[rogers18445]

What would you do with the g measurement to get your location? Can g deltas be matched to sea floor contour maps? Otherwise it seems like subs would have to be following g-mapped routes.

[etrautmann]

With a highly detailed map, it becomes a statistical fitting problem, to locate yourself based on an estimate of position history using intertidal measurements and your time history of gravitational measurements, starting from an initial position estimate. I would imagine this could be quite accurate.

[quickthrower2]

Dead reckoning with gravitational adjustments?

[tehlike]

Pretty much a slam application.

[etrautmann]

Oops fat fingers - s/intertidal/inertial

[jchallis]

A typical sub drifts by only a few tens of meters per day - obviously this builds up . The point of the BEC gravity map is to reduce this inaccuracy an order of magnitude, thereby extending the subsurface range of the subs.

[rogers18445]

But how would you do this? Sensor fusion with inertial sensors is well understood, how do you fuse g into the equation to improve things?

[wongarsu]

I'd assume features below the sea floor have a much bigger impact than surface features. Large iron deposits, volcanic activity, natural gas deposits, whatever the LLSVPs are, etc. But you could use ships to correlate gravity measurements with GPS locations and make an accurate map that way. You don't need to map the entire ocean, just enough locations to allow subs to occasionally recalibrate their position.

[closewith]

>LLSVP

Large low-shear-velocity provinces, for anyone else wondering: https://en.wikipedia.org/wiki/Large_low-shear-velocity_provi...