kjorn wrote: ↑
Mon Sep 25, 2017 10:26 pm
How come the woo, piq and xsensor all show the same height?
You say these devices dont work, when just using one shows that they do work. I dont understand what point you are trying to make. You wish they didnt work? You dont believe they work?
How do you think VR and AR work using phones as screens? Sub-millimeter tracking in a 3d space. Using all the sensors you seem to think dont work.
They don't show the same height. Woo1 users were complaining Xensr giving higher readings. Now Woo2 is giving different readings (closer to Xensr I believe)..
The point I was saying that verifying algorithm to detect height is just not tossing IMU airborne against tape measure like you suggested. Because for the reasons I mentioned. If you develop freefall IMU you get a freefall IMU which is unable to approximate kiteboarding jumps. For example Recon Instruments' hud is a freefall device and it cannot detect my snowkite jumps nevertheless how high I jump.
Augmented reality relies heavily on camera tracking. Look at Blender AR videos already years back where CG is on top of normal video (non-IMU) camera footage. IMU-sensors of course can be used to aid camera tracking but AR is not based on IMU. Consumer VR (Gear VR, Rift) is more based on gyroscopes although accelerometer can be used to detect "initial" orientation. Some VR devices use cameras and beacons as well where you also need to capture movement. But in the end AR & VR proves nothing about kiteboarding IMU.
Then you put "sensors don't work" into my mouth. I never said sensors not working. Only thing about sensors I said was the noise which is normal on analog-to-digital devices. But if noise has "bias" the device needs to cancel it. Also sensors have different accuracies, some are accurate enough for mobile phones, some are accurate enough for IMU device and some are accurate enough for space rockets and ballistic missiles.
What I said was the algorithms inside the device. With kiteboarding device you need totally different algorithms to freefall device. And you cannot simplify it to calibrate your device height with tossing device against tape measure. You need to verify that device does correct in flat water, chop and waves. Accelerometer is constantly pounded, gyroscope rotated and from this environment you need to detect significant upwards speed to detect jump (because you cannot detect jump from 0g MEMS accelerometer is giving on freefall).
I had 1 hour session and Woo1 shows 27 jumps and Woo 2 77 jumps.
On the highest jump Woo1 7,9 m and Woo2 8,7 m .
Seems like Woo 2 process and record all data faster and more exact while Woo 1 kind of rounding up everything .
And this proves my point also between Woo1 and Woo2. Same company providing solution for kiteboarding height measuring and they are 10% off from each other which is not "sub-millimeter accuracy you state". Both devices could
use same knowhow and software algorithm (sw upgrade to Woo1 to Woo2 level), but Woo was forced to upgrade hardware? Better sensors? Faster CPU? More memory? Bigger battery? Better waterproofing? Or all mentioned?
Then what is verifying that Woo2 8.7m is correct? That's why I like to call it "woo-meters".