Underwater GPS

[jseyfert3]

Any electronic componentry that currently exists in the consumer does not have the accuracy needed for a device like this, but one day I'm sure it will have..

The interesting thing they kept mentioning was tracking the diver's movements. Thinking about this more since I posted, I could see how doing so could help the accuracy of an inertial tracking system immensely, and also explains why you need to compensate for currents by activating it and swimming at a pace that kept you from moving.

They claim they already have the system working, and only need to shrink the wrist unit, which is a huge box as shown in the video(s). They also claim a Q4 release date, so we shall see soon if their claims hold up...

 

[lowviz]

...//... and also explains why you need to compensate for currents by activating it and swimming at a pace that kept you from moving. ...

That is a necessary "cheat" to make it work. You really want to dive like that?

I actually thought about doing something similar to this using an Arduino based controller and a variety of sensors. There is substantial maths involved, but not determine the motion vectors from the sensor inputs - more so to compensate for drift. ...//...

Yes, drift (being in a constant translational reference frame) with respect to the bottom or a tied-in boat would be a killer...

...//... Once you start tracking motion vectors in 3D, error margins are multiplied exponentially very quickly, and become difficult to account for.

Add a decent depth sensor as an error limiter, but even so, I totally agree. But then, all my experience is with older Kulite sensors and most certainly things have improved.

...//... Any electronic componentry that currently exists in the consumer does not have the accuracy needed for a device like this, but one day I'm sure it will have..

Again I agree. I would be the very last person to dampen the innovator's spirits. Go for it! But I'm not about to buy one just now, either.

Would be fun to dive and mess with a prototype, though...

 

[Nemrod]

Apollo used a Star Tracker system integrated in to the nav system and guidance was updated by the astronauts using a sextant also integrated in to the guidance system to re allign the Inertial system periodically through the mission.
Submarines have a much more sophisticated system than can be strapped to a wrist.

Okay. N

 

[stiebs]

They claim they already have the system working, and only need to shrink the wrist unit, which is a huge box as shown in the video(s). They also claim a Q4 release date, so we shall see soon if their claims hold up...

Have you seen how many failed Kickstarter and Indigogo projects say exactly th4 same thing.. we have a working prototype, we just need your help to bring it into production... we've proven the technology, we just need to get it out to the world..

Remember this one?
https://www.indiegogo.com/projects/triton-world-s-last-artificial-gills-scam#/

I will believe it when I see it, but in the meantime, I won't be holding my breath...

 

[stiebs]

Yes, drift (being in a constant translational reference frame) with respect to the bottom or a tied-in boat would be a killer...

My reference to drift in that context was electronic sensor drift, not drift as a result of water current...

If you take a sensor, or combination of sensors including gyroscopes, accelerometers and IMUs, you can fairly trivially track the movement of an object in a 3D space.. say down X units, across Y units and up Z units, and be pretty sure where it ended up within a margin of error. There will be some sensor drift meaning that it won't be exactly where you think it should be.

Multiply that sensor drift over hundreds of lateral and vertical movements, and you will be nowhere near where you think you should be.

Another option could use some of sort sonar mapping for course resolution, and inertial movement to fill in some of the gaps

 

[lowviz]

My reference to drift in that context was electronic sensor drift, not drift as a result of water current... ...

OK, that too.

Let me try to explain my issue, correct me where I need it:

Imagine two nonaccelerating inertial reference frames. One is fixed at the boat's tie-in and never moves. The second is fixed on a particle in the current that is flowing past the tie-in. The current has constant velocity, no acceleration. Same set of sensors at each datum. Everything reads zero on both sensor arrays. How does one track the widening distance between the two? You don't. Your only hope of getting back is to record the acceleration of going from one frame to the other and multiplying it by time. You get one and only one chance at this.

So let's say I jump off a tied-in boat. I will very quickly come to the current's velocity and direction, it will happen during all the sensor overload of the splash itself. The sensors on me need to balance all the forces of the splash and come up with a resultant, that of the current itself. That is the only hope of tying the two frames together with respect to distance.

Is this even remotely possible IRL?

... If you take a sensor, or combination of sensors including gyroscopes, accelerometers and IMUs, you can fairly trivially track the movement of an object in a 3D space.. say down X units, across Y units and up Z units, and be pretty sure where it ended up within a margin of error. ...

Every one of those requires an acceleration. You can be pretty sure how you moved about in your reference frame, but also need to know the really big detail of where the boat is...

 

[stiebs]

I think our thinking is aligned

Although my knowledge of this field is now being stretched... I believe it is possible, but at least within our current technology capabilities, improbable, to track movement without a fixed frame of reference, but within somewhat narrow environment constraints.

You're right in saying that as soon as the subject reaches a constant velocity, eg drifling in a current or swimming at a constant speed and direction, all reference is lost.

With little to no current, or possibly surge rather than current, and a diver changing direction sufficiently, it is theoretically possible. I think. Albeit with the large error margins I've mentioned above

 

[NAM001]

Interesting discussion. A lot of reference to submarine navigation was used. Probably a convienient comparable platform because of non availability to gps data while submerged. other than that there is no difference to a sub over a surface vessel or plane. A device like this has to run on accelerometers and not gyro's. no change of force no change in calculated values. The water current issue. It also has to have a stable platform which is the most difficult part of getting correct force reading appplicatons, or the application of those forces in the correct x,y,z axis. That stable platform is the hardest to achieve. additional sensors are needed to correct for the constant instability of the platform (your wrist). IE... a level device going 10 kt forward will record forces of 10 forward and 0 vertical motion. tip your wrist a little and those forces will be perhaps applied as if you were going 8kt forward and 5 vertical. swing your arm a little and portion in the N direction becomes NNE. In the case of the first,,,, a depth measurement can be used to cancel out the arm tilt. depth input would be used as a check and balance for the measurement/ appplication validity. in the case of the second I don't know off hand how that would be corrected for. Also for a device like this with all the sensors I would imagine it uses a lot of power. for drift/current calibration as long as you are DIW you can use that value in the computer and then commence drifting and establish the drift speed and direction, From that point you can use it to correct calculated data readout. If set and drift changes during the dive, you again introduce error. Get behind a rock where the current does not move you and the unit would continue to show you moving. Currents definitely complicate the accuracy of this device.

Someone mentioned your cell phone showing distance, direction and speed. I suspect that those phone calculations are done from a GPS input and not an accelerometer input so the cell phone may be a bad device for comparison.

Looking forward to reading more discussion.

 

[stiebs]

A few errorseconds in your response KWS. First off, it might be worth your while reading up on gyros, accelerometers and IMUs.

Gyros detect rotational movement, and quite easily correct for tilt and rotational movement such as on your arm.

And most phonestablished absolutely have accelerometers or IMUs. Look up the android app "send me to heaven"... if it's still available. It usespecially the phones accelerometer to determine how far you can throw it into the air and catch it again. It does sanity checks against gravitational forces to ensure it is thrown and not launched by some other mechanism. If you can still download it, I don't suggest you try to top the leader board

The other much used example of a phone's accelerometer is with the Samsung VR headset... it's the accelerometer that maps your head movement into movement in the VR world that you're immersed in.

And while I'm at it, if you use Facebook, you might have seen people posting panoramic photos, and as you move your phone left and right and tilt up and down it moves around the photo. That's the accelerometer at work,

 

[jseyfert3]

Have you seen how many failed Kickstarter and Indigogo projects say exactly th4 same thing.. we have a working prototype, we just need your help to bring it into production... we've proven the technology, we just need to get it out to the world..

Remember this one?
https://www.indiegogo.com/projects/triton-world-s-last-artificial-gills-scam#/

I will believe it when I see it, but in the meantime, I won't be holding my breath...

Yeah, I'm well aware. This one seems different though. Definitely plausible, unlike the artificial gills. And unlike the artificial gills they don't seem to have ever asked for money on Kickstarter or Indigogo, so if they are working on this like they say they are getting their money from elsewhere.