a radio based navigation system

[Stratennotblitz]

so my idea is to use 3 axis accelerometers and star trackers to create a homemade navigational system, it would use the star trackers at night to give the user a "position 0" and the accelerometers, as the user walks would be able to pick up rotation(you could use a compass as backup), speed (thus position relative to position 0) etc, you could overlay that info on a pre-existing map and make some sort of gps system homemade, 

a version 2.0 would use radio signals with broadcasting towers and their exact coordinates to triangulate a more precise position, per example you have a rotating antenna that picks up and bounces signals to other radios and measures the intensity of the signal (you could use a barometer and altimeter to correct atmospheric distortion etc)

if you have a stronger signal from 3 radio sources relative to you position 0 and you, you can use that to correct your actual position more correctly by figuring out the different signal intensities between them (distance) and comparing it to your distance from the position 0 and their distance from it.

(https://arstechnica.com/gadgets/2017/08/radio-navigation-set-to-make-global-return-as-gps-backup-because-cyber/)

 

essentially this image but with modern additional equipment

[cubinator]

Like a ground-based GPS backup? The towers are in known positions on the Earth, and so can be used as reference. 

Star tracking doesn't really need this, though, all you need is an accurate clock and a view of the sky. Even works during the day. "For what latitudes and longitudes is the Sun supposed to be as high as I measure it right now?" Repeat over the course of a few hours and you'll have your position.

[JoeSchmuckatelli]

Far easier is to write an app that leverages the radio receiver on your phone and measures the relative distance to known towers. 

 

By the way - I officially copyright this idea 

[kerbiloid]

You should do this ~12 terrestrial years earlier.

Edited August 15, 2021 by kerbiloid

[Shpaget]

VOR, NDB, DME, TACAN

That should get you started.

[cubinator]

10 hours ago, JoeSchmuckatelli said:

Far easier is to write an app that leverages the radio receiver on your phone and measures the relative distance to known towers. 

Isn't that how "network location" works to get your approximate location? Don't phones already do this (or something similar)?

[JoeSchmuckatelli]

34 minutes ago, cubinator said:

Isn't that how "network location" works to get your approximate location? Don't phones already do this (or something similar)?

Yes - but it's the reverse.  The company (and government) can get an approximate lock on you... But AFAIK there are no apps giving you a relative position b/c GPS 

 

Edit - if you think about it, the app would need a list of GPS locations for each tower to compare against.  With that it's pretty simple triangulation. 

Edited August 15, 2021 by JoeSchmuckatelli

[Shpaget]

47 minutes ago, JoeSchmuckatelli said:

But AFAIK there are no apps giving you a relative position b/c GPS 

I distinctly remember some of my pre-smart phones having the ability to show my approximate location.

No apps involved, just a network/phone capability.

[kerbiloid]

Old phones are equipped with a built-in function of their location relative to the closest audio dynamics.

The dynamics start buzzing when the phone is near.

[wumpus]

2 hours ago, Shpaget said:

I distinctly remember some of my pre-smart phones having the ability to show my approximate location.

No apps involved, just a network/phone capability.

There was also a lot of politicing in upgrading 911 (or whatever you local emergency phone number is) to get the GPS information from smartphones.  There was a lot of push to require dumbphones (mobile, not landline) to include this as well but I'm not sure what they ended up with.  I think that they use tower triangulation and that is good enough.

If your phone can quickly get GPS coordinates, it either never turns off (great for emergencies,  lousy for battery life) or uses tower triangulation to determine a good approximation and lets the GPS start computing from there.

[magnemoe]

6 hours ago, cubinator said:

Isn't that how "network location" works to get your approximate location? Don't phones already do this (or something similar)?

The phone service know and need to know because they need to switch you. Many apps uses this, its not 1 meter accurate but good enough for most uses. 
Cell tower lock has also been uses in lots of court cases. Stupid criminals claim they was far from the crime scene but they was close all the time. An smart criminal could give his phone to an friend and tell him to do stuff he would do but luckily most criminals are stupid. 

[Meecrob]

21 hours ago, Stratennotblitz said:

so my idea is to use 3 axis accelerometers and star trackers to create a homemade navigational system, it would use the star trackers at night to give the user a "position 0" and the accelerometers, as the user walks would be able to pick up rotation(you could use a compass as backup), speed (thus position relative to position 0) etc, you could overlay that info on a pre-existing map and make some sort of gps system homemade, 

a version 2.0 would use radio signals with broadcasting towers and their exact coordinates to triangulate a more precise position, per example you have a rotating antenna that picks up and bounces signals to other radios and measures the intensity of the signal (you could use a barometer and altimeter to correct atmospheric distortion etc)

if you have a stronger signal from 3 radio sources relative to you position 0 and you, you can use that to correct your actual position more correctly by figuring out the different signal intensities between them (distance) and comparing it to your distance from the position 0 and their distance from it.

(https://arstechnica.com/gadgets/2017/08/radio-navigation-set-to-make-global-return-as-gps-backup-because-cyber/)

Your first idea is essentially an inertial navigation system, which is far superior to radio beacons/broadcasting towers. Commercial aviation ditched radio for INS decades ago.

[K^2]

On 8/15/2021 at 4:04 PM, Meecrob said:

Your first idea is essentially an inertial navigation system, which is far superior to radio beacons/broadcasting towers. Commercial aviation ditched radio for INS decades ago.

Commercial aviation "ditched" radio for INS because flying over oceans made using radio towers for navigation a little difficult. It took decades of improvements before INS was good enough to get you to a specific location, rather than just a vague area, because it's actually a very hard problem. Typical off-the-shelf accelerometers aren't remotely good enough for such work. And even with commercial INS systems, your terminal approach is still by-radio. The INS is still only meant to get you to a radio system that's going to guide you to the landing strip. Not to mention that GPS has become the primary system now, with INS being effectively a backup in case GPS goes down.

If you shop for high precision accelerometers for UAVs etc, you'll see VRW on the order of 0.05mg/√Hz. This adds up to about 3cm/s per root hour. After a 9 hour flight, your position error is going to be about 1.6km. Which is actually quite amazing, but you don't want to be a mile off course if you are trying to find an airport in low visibility. You will have to rely on radio navigation. And this is with top shelf component that you have to sample and integrate at nearly 1kHz to get that level of precision, and we haven't even talked about orientation of the INS. Because if your axes are misaligned even the tiniest bit, the component of the gravity onto your XY plane will put you way, way off course. Not to mention the fact that gravity itself isn't uniform around the world, so you are going to get a drift in the Z direction as well.

You can improve on that by combining input from several accelerometers and optical gyros and running output through an optimal filter. That's effectively what a modern INS does. But if you are anywhere over land, it's far easier and far cheaper to use existing radio stations for precise positioning.

Or better yet, just buy a cheap GPS chip and read from it using an Arduino or something. Unless you are explicitly building something for safety and redundancy as a backup, GPS is by far your best option.

 

Edit: Only marginally related, but I it's just too awesome not to mention: Pulsar-based Navigation. It still only puts you within a few km, but it will work anywhere on Earth, and on a flight to the Moon, and on a voyage to any other part of the Solar System, and on an interstellar trip to basically any star you can see in the sky with an unaided eye... And it will still be good to within a few km after you've traveled lightyears, because you're just counting the pulses and comparing it to frequencies observed from reference location.

Edited August 18, 2021 by K^2

[mikegarrison]

I would like more information from the OP about the goal of this work.

Is this simply for the experience of doing it? Are you trying to do it with real hardware or just figure out how it is done? What kind of error rates are acceptable?

In the real world, GPS is the best and also the cheapest option. Airplane-quality inertial navigation systems are expensive and these days almost always involve laser ring gyros (which are not actually gyros but have a similar functionality of detecting rotational rates).

This guy actually did a short video about the subject which seems to me like a good intro, though of course doesn't really tell you how to build one.

 

[magnemoe]

I was on an transatlantic from US to Europe around 6 years ago who had to abort because one of the navigation systems failed and they was not allowed to cross the Atlantic on one. 
They flew to the coast but then had to turn back, probably trying to hot patch it and knowing they would loose the fuel and needed another plane if not. 
Long distance flights can not land with the fuel they take off with. 
Another flight who was around 2000 km inside Norway just turned back and landed then they could not retract the nose wheel. 
Obviously we passengers was more interested if they could get it out again for landing.
 

[mikegarrison]

1 hour ago, magnemoe said:

Long distance flights can not land with the fuel they take off with.

Well ... sort of. If you land overweight there is a risk of structural damage and many inspections that have to be done. But if you really, really need to land, then you land.

[Stratennotblitz]

15 hours ago, K^2 said:

Commercial aviation "ditched" radio for INS because flying over oceans made using radio towers for navigation a little difficult. It took decades of improvements before INS was good enough to get you to a specific location, rather than just a vague area, because it's actually a very hard problem. Typical off-the-shelf accelerometers aren't remotely good enough for such work. And even with commercial INS systems, your terminal approach is still by-radio. The INS is still only meant to get you to a radio system that's going to guide you to the landing strip. Not to mention that GPS has become the primary system now, with INS being effectively a backup in case GPS goes down.

If you shop for high precision accelerometers for UAVs etc, you'll see VRW on the order of 0.05mg/√Hz. This adds up to about 3cm/s per root hour. After a 9 hour flight, your position error is going to be about 1.6km. Which is actually quite amazing, but you don't want to be a mile off course if you are trying to find an airport in low visibility. You will have to rely on radio navigation. And this is with top shelf component that you have to sample and integrate at nearly 1kHz to get that level of precision, and we haven't even talked about orientation of the INS. Because if your axes are misaligned even the tiniest bit, the component of the gravity onto your XY plane will put you way, way off course. Not to mention the fact that gravity itself isn't uniform around the world, so you are going to get a drift in the Z direction as well.

You can improve on that by combining input from several accelerometers and optical gyros and running output through an optimal filter. That's effectively what a modern INS does. But if you are anywhere over land, it's far easier and far cheaper to use existing radio stations for precise positioning.

Or better yet, just buy a cheap GPS chip and read from it using an Arduino or something. Unless you are explicitly building something for safety and redundancy as a backup, GPS is by far your best option.

 

Edit: Only marginally related, but I it's just too awesome not to mention: Pulsar-based Navigation. It still only puts you within a few km, but it will work anywhere on Earth, and on a flight to the Moon, and on a voyage to any other part of the Solar System, and on an interstellar trip to basically any star you can see in the sky with an unaided eye... And it will still be good to within a few km after you've traveled lightyears, because you're just counting the pulses and comparing it to frequencies observed from reference location.

yup

 

15 hours ago, K^2 said:

Commercial aviation "ditched" radio for INS because flying over oceans made using radio towers for navigation a little difficult. It took decades of improvements before INS was good enough to get you to a specific location, rather than just a vague area, because it's actually a very hard problem. Typical off-the-shelf accelerometers aren't remotely good enough for such work. And even with commercial INS systems, your terminal approach is still by-radio. The INS is still only meant to get you to a radio system that's going to guide you to the landing strip. Not to mention that GPS has become the primary system now, with INS being effectively a backup in case GPS goes down.

If you shop for high precision accelerometers for UAVs etc, you'll see VRW on the order of 0.05mg/√Hz. This adds up to about 3cm/s per root hour. After a 9 hour flight, your position error is going to be about 1.6km. Which is actually quite amazing, but you don't want to be a mile off course if you are trying to find an airport in low visibility. You will have to rely on radio navigation. And this is with top shelf component that you have to sample and integrate at nearly 1kHz to get that level of precision, and we haven't even talked about orientation of the INS. Because if your axes are misaligned even the tiniest bit, the component of the gravity onto your XY plane will put you way, way off course. Not to mention the fact that gravity itself isn't uniform around the world, so you are going to get a drift in the Z direction as well.

You can improve on that by combining input from several accelerometers and optical gyros and running output through an optimal filter. That's effectively what a modern INS does. But if you are anywhere over land, it's far easier and far cheaper to use existing radio stations for precise positioning.

Or better yet, just buy a cheap GPS chip and read from it using an Arduino or something. Unless you are explicitly building something for safety and redundancy as a backup, GPS is by far your best option.

 

Edit: Only marginally related, but I it's just too awesome not to mention: Pulsar-based Navigation. It still only puts you within a few km, but it will work anywhere on Earth, and on a flight to the Moon, and on a voyage to any other part of the Solar System, and on an interstellar trip to basically any star you can see in the sky with an unaided eye... And it will still be good to within a few km after you've traveled lightyears, because you're just counting the pulses and comparing it to frequencies observed from reference location.

yup! had this idea from it, watched a documentary about redundant nav systems

On 8/14/2021 at 10:51 PM, cubinator said:

Like a ground-based GPS backup? The towers are in known positions on the Earth, and so can be used as reference. 

Star tracking doesn't really need this, though, all you need is an accurate clock and a view of the sky. Even works during the day. "For what latitudes and longitudes is the Sun supposed to be as high as I measure it right now?" Repeat over the course of a few hours and you'll have your position.

damn, I didn't know that, I wonder if someone could pull full nav off with good enough software and day/night trackers

8 hours ago, mikegarrison said:

I would like more information from the OP about the goal of this work.

Is this simply for the experience of doing it? Are you trying to do it with real hardware or just figure out how it is done? What kind of error rates are acceptable?

In the real world, GPS is the best and also the cheapest option. Airplane-quality inertial navigation systems are expensive and these days almost always involve laser ring gyros (which are not actually gyros but have a similar functionality of detecting rotational rates).

This guy actually did a short video about the subject which seems to me like a good intro, though of course doesn't really tell you how to build one.

 

thanks for the answer, I just wanna experiment with navigation systems for developing countries

On 8/15/2021 at 7:04 PM, Meecrob said:

Your first idea is essentially an inertial navigation system, which is far superior to radio beacons/broadcasting towers. Commercial aviation ditched radio for INS decades ago.

yes

On 8/14/2021 at 11:23 PM, JoeSchmuckatelli said:

Far easier is to write an app that leverages the radio receiver on your phone and measures the relative distance to known towers. 

 

By the way - I officially copyright this idea 

already patented (kinda) https://patents.google.com/patent/US7945271B1/en

[cubinator]

36 minutes ago, Stratennotblitz said:

damn, I didn't know that, I wonder if someone could pull full nav off with good enough software and day/night trackers

Considering you can do it with a big book of tables and holding your hand up to the sky, I'd say all it takes is a sensor that can figure out which star it's looking at. I'd start it off with the sun for an easy estimate. Pretty sure spacecraft/people who drive spacecraft do this with the starfield. 

[mikegarrison]

Yeah, celestial navigation is among the oldest ways of navigating.

You only need an accurate clock for longitude. Latitude you can get without knowing exactly what time it is.

[YNM]

On 8/15/2021 at 11:28 AM, Shpaget said:

VOR, NDB, DME, TACAN

LORAN too ?

 

But actually even back in the age of CDMA and GSM mobile signals (so 2G) it was possible to roughly locate where you are. I still remember during our yearly mudik trip those nokia phones from the early 2000s would tell you which city are you in/passing through (I believe it was down to which district but not sure).

These days I presume if you designate every 4G and 5G transmitter with a GPS coordinate and/or equip it with a GPS base receiver (the stuff that calculates their position for years on end stationarily to achieve extremely high precision) you can easily lose the GPS satellite and just rely on that maybe for a decade or two.

 

But GPS/GNSS is always going to be far superior and cheaper. There's a reason GNSS approach minimums are equal to ILS Cat I approach minimums.

Edited August 19, 2021 by YNM

[Stratennotblitz]

On 8/18/2021 at 10:51 PM, cubinator said:

Considering you can do it with a big book of tables and holding your hand up to the sky, I'd say all it takes is a sensor that can figure out which star it's looking at. I'd start it off with the sun for an easy estimate. Pretty sure spacecraft/people who drive spacecraft do this with the starfield. 

yup, they're called star trackers : 

 

On 8/19/2021 at 2:27 AM, YNM said:

LORAN too ?

 

But actually even back in the age of CDMA and GSM mobile signals (so 2G) it was possible to roughly locate where you are. I still remember during our yearly mudik trip those nokia phones from the early 2000s would tell you which city are you in/passing through (I believe it was down to which district but not sure).

These days I presume if you designate every 4G and 5G transmitter with a GPS coordinate and/or equip it with a GPS base receiver (the stuff that calculates their position for years on end stationarily to achieve extremely high precision) you can easily lose the GPS satellite and just rely on that maybe for a decade or two.

 

But GPS/GNSS is always going to be far superior and cheaper. There's a reason GNSS approach minimums are equal to ILS Cat I approach minimums.

I doubt gps/gnss are going to be cheaper in the future, we could use atmospheric satellites in the future, just like Facebook did with their flying internet planes to beam down positional data, it's far cheaper and reliable, since there are thousands of those (possible) compared to only a few gps systems, not to mention using radio antennas with better inertial measurement units to communicate, the us military already wants a better gps for national security reasons so that industry might develop in the future, the optimal system would be a mix of all the previous nav methods

[YNM]

16 hours ago, Stratennotblitz said:

we could use atmospheric satellites in the future

If it's not fixed to the ground, or it can't be expressed in TLE that doesn't need to be changed every so often, I don't think you can use it to calculate your position wrt the Earth, unless you only want relative positions to the beacon or something.

Also by "cheap" I mean to the end user. Sure yes some government's tax money had to pay for the satellites and maintaining them, but even if it never sees direct financial returns, much like the idea of weather reconnaisance satellite/stations - or emergency and disaster relief - it's just one of those things you don't do and expect any sort of financial incentives back. Though arguably it's still better than most other alternatives (like with ground stations who owns the land becomes important).

Edited August 22, 2021 by YNM

[Stratennotblitz]

On 8/22/2021 at 12:57 AM, YNM said:

If it's not fixed to the ground, or it can't be expressed in TLE that doesn't need to be changed every so often, I don't think you can use it to calculate your position wrt the Earth, unless you only want relative positions to the beacon or something.

Also by "cheap" I mean to the end user. Sure yes some government's tax money had to pay for the satellites and maintaining them, but even if it never sees direct financial returns, much like the idea of weather reconnaisance satellite/stations - or emergency and disaster relief - it's just one of those things you don't do and expect any sort of financial incentives back. Though arguably it's still better than most other alternatives (like with ground stations who owns the land becomes important).

gps satellites move, in fact they move faster than atmospheric electric planes for navigation, about 2 times circling the earth in a day compared to more than 3 days for electric planes, meaning they can stay in a specific area longer, the only issues would be atmospheric turbulences, but that could be fixed by using topographic maps to adjust craft error 

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

[YNM]

2 hours ago, Stratennotblitz said:

gps satellites move, in fact they move faster than atmospheric electric planes for navigation, about 2 times circling the earth in a day compared to more than 3 days for electric planes, meaning they can stay in a specific area longer, the only issues would be atmospheric turbulences, but that could be fixed by using topographic maps to adjust craft error 

Yeah but you can very easily describe their trajectory using a geodesic, which you can't do for a flying or floating thing in the air. That's why TLEs have been around since ages ago.

2 hours ago, Stratennotblitz said:

the only issues would be atmospheric turbulences, that could be fixed by using topographic maps to adjust craft error 

And what are you basing that maps on ? Ground beacons measured with chains and triangulation ? It's better to just use ground beacons then...

GPS/GNSS literally eliminates all of this, look up the recent proposed changes on geodectic systems (at the very least gravitational models), it's all based on the satellites' geodesics.

Edited August 26, 2021 by YNM

[magnemoe]

On 8/19/2021 at 4:51 AM, cubinator said:

Considering you can do it with a big book of tables and holding your hand up to the sky, I'd say all it takes is a sensor that can figure out which star it's looking at. I'd start it off with the sun for an easy estimate. Pretty sure spacecraft/people who drive spacecraft do this with the starfield. 

Did not the pluto nuclear ramjet plan to use star tracking for navigation? It works on earth as its basically old style navigation, not sure how accurate it is. They gave it up on later cruise missiles who used ground features to find its position, or did the 1980s cruise missiles also use it?