  They can be defined in various ways, and the notation has not been standardized.  See for example http://mathworld.wolfram.com/EulerAngles.html. The choice of Euler angles to be made in what follows is referred to in http://www.martinb.com/  as the NASA Standard Airplane convention.  The orientation matrix obtained in this last reference has a sign error.

Let   and  be sets of orthonormal right-handed base vectors corresponding to axes (x, y, z) and (xo, yo, zo), respectively.  The process of rotating   into   uses two intermediate sets of axes, having base vectors  and, respectively.  Let

b3 = rotation angle about  bringing  into

b2 = rotation angle about  bringing  into

b1 = rotation angle about  bringing  into

The way in which these three angles are defined from the geometry of given frames and  will be dealt with after obtaining the rotation matrix resulting from the successive application of the three rotations.  We have

 

           (1)                

 

 

 

 

           (2)           

 

 

 

               (3)

 

 

Substituting from (1) into (2), and from (2) into (3) yields

                                                      (4)

where

                                                                    (5)

Performing the matrix product, and adopting the abbreviating notation

(c1,  c2c3) = (cos b1, cos b2, cos b3)                                 (6)

(s1s2,  s3) = (sin b1, sin b2, sin b3)                                      (7)

yields

              (8)

 

The figure shows how angles b3, b2, and b1 are defined from given axes (x, y, z) and (xo, yo, zo).  Plane (y, z) intersects plane (xo, yo) along a line oriented and labeled as axes y1 and y2.  Axis z1 coincides with axis zo, and axis x1 is on the perpendicular to axes y1 and z1, and is oriented such that the system (x1, y1, z1) is right-handed.  Being perpendicular to axis zo, axis x1 is in the (xo, yo) plane.  b3 is the angle of rotation about axis zo that rotates axes (xo, yo, zo) into axes (x1, y1, z1).  Axes (x2, z2) are defined in the plane perpendicular to axis y2, with axis x2 coinciding with the x axis.  Axes (x2, z2) and (x1, z1) are coplanar since they are all perpendicular to the same axis

 

 

 

 

y2 or y1.  b2 is the angle of rotation about axis y1 that rotates axes (x1, y1, z1) into axes (x2, y2, z2).  Finally, axes (y2, z2) and (y, z) are coplanar since they are all perpendicular to the same axis x2 or x.  b1 is the angle of rotation about axis x that rotates axes (x2, y2, z2) into axes (x, y, z).  All rotation angles are defined according to the right-hand rule.

The preceding geometric constructions fails if plane (y, z) coincides with plane (xo, yo).  In that case, axes x and zo are collinear.  Two cases need to be considered

 

Case

If , axes x and zo coincide, and axes xo, yo, y, and z are coplanar.  Let

b3 = rotation angle about the zo axis that brings axis yo into axis y.

Axis xo is thereby rotated into axis x1, which is collinear with axis z, but oriented in the opposite direction. Thus, noting that, frame  is rotated into.  A rotation of  about the y axis by angle

b2 =  -p /2                                        (9)

brings axis x1 into axis x, and axis z1 into axis z, thus frame  into frame.  We can apply the general results to the present case by letting R1 be the identity matrix, and substituting b2 =  -p /2  into the expression of R2 in Eq. (2).  This yields

 

                               (10)

and

                         (11)

We thus obtain

                     (12)

 

Case

If , axis x is collinear with zo and opposite in direction, and axes xo, yo, y, and z are coplanar.  Let

b3 = rotation angle about the zo axis that brings axis yo into axis y.

Axis xo is thereby rotated into axis x1, which coincides with axis z. Thus, noting that, frame  is rotated into frame.  A rotation of  about the y axis by angle

b2p /2                                             (13)

brings axis x1 into axis x, and axis z1 into axis z, thus frame  into frame.  Substituting b2p /2  into the expression of R2 in Eq. (2), then applying Eq.(11) yields

 

                                   (14)

 

                 (15)         

 

 

1.2 Angular Velocity

Let (b1, b2, b3) be functions of time, and  be the corresponding angular velocity of  relative to.  The motion of  relative to, or its absolute motion, may be considered as composed of three relative motions:

a) the motion of  relative to, whose angular velocity is

b) the motion of  relative to, whose angular velocity is

c) the motion of  relative to, whose angular velocity is

By a theorem of rigid body kinematics, the absolute angular velocity is the sum of the relative angular velocities.  Thus

                                         (16)

To obtain the components, of on, we have

                 (17)

From (3) and (2) we have  and.  Eq. (17) turns into

                  (18)

or, with reference to Eqs. (3) and (4),

 {1 0 0}  + {second column of R1} + {third column of R}

From Eqs. (3) and (8),

 {1 0 0} + {0  cos b1 -sin b1} + {-sin b2    sin b1cos b2   cos b1cos b2}    (19)

    = {-sin b2     cos b1+sin b1cos b2     -sin b1+cos b1cos b2}            (19)

For the components, of  on, we have

                  (20)

 

or, with reference to Eqs. (4) and (1),

 [first row of R] T + [second row of R3] T + {0 0 1}

From Eqs. (8) and (1),

 {cosb2cosb3    cos b2sin b3  -sin b2}  + {-sinb3  cosb3   0} + {0 0 1}  (21)

       =  {cosb2cos b3 -sin b3     cosb2sin b3+cos b3     -sin b2+}       (21)

 

Case : b1 = 0,  b2 = -p/2

                                        (22)

Case : b1 = 0,  b2 = p/2

                                      (23)

 

2. Nutation, Precession, Spin

A traditional way of defining Euler Angles is associated with the analysis of a body rotating about a body-fixed axis, chosen as the z axis, which is itself rotating about a fixed axis, chosen as the zo axis.  The problem is to describe the orientation of body-fixed axes (x, y, z) with respect to fixed axes (xo, yo, zo).   An intermediate axis x' is defined along the perpendicular to axes z and zo.  Axis x' thus lies in the (xo, yo) plane, and is also along the intersection of the (x, y) plane with the (xo, yo) plane.  The Euler angles are:

q  = angle between the z and zo axes

y  = angle between the xo and x' axes

j  = angle between the x' and x axes

Two intermediate sets of axes are defined by successive rotations as follows:

a) rotation about axis zo by angle y brings axes (xo, yo, zo) into axes (x1=x', y1, z1=zo).  

b) rotation about axis x' by angle q brings axes (x1=x', y1, z1=zo) into axes (x', y', z'). 

c) rotation about axis z by angle j brings axes (x', y', z') into axes (x, y, z).

In the context of a body rotating about a body-fixed z axis, such as a top, y, q, and j are called, precession, nutation, and spin, respectively. For a fixed q, the precession describes the rotation of the axis of the top about the vertical zo axis.  The nutation represents the inclination of the axis of the top with respect to the vertical, and the spin represents the rotation angle of the top about its axis.

 

2.1 Rotation Matrix

As before, let   and  be sets of orthonormal right-handed base vectors corresponding to axes (x, y, z) and (xo, yo, zo), respectively, and let  and correspond to axes (x1=x', y1, z1=zo) and (x', y', z'), respectively.  We have

 

 

 

and

 

We obtain

 

 

2.2 Angular Velocity

The angular velocity of the (x, y, z) frame is the sum of the intermediate relative angular velocities, i.e.

    

The components of the angular velocity on the fixed axes are

Thus

or

                                   

The components on the body-fixed axes are

 

Thus

 

or

The components on the (x', y', z') axes are

 

Thus

 

or,