0

Now I have quaternion data representing the orientation of wearable devices. Because people wear it differently, the initial quaternion is different for each person. I want to align their movement data based on their initial quaternions.

For participants 1 and 2, their initial orientations are p1 and p2, their movement quaternions are q1 and q2. I want to align q2 to q1, so I tried:

r = p1 * p2', where p2' is the conjugate of P2

q1 = r * q2

based on another post, but it doesn't work.

Russell
  • 21
  • 2
  • Please give us the actual quaternions you are using (i.e., post the numbers). And then tell us the numerical result you want. And can you explain in a bit more detail what the phrase "align their movement data" means to you? – James Tursa Nov 10 '21 at 18:56
  • Because people wear the device differently, the final quaternion we receive are very different, for example, first people's quaternion q1 as the movement data is [0.7657 0.6394 0.0685 0.0032], but the second people's quaternion q2 is[0.7141 0.3338 -0.6152 0.0027], their coordinates have different ranges. I want to rotate q2 to q1, so that when they share similar ranges, I can do data mining on it. I assume that when they stand still, the quaternion data can reveal the orientation of the wearable device, so I hope to use the rotation from p2 to p1 to rotate q2 to q1. – Russell Nov 11 '21 at 02:44

0 Answers0