3D Orientation: Pitch, Yaw & Roll Detection | Marvelmind

0:01 Hello colleagues. Today I will be talking about pitch, yaw, and roll for the system on the example of a camera. We just received one more time a request: is it possible to do with the system or not? And if it's possible, what is required? We decided, no, let's give a video explanation. So yes, it is possible to have not only precise location of your, for example, camera or any other object, but also pitch, yaw, and roll, and a very precise one. How it will work: the recommended solution. Let me give an example of the camera. When the camera is in the middle of, let's say, 20 by 20, the distance between the stationary beacons is, for example, 20, 20, 20 meters.

0:59 A basic simple 3D submap, and somewhere in the middle you have a camera. The task is to know the very precise angle, let's say 3D direction of the camera. How could it be done? Now, first of all, you would need to have a few location points with the base that would let you measure this direction. The direction would be measured as the very precise measurement of the location of each of these points. And so, effectively, you would need to install three or more—in this case, three or four—stationary beacons around the camera inside the room or inside the area where you're shooting. And on the camera, you would

1:57 need to install Super-Beacon or, if you use Super-Beacons with external microphones, it's possible. Okay, I'm using a basic beacon, Beacon HW v4.9, but since they look the same, it would be the same. Why I'm suggesting to use the external microphone instead of the microphone on the Super-Beacon is simply for convenience. You could use the Super-Beacon itself. And remember that the location—because in this case the location would be measured with a very, very high precision—it's very, very important that you would be measuring the location of the microphone, not the center of the sensor or not the middle, but the microphone. In this case, it's easy. Simply, again, for practical purposes, to have an external microphone, like in this case. So you have a microphone installed on

2:57 whatever, on a small stick with the base from the camera. So you have a couple of things to remember. One is that you need to have as large a base as you can afford. Usually it's limited by your camera or by the size of your object. Second, they must not be, let's say, in a line. Because you need 3D angle pitch, yaw, and roll, it means that you cannot install them in one line. In this case, you wouldn't be able to measure, for example, the Z position. Because in this case, you would need to have a third microphone not on the line. So they must be in volume. They must have a triangle. And what happens is you would be able to measure, in this perfect condition, the location of each of these points with plus or minus roughly plus or minus one

3:56 centimeter or even better. Because there is a trade-off between the location update rate and the accuracy. With high update rate, the noise and the precision of measurement would increase significantly. But the latency would increase significantly. So it meant that for static objects, you may have accuracy higher than one centimeter—plus or minus one centimeter. But if you need movement and you want to minimize the latency, then it could be higher than plus or minus two, plus or minus one centimeter, like our regular plus or minus two centimeters. So it means that the location of each of these microphones—external microphone—would be measured with, let's say, let's assume plus or minus one centimeter. And if you have a 50-centimeter base, which is typical for, let's say, a small to medium-sized camera which you can afford, so in this case you would be able to measure

4:55 this with plus or minus two degrees accuracy. Overall accuracy of pitch, yaw, and roll—it's a bit too complex to calculate because it depends on their relative position. Is the triangle, was the difference, was there, let's say, not even ratio, but what's the relative position between this triangle of the microphones and the triangle of those? The mathematics would be extremely complex and not practical to use. This is why I'm giving just an idea of what kind of precision and accuracy you would be able to get. So in this case, you would be able to get plus or minus two degrees, but that's for a worst-case scenario when you have plus one centimeter accuracy in this microphone and minus one centimeter in this microphone. But that would be one of their angles.

5:55 Pitch, yaw, roll—depending on where they are located—but you would have all three. And all three will be sensor fused with the IMU which is inside the beacon. So it meant that you would have not only very, very, very precise measurements in terms of measurement based on IMU or gyroscope in this case inside, but also these microphones would eliminate the drift of the gyroscope. But the static maximum static drift you could expect is around this value. So for the base of 50 centimeters, around two degrees for this particular angle. For other angles, no—it would be corresponding to this maximum difference. So, answering overall: yes, it is possible to have pitch, yaw, and roll

6:53 based on our system. That would require at least three external microphones. And in order to connect the three external microphones, you have two options. Now, first of all, you can connect to one Super-Beacon. But in this case, this one Super-Beacon would effectively receive three absolutely independent channels. So it means that if in general you would be able to measure up to 30 meters with three, you would reduce the maximum distance to 10 because it would be 10 plus 10 plus 10—three independent microphones. If the size of your room—let's say the maximum distance between this and this, this and this, this and this—is below 10 meters, then one Super-Beacon and three external microphones would be sufficient for the system. If it's not, and you have a larger room, then you would need to have one Super-Beacon per each

7:51 external microphone. As mentioned, you could use the microphone of the Super-Beacon, but in many cases it's simply inconvenient because the external microphone is tiny. It's not obstructing other microphones, and you could put it anywhere. But the Super-Beacon placed here is simply not so much convenient because you can create obstruction undesirably, and the shadow from this, for example, beacon to this beacon—because there will be a frame that would be holding this beacon, and the frame for a Super-Beacon would be, you know, larger than the frame for the microphone. So from a practical point of view, we suggest to use the external microphones, but you could use basic Super-Beacons as well. Again, you have two options: either one Super-Beacon with three external microphones, but in this case the maximum distance would be below 10 meters, or three Super-Beacons. But in this case

8:51 the maximum distance would be 30 meters. Now, that's basically it because it's relatively simple—a simple configuration, but a rather special configuration in that regard. Usually we either provide the location or location plus direction, but in this case it's not only location plus direction but 3D direction, and a pretty special arrangement in order to receive both compensated-by-drift direction and location at the same time. So if you ask questions like this, we are happy to help and explain the details. And the very short answer to this—one more time: yes, it's doable. Yes, it's not so complex a configuration after all. Thank you very much.