VRPN quaternion conversion from PPT from Intersense InertiaCube

[willemsn]

In our application, we've used VRPN and our PPT computer to communicate the position of our tracked points. For orientation, we've mostly accessed our Intersense InertiaCube orientation information directly via a serial port on our Linux box. This has worked fine, and when we've done this, we pulled Euler angles off the Intersense device.

We finally re-arranged the lab space and connected the Intersense device to our PPT system. It recognized the device and we've added the post-processing Intersense module. Our tracked point is now sending out position and a quaternion over the VRPN network to our workstations.

So far so good... however, when we used VRPN's API and specifically, the q_to_euler conversion function to do a sanity check between the yaw, pitch, and roll as seen on our rendering system and that of PPT, we get different values.

Shouldn't we get the same values for yaw, pitch, and roll if these values as seen by PPT are being packaged up correctly into VRPN's quaternions?

We're puzzled by this behavior and thought we 'd ask if anyone has an idea about what is happening on the PPT side for packaging up Yaw, Pitch, and Roll of intersense devices into a VRPN quaternion.

Many thanks!

Pete