Hacker News Reader

Comment by kumarharsh | Hacker News Reader

Kinda funny that the aviation industry cannot implement the more accurate inertial navigation systems (INS) right now because:

"The biggest single problem in trying to implement this change worldwide would be inertia"

View on HN

INS that doesn't drift pretty much immediately is expensive and needs periodic fixups, and the moment GPS became available it pretty much dropped off anyones purchase plans (as it can't be used as backup for GPS unlike VOR/DME)

I think they may have been joking since INS - is inertial navigation.

While INS is not simple, it's also not that difficult to implement. As even the cheapest drones that can hover show new MEMS sensors are accurate enough to work with a proper filtering.

I did some prototype INS system as my master's thesis 10 years ago, the code was quick and dirty and even then the accuracy was like 30 meters after an hour of walking around with the device.

Thing is, it has to work over atlantic/pacific/cross north pole routes, etc. We're talking multiple hours and long distance, possibly with a lot of turbulence (even for GA, which actually gets less steady flight so...).

And then it needs to provide guarantees about said navigation, guarantees that those drones do not need.

That is surprising, you must have had an extremely accurate accelerometer and gyro, or employed tricks [1]. There are two problems with determining position from acceleration measurements:

1. You're integrating twice (acceleration to obtain velocity, then velocity to obtain position). So if you have any noise or error, you're integrating that, and integrate that again. Hello, parabola.

2. Gravity. It's strong. So you have to subtract it (as it induces an apparent acceleration upwards).

If the difference between actual down and where your model thinks is down is just a fraction of a degree, you'll be totally off within minutes.

See eg here: https://www.youtube.com/watch?v=C7JQ7Rpwn2k&t=1401s

Or here: https://www.cl.cam.ac.uk//techreports/UCAM-CL-TR-696.pdf

> As a concrete example consider a tilt error of just 0.05 [degrees]. This error will cause a component of the acceleration due to gravity with magnitude 0.0086 m/s2 to be projected onto the horizontal axes. This residual bias causes an error in the horizontal position which grows quadratically to 7.7 m after only 30 seconds [and thus to 770 m after 5 minutes, unless I'm mistaken, and 110 km after an hour]

Or here: https://liqul.github.io/blog/assets/rotation.pdf (search for "Accuracy of Velocity and Position Estimates").

[1] such as assuming that your foot has velocity zero while on the ground, which does not hold when you're in an elevator, for example, and which you can't use in a drone without some serious sensor fusion.

I had 2 tricks - really high-order filter borrowed from some paper and the assumption that a walking human doesn't gain constant acceleration. Even in an elevator you stop on a floor which resets your speed to 0.

I agree that there is no way you could extend this directly for flying, but with modern devices and things like ground-distance radar, relative airspeed indicators and so on I don't think it is beyond the realm of possibility. Plus we have detailed hightmap of the world, which, when combined with a radar should allow for terrain tracking. That makes the accuracy of sole INS much less crucial.

This also depends on the plane. For larger commercial applications, you need to have INS. You want to use a combination of sources ideally, since INS is accurate for a shorter period and GNSS/GPS for longer periods(often GPS/GNSS is only updated every second). Plus a lot of newer units have a much smaller drift. This FOG has a drift of 0.1 deg per hour, which is quite good (https://www.advancednavigation.com/solutions/spatial-fog-dua...).

Outside of military uses when GNSS can be jammed INS is pointless when GNSS is available.

INS that is accurate is very expensive to build and maintain. INS that isn’t reliant on external inputs including from a magnetic compass for calibration is even more so.

And even top-grade military INS does external checkpoints where possible (although it's probably pretty accurate in any case).

Not saying you're wrong, but. . . who actually knows what actual top-grade INS is doing these days? Those kinds of capabilities would probably be Top Secret or SCI.

The SR71 was using stars to navigate!

You need clear view of the sky and big lenses (you can always use more compact metamaterial lenses and infrared imaging for cloudy days but they are as expensive if not more than INS). Star charts also drift over the years and you will have the same problem as magnetic true north where the database needs to be updated regularly. Fine for airliners, not so great for skyhawks.

Yeah I guess that wouldn’t work during takeoff and landing. But for the most common airliners (a320 to a380 and boeing equivalent): are there clouds above their cruising altitude? By day I presume you could use the sun.

Some early jetliners had special cupola for star navigation, but it soon fell out of use because NDB, VOR, DME and other radio systems were easier and better.

I'm imagining a navigation system by trying to pattern-match the earth terrain with a downward-facing camera from the plane.

A suitably large database of satellite photos covering various conditions, day, night might work for all cases but cloudy (when the plane is above/in the clouds).

Already exists: https://en.wikipedia.org/wiki/TERCOM Since the 1950s!

Oh yeah, contour-matching. I remember this being touted with regard to U.S. cruise missiles in the 80's, 90's.

Guess radar + countour is more reliable than camera + imagery.

(I'll crawl back to the 1800's now.)

They've used that landing on Mars as well, Terrain Relative Navigation.

An automatic celestial navigation system for navigating both night and day by observation of K-band or H-band infrared light from multiple stars.

https://patents.google.com/patent/US7349803B2/en

In this celestial map, the bodies of the solar system are placed so exactly that those versed in astronomy could calculate the precession (progressively earlier occurrence) of the Pole Star for approximately the next 14,000 years. Conversely, future generations could look upon this monument and determine, if no other means were available, the exact date on which Hoover Dam was dedicated.

https://www.usbr.gov/lc/hooverdam/history/essays/artwork.htm...

That worked in the SR71's case because, flying high as it did, the sky above is black and you can see stars during the day.

They were teaching cadets at the Air Force Academy celestial navigation in the 1980s.

This guy

https://en.wikipedia.org/wiki/Guy_Murchie

taught celestial navigation to navigators flying across the Atlantic in WWII.

Exactly what I was thinking about:

https://www.thedrive.com/the-war-zone/17207/sr-71s-r2-d2-cou...

https://www.thedrive.com/the-war-zone/41287/r2-d2-spotted-on...

If something unforeseen happens an the GNSS systems go down for whatever reason we might be set up for a terrifying few months of disruption. Coronavirus was bad enough, and most of the disruptions were voluntary to some extent.

We still have VOR network and other legacy systems. Some are no longer integrated into parts of the avionics but you can still plan routes and use them to fly. Plus Primary Surveillance Radar will see you anyway.

I wonder how long it will take to get the VOR network serviceable. Many of them are no longer maintained.

some dropped off maintenance, but not all. NDB is pretty much dead outside of few rare spots, though, and not everywhere had DME even at best time. But nav points used with GPS still often are based on VOR locations.

If all the various GNSS constellations went down simultaneously we'd definitely be in the shit. Not just because we lost GNSS, but we'd also probably be fighting a war or Kessler syndrome would be in effect. Either way, losing GNSS would be just one of many problems!

Operation Black Buck [1] was done using INS [2] as GNSS was not yet available.

[1] https://en.wikipedia.org/wiki/Operation_Black_Buck [2] https://en.wikipedia.org/wiki/Delco_Carousel

Reply on news.ycombinator.com