Flying Indoors, Technical Aspects

[Dogpilot]

Flying indoors can possibly be an iffy proposition. You loose GPS signal, your essentially in attitude mode. I recently snagged a Mini 2 to add to the collection and thought this would be quite nice indoors, some prop guards and rock on. So in my first attempt, it would never get more than 5-6 satellites and considering I live in the mountains at around 2200 meters, it puts the drone in High Altitude Mode warning. It is a bit wobbly. So I thought, "hey, I have this ultra cool GPS repeater to bring my aircraft systems online in the hanger without dragging it out into the snow!" I have to have complete lock on to make some adjustments and corrections to the aircraft system software.

So, it stopped snowing today and I motored out to the hanger, dug out the door (we got just over a meter of snow). Packed up the device and hauled it home. It has to have the repeater box and the transmitter about 10-15 meters apart with the head having a clear view of the sky. I set the beastie up and Lo! It works, 14 satellites, solid strong signal. The mini was ultra happy, stable. The cat is in big trouble now. Now if I can suss out how to make it hover less than ½ meter it would be even cooler.

The gizmo is not terribly expensive, around $100 on eBay.

 

[nathankw]

But how can that work? Surely GPS reading indoors will always read the position that the repeater antenna is at - so it won't give the true position / movement of the drone.

I'd have thought that would just confuse the matter?

 

[Dogpilot]

GPS position really has nothing to do with the position of the repeater. GPS position is function of very precise clocks and orbital parameters. So the repeater is re-transmitting this information, not any kind of position data. Each satellite gives you a “line of position” which as you add more you get a 3D position with great accuracy.

Just happen to live next to the Naval Observatory and know the director, we flew the same plane in the Navy. It injects all the orbital corrections up to the constellation. If you are really curious, get a copy of “American Practical Navigator” and read how it works. Also good if you have a sleep disorder. I had to take a nauseatingly long class at UCLA about it and celestial navigation. it actually proved useful, as I sailed from San Diego to Sydney and thankfully didn’t miss Australia as A Midshipman.

 

[nathankw]

I do have a reasonable understanding of how GPS works but I can't work out how you can use a repeater and still get an accurate location for the drone (and not the repeater).

Suppose (for the sake of argument), I'm standing next to my repeater and we're exactly 1000 miles from a satellite to the west and 1000 miles from a satellite to the east.

The signals therefore take the same time to reach me from either satellite.

Now I move 1 mile east.
If I were getting the signals direct from the satellites the east signal would reach me sooner since I'm now closer to that one.

But if the signals are coming via the repeater - both signals have traveled the same distance (1000 miles) to reach the repeater then the same distance (1 mile) to reach me.

So I would have thought I'd receive both signals at the same time and would appear to the GPS unit to be at the same position as the repeater antenna.

Am happy to stand corrected but I just can't see how that can work.

Of course there are indoor positioning systems using multiple sources to triangulate position in the same way that GPS does. But I presume that's not what you're talking about?

 

[DougMcC]

I’m going to watch this and learn something. Great discussion.

 

[Rich QR]

For any who want a brief and fairly nontechnical explanation of how GPS works:

Satellite Navigation - GPS - How It Works

www.faa.gov

As nathankw explains, it works by precisely measuring the time it takes a signal to travel (at the speed of light) from several satellites to the receiver. By measuring the differences in times of arrival from several satellites, the receiver can work out the differences in distances between the GPS antenna and each of the various satellites. Combine this with knowing the precise positions and velocities of the satellites as they're orbiting the Earth, and the computer can figure out the position of the receiving antenna on Earth.

In addition to transmitting the time very precisely, each satellite also transmits some digital data, rather slowly. The digital data allows the GPS receiver to know the positions and velocities of the satellites. One piece of data that it transmits is called the almanac. The almanac tells the orbital parameters of every one of the satellites, well enough to find them in the sky, but not well enough to use them for precise navigation. Another piece of data is called the ephemeris. The ephemeris gives very precise orbital parameters of the satellite, precise enough to use for navigation. Each satellite broadcasts the ephemeris only for itself, but broadcasts the almanac for all of the satellites. The ephemeris is only accurate for a short period of time, a few hours. The almanac, on the other hand, is good for 90 days.

The almanac takes 12.5 minutes to transmit, and the ephemeris data for a satellite takes 30 seconds to transmit. These transmission times account for the time it takes a GPS receiver to "lock on" to satellites and use them for navigation. If you shut off your drone to change batteries, when you restart, the GPS can receive multiple satellites almost immediately, because it already has valid almanac and ephemeris data. The first flight of the day, the GPS needs to "listen" to the satellites for around 30 seconds, as it downloads each satellites ephemeris data. If you left your GPS off for a year, it would take even longer to get a good lock on the satellites, because the old stored almanac data would be stale.

I suspect the repeater is simply amplifying and repeating everything it hears. In that case, any receiver hearing the repeated signal is going to be fooled into reporting its position fixed at location of the repeater's receiving antenna, since the differences in arrival times at the antenna are the only things that the GPS receiver has to go on. This would allow a GPS on an airplane inside a hangar to receive a position, and also allow that receiver to get the almanac and ephemeris data, so it's immediately ready to get an accurate position as soon as it's rolled out under the open sky.

 

[nathankw]

"I suspect the repeater is simply amplifying and repeating everything it hears. In that case, any receiver hearing the repeated signal is going to be fooled into reporting its position fixed at location of the repeater's receiving antenna, since the differences in arrival times at the antenna are the only things that the GPS receiver has to go on"

Exactly what I would have thought. That said - I've no actual experience of using a GPS repeater so happy to stand corrected by someone with direct experience of how they work.

Back to to the original point of the thread about flying doors. I've done a bit of flying inside with my Mini 2 and actually found it very stable. I've flown with no GPS but in that case the drone can use its visual sensors and those work very well if the environment is well lit.

One possible explanation for what the OP has found - I suspect that poor GPS is worse than no GPS.
The OP says he was originally getting 5-6 satellites. Indoors the GPS would be prone to reflections and other problems. So the drone may have been receiving unstable signals causing it to jump about.

Apparently you can get problems inside when the drone suddenly gets a signal through a window causing it to jump.

With the repeater the drone is getting a bogus signal (because the position won't change when the drone moves). So the drone effectively ignores GPS altogether and uses visual feedback - hence is more stable.

So the repeater might actually be working as a GPS blocker!
Just a theory.

 

[Dogpilot]

What you are missing is a AGC function on your GPS receiver. It will lock onto the strongest signal. GPS signals are very weak, ergo why trees and roofs block out most of the signal. So when the repeater is online it is naturally the strongest signal and your receiver ignores the other signals. The total distance on the repeater setup is only 50 m. The transmitter and receiver must be at least 15 m apart or the receiver has something blocking its own signal, like lets say a metal hanger. Otherwise it is receiving what it transmits and you get garbage. It is very accurate when running, no difference than regular sky view. It has to be so I can adjust settings on the Synthetic Vision System on my airplane. The autopilot and TAWS will not function until it gets WAAS signal or differential nav as well. (I actually have my Engineer's license, A&P, for aircraft, I wouldn't fly it if I can't fix it).

So the drone's GPS sees accurate position, since while the receiver is offset, lets say the full 50 m of the cable. It still see the time difference the same since the radio signal is received by the receiver and transmitted the additional 50 m making the total signal distance time offset the same. No it will not be as accurate as total sky view and I will not let it do positioning to drill into my eyeball. It is good enough for accurate flying. You accomplish similar corrections when you use a psudolite over a benchmark to give high precision to a survey rover. It transmits a very accurate position, surveyed to high precision, which is used to correct tiny differences in the TD from the regular sats. WAAS and differential provide much the same to us in the air for precision approaches that have LPV (GPS glide slope). You can up that precision by adding ground psudolites for higher precision on approach which will eventually supplant our ILS radio based approaches.

Of course this all goes down the tube when the military plays EW with the GPS system. Living in the SW, near Area 51 and all the military play areas. They test this tech from time to time, actually an annoying number of times recently. They NOTAM it but it still makes your entire NAV system go into DR mode and you have to go back to green needles. It will also make your drone fairly useless and put you in att mode.

 

[nathankw]

I think maybe we're talking at cross purposes. Of course, if the drone can still receive signals direct from the satellites then you will get an accurate position. That's case with systems like WAAS and differential GPS - the mobile unit is still receiving signals directly from the GPS satellites but also receiving additional data from the fixed unit that enables it to remove errors and improve accuracy.

I was talking about the situation where a drone cannot receive any direct signals (eg because it's indoors) - but is receiving GPS signals relayed from a fixed point (the repeater). Obviously in this situation there's no way for the drone to accurately know it's position relative to the repeater. As Rich says, it can still receive almanac and ephemeris data.

It's interesting to know what the drone would do in that situation when its GPS location would appear to be fixed despite its other sensors detecting movement.

 

[Brockrock]

It's interesting to know what the drone would do in that situation when its GPS location would appear to be fixed despite its other sensors detecting movement.

And for this, kind Sir, we need the Op to post a video, complete with setup, startup, and flight telemetry overlays..

 

[NLOL]

Hmmm, seems to me that, the GPS position inforation received by the drone is incorrect - but so what. So what if it's actual position on a map is not the same as you know where your drone is. IN fact there are more than the current WGS84 reference systems available on GPS's. What counts for flying in some local area is simply that there is a reference. The dron knows where it started, it knows if it is drifting from that location, etc. The repeater is to give the drone position reference information during local visual flight. So your map- really, just the background on your screen, may be wrong but all the flight track and position data relevant the the home point is still good.

 

[IslandBarnDave]

For any who want a brief and fairly nontechnical explanation of how GPS works:

Satellite Navigation - GPS - How It Works

www.faa.gov

As nathankw explains, it works by precisely measuring the time it takes a signal to travel (at the speed of light) from several satellites to the receiver. By measuring the differences in times of arrival from several satellites, the receiver can work out the differences in distances between the GPS antenna and each of the various satellites. Combine this with knowing the precise positions and velocities of the satellites as they're orbiting the Earth, and the computer can figure out the position of the receiving antenna on Earth.

In addition to transmitting the time very precisely, each satellite also transmits some digital data, rather slowly. The digital data allows the GPS receiver to know the positions and velocities of the satellites. One piece of data that it transmits is called the almanac. The almanac tells the orbital parameters of every one of the satellites, well enough to find them in the sky, but not well enough to use them for precise navigation. Another piece of data is called the ephemeris. The ephemeris gives very precise orbital parameters of the satellite, precise enough to use for navigation. Each satellite broadcasts the ephemeris only for itself, but broadcasts the almanac for all of the satellites. The ephemeris is only accurate for a short period of time, a few hours. The almanac, on the other hand, is good for 90 days.

The almanac takes 12.5 minutes to transmit, and the ephemeris data for a satellite takes 30 seconds to transmit. These transmission times account for the time it takes a GPS receiver to "lock on" to satellites and use them for navigation. If you shut off your drone to change batteries, when you restart, the GPS can receive multiple satellites almost immediately, because it already has valid almanac and ephemeris data. The first flight of the day, the GPS needs to "listen" to the satellites for around 30 seconds, as it downloads each satellites ephemeris data. If you left your GPS off for a year, it would take even longer to get a good lock on the satellites, because the old stored almanac data would be stale.

I suspect the repeater is simply amplifying and repeating everything it hears. In that case, any receiver hearing the repeated signal is going to be fooled into reporting its position fixed at location of the repeater's receiving antenna, since the differences in arrival times at the antenna are the only things that the GPS receiver has to go on. This would allow a GPS on an airplane inside a hangar to receive a position, and also allow that receiver to get the almanac and ephemeris data, so it's immediately ready to get an accurate position as soon as it's rolled out under the open sky.

Really interesting. Thanks for taking the time to explain this.

 

[Dogpilot]

As mentioned by another poster, you need to go and then understand how the system works. It will work in a completely shielded environment, like my hanger as long as it receives the repeated signal. The hanger is a faraday cage, a metal building, signals do not enter it and it can see no direct GPS signals. I would love to fly the drone in my hanger, but I am on a controlled airport and no flying permitted without a waiver, so I won't be able to demo there. Nonetheless, the signals work the same. They are simply repeated. The GPS computer does the same math with the times as it does with a direct signal. Nothing changes. No direct is required or received. I follow procedures set out by the Garmin's (maker of a metric ton of things GPS) line maintenance manual. So I don't think those wizards are deceiving themselves. The position does not jump when I switch on. DJI's little program on your phone also gets position from cell towers, but that information is not shared with the drone, but updates the map's home point.

 

[NLOL]

"I suspect the repeater is simply amplifying and repeating everything it hears. In that case, any receiver hearing the repeated signal is going to be fooled into reporting its position fixed at location of the repeater's receiving antenna," is not correct. To esily understand this imagine two repeaters side by side but with different delays in their signal processing. Clearly the drone would not think both signals are coming from the same location.

 

[nathankw]

Since my knowledge is limited I thought it would be good to get some clarification from someone who really knows about repeaters, so I emailed www.gps-repeaters.com who sell and install repeated systems.

Here's their response which I think clears things up:

>

"I get quite a few enquiries about using a GPS repeater to operate a UAV indoors.

I'm afraid it hasn't worked to date; the reason is that the system only relays the outdoor GPS signal to the inside of a building.

A single repeater unit, mounted 10 meters up in the roof, facing down, can easily provide signal over a large area (approx. 40-50 metres diameter).

Thing is, the signal you get within that coverage area is always the same - that is to say wherever you are within the coverage area you will only pick up the positional coordinates of the outdoor antenna.

This doesn't provide the means to navigate inside the building as the signal is the same wherever you are.

A drone can get confused, because when it moves it is still receiving the same GPS coordinates as if it was standing still.

You can use a repeater for functional testing but usually not for indoor flight/navigation.

Best regards

Phil"

 

[NLOL]

What then does the 'repeater' do? It is not just boosting signal strength - If I got a stronger signal - that is if the signal reached my drone - the drone would fly fine. Your 'repeater' is not just repeating/boosting the signal and adding the time delay of its processing. It is something other than a repeater.

 

[nathankw]

What then does the 'repeater' do? It is not just boosting signal strength - If I got a stronger signal - that is if the signal reached my drone - the drone would fly fine. Your 'repeater' is not just repeating/boosting the signal and adding the time delay of its processing. It is something other than a repeater.

I don't quite understand your question.

I think Phil explains it very well.

The repeater is boosting the signal strength (or rather rebroadcasting it in an area where you wouldn't otherwise be able to receive GPS signals). However, because it's boosting the signals as they're received at the repeater - the drone will always think it's at that fixed position. So it won't get the usual position feedback that allows it to stay stable, not drift, perform maneuvers etc.

Such a signal is still useful for other purposes - for instance testing equipment.
Emergency services use it them so when they leave the building their GPS units are already locked on and work straight away.

 

[Rich QR]

What then does the 'repeater' do? It is not just boosting signal strength - If I got a stronger signal - that is if the signal reached my drone - the drone would fly fine. Your 'repeater' is not just repeating/boosting the signal and adding the time delay of its processing. It is something other than a repeater.

No, the problem is that a repeater repeats the signals exactly as they were received by the repeater's antenna.

The only way a GPS has of figuring out its location is by looking at the differences in timing of signals from the various satellites. Those differences will be preserved by the repeater, so the GPS receiver will think it's at the repeater's location.

For a simplified hypothetical explanation, pretend there are four GPS satellites: one due north, one due east, one due south, and the final one due west. Pretend they're not moving (ignore the fact that that would never happen with real orbiting satellites). A receiver happens to hear all four timing signals pinging at exactly the same time, perfectly synchronized to each other. So it knows that each of the satellites is exactly the same distance away, and it knows the GPS receiver's location is in the center point between all four satellites.

(Side note: the GPS receiver has a clock inside, but the clock is not able to maintain accuracy to within a few nanoseconds. The satellites have very expensive, precise atomic clocks that maintain superb accuracy. So the GPS receiver doesn't use the absolute time of a signal's arrival, it looks only at the differences between the arrival times of signals from the different satellites.)

Light travels at close to one foot per nanosecond. Let's move the receiver ten feet to the east. It will hear the eastern satellite's time ping first, because that's closest. Ten nanoseconds later, it will hear the ping from the north and south satellites simultaneously, and ten nanoseconds after that, it will hear the ping from the western satellite.

By the timing of the signals, the receiver knows that the distance to the eastern satellite is ten feet less than the distance to the north and south satellites, and it knows that the distance to the western satellite is ten feet greater than the distance to the north and south satellites.

If the signal had gone through a stationary repeater, the repeater isn't going to speed up or slow down the signal from any individual satellite relative to the others. If the repeater had received all the signals simultaneously, it will repeat all of them simultaneously. The receiver will receive them with some delay, depending on the distance between the repeater and the GPS receiver, but the absolute delay doesn't matter; what matters is the difference between the arrival times of the signals from various satellites.

The real situation is more complicated than this hypothetical example, because the satellites are at various angles, in 3-d space, and moving rapidly. But the point is, the GPS only knows its position by looking at the differences in timing signals from the various satellites. If the timing signal from satellite "A" starts "gaining" on the signal from satellite "B", it knows that it's getting relatively closer to satellite "A" and/or farther from "B". Since a repeater is incapable of speeding up or slowing down the relative timing between the various satellite signals, the GPS gets the relative timings as they were received by the repeater's antenna, so that's the position it computes.

 

[NLOL]

Well, yes - that is exactly what I think a repeater does - and thus, ah ha, the time differences are 'frozen' at the position of the repeater and it does not work to fly the drone. Got it.

 

[Dogpilot]

Well I was going to go back to my hanger and snag one of my very precise GPS, one that show movement in very small increments, a meter or so. The WX gods disagreed and blanketed us with more snow. So been on the end of a snow shovel a lot today. I'll check this out when I get a chance to drive again. Heck even the receiver head is buried, so nothing is working.