Universal Robots pose format

The pose format that is used by Universal Robots consists of a position \(XYZ\) in millimeters and an orientation in angle-axis format \(V=(\begin{array}{ccc}RX & RY & RZ\end{array})^T\). The rotation angle \(\theta\) in radians is the length of the rotation axis \(U\).

\[\begin{split}V = \left(\begin{array}{c}RX \\ RY \\ RZ\end{array}\right) = \left(\begin{array}{c}\theta u_x \\ \theta u_y \\ \theta u_z\end{array}\right)\end{split}\]

\(V\) is called a rotation vector.

Conversion from angle-axis format to quaternion

The conversion from a rotation vector \(V\) to a quaternion \(q=(\begin{array}{cccc}x & y & z & w\end{array})^T\) can be done as follows.

We first recover the angle \(\theta\) in radians from the rotation vector \(V\) by

\[\theta = \sqrt{RX^2 + RY^2 + RZ^2}\text{.}\]

If \(\theta = 0\), then the quaternion is \(q=(\begin{array}{cccc}0 & 0 & 0 & 1\end{array})^T\), otherwise it is

\[\begin{split}x = RX \frac{\sin(\theta/2)}{\theta}\text{,} \\ y = RY \frac{\sin(\theta/2)}{\theta}\text{,} \\ z = RZ \frac{\sin(\theta/2)}{\theta}\text{,} \\ w = \cos(\theta/2)\text{.}\end{split}\]

Conversion from quaternion to angle-axis format

The conversion from a quaternion \(q=(\begin{array}{cccc}x & y & z & w\end{array})^T\) with \(||q||=1\) to a rotation vector in angle-axis form can be done as follows.

We first recover the angle \(\theta\) in radians from the quaternion by

\[\theta = 2\cdot\text{acos}(w)\text{.}\]

If \(\theta = 0\), then the rotation vector is \(V=(\begin{array}{ccc}0 & 0 & 0\end{array})^T\), otherwise it is

\[\begin{split}RX = \theta \frac{x}{\sqrt{1-w^2}}\text{,} \\ RY = \theta \frac{y}{\sqrt{1-w^2}}\text{,} \\ RZ = \theta \frac{z}{\sqrt{1-w^2}}\text{.}\end{split}\]