Android Question Time of Arrival GPS alternative

[solfinker]

Hello.
Is there a way to determine the elapsed time between the reception of positioning signals coming from different base stations?
Is there a time of arrival (ToA) measurement that calculates the time lapse between the transmission and the reception of a particular signal?
That would give us a way to determine the distance between the transmitter and the receiver.
The precision of cell ID positioning can be improved if the base station takes into account the timing advance (TADV or TA) measurement. Since GSM is a
TDMA network, it requires to know very accurately at what time stamp the signal from a particular UE (user equipment) will be received.
The TADV measurement is used by the network to schedule the UE uplink signal to arrive at a particular time slot. The TADV value is the time that an uplink signal coming from the UE needs to reach the base station. Hence, the TADV is a ToA (time of arrival) measurement and it can be useful to locate the UE within a circumference centered at the base station position.

Thank you.

 

[agraham]

I don't think you mean GPS

 

[zed]

TDMA (Time-Division Multiple Access) allows several people to have telephone conversations at the same time.
It is a type of digital modulation. Nothing to do with GPS

 

[agraham]

Yes, I know about GSM and TDMA. The question is why is GPS in the thread title?

 

[solfinker]

In reference to Raw GNSS Measurements provided by Android framework and enabled by the GPS library and the idea of using ToA for improving the real position of the user terminal made me believe that adding GPS to the title would be a good idea for fining the tune a little more.
By your comments, I can see I have failed. I guess I should have used something like LOCATION SENSORS.

The thing is, is there a way to determine the elapsed time between the reception of positioning signals coming from different base stations?

 

[agraham]

Not as far as I know but I haven't delved too deeply into the telephony APIs.

 

[BlueVision]

I don't really see the point of such an application. It might be possible with the corresponding functions that a telephone provider has available, as far as GSM is concerned. He could use them to optimise his network by comparing several transmitting stations with each other. For an end user, this possibility does not exist at all.

Counter question: How do you want to determine the distance to a GSM transmission point if you do not know the exact location of that point?

 

[solfinker]

We do know the location of the antennas in our area.
I am just trying to improve the position of the mobile phones given by GPS - oh, the word! - with a simple algorithm.
Thank you for your help.

 

[BlueVision]

Hmm...
Understood, please allow me one last hint in the direction of physics:

Ultimately, both mobile phone waves and satellite signals propagate at the speed of light. In the case of GPS, this can be quite time-critical (signal reflections), although the signal sources (satellites) are very far away for triangulation, which leads to significant differences in the propagation time of the signals. I fear that the propagation times of the signals for GSM will hardly show any differences due to the short distance between the mobile masts. This is related to the measurement accuracy of commercially available mobile radio devices and their processing speed with a programme.

From this point of view, I do not believe in a practicable improvement of a GPS position by measuring the propagation time of GSM signals.

 

[Computersmith64]

We do know the location of the antennas in our area.
I am just trying to improve the position of the mobile phones given by GPS - oh, the word! - with a simple algorithm.
Thank you for your help.

Improve it by how much? Bearing in mind that high precision GPS is achieved by having a base station receiving GPS positions, comparing them to its known position & then transmitting the corrections to the rover receiver(s), you aren't going to get any more accuracy by using the known position of a cell tower & measuring how long the signal takes to get to your phone. Since selective availability was switched off decades ago, uncorrected cell phone GPS XY accuracy is generally pretty good (within 5-6m), however Z accuracy will be 3 times worse - so around 15m. Even if you could improve your XY accuracy by using GSM signal (you can't), the Z accuracy would still be way out. Honestly, if it were that simple companies like Trimble, Hemisphere, et al wouldn't have spent billions of $ over the decades developing RTK systems (which in the early days were site-specific & could cost $100K+ to set up) & then improving them to the point where you can now subscribe to wide area correction services.

Even if you wanted to use a "proper" correction source, you couldn't - because cell phones don't have the functionality built in to be able to receive the corrections & then process them to derive a HPGPS position. There's plenty of good info on GPS & the various means of improving accuracy in the public domain if you want to Google it.

**** I was involved in setting up RTK GPS systems for mining & agriculture from the mid 1990s until about 2014 & worked for companies that specialized in these systems, including Trimble Navigation, Caterpillar & Aquila Mining Systems - all pioneers in applying GPS to machine control, guidance & autonomy systems.

- Colin.

 

[solfinker]

Agreed that we are not going to improve GPS location, that's not the point. I wrote it quite wrongly. But we want to use the antenna's locations in cases where there is no horizon, when the GPS signal is not accessible.
It is then an alternative to GPS since 4G or 5G signal's penetration is deeper that GPS's.

Thank you.

 

[Computersmith64]

Agreed that we are not going to improve GPS location, that's not the point. I wrote it quite wrongly. But we want to use the antenna's locations in cases where there is no horizon, when the GPS signal is not accessible.
It is then an alternative to GPS since 4G or 5G signal's penetration is deeper that GPS's.

Thank you.

It'll be interesting to see how successfully you can do that. I think your biggest issue will be getting any kind of useful accuracy. GPS has the advantage of the satellites having atomic clocks & the receivers being syncronized to them. Without that, I think you'll probably struggle to get any better accuracy than you already get from using the ACCESS_COARSE_LOCATION permission - but good luck with it. Let us all know how you go.

- Colin.

 

[Adrian Jansen]

Hello.
Is there a way to determine the elapsed time between the reception of positioning signals coming from different base stations?
Is there a time of arrival (ToA) measurement that calculates the time lapse between the transmission and the reception of a particular signal?
That would give us a way to determine the distance between the transmitter and the receiver.
The precision of cell ID positioning can be improved if the base station takes into account the timing advance (TADV or TA) measurement. Since GSM is a
TDMA network, it requires to know very accurately at what time stamp the signal from a particular UE (user equipment) will be received.
The TADV measurement is used by the network to schedule the UE uplink signal to arrive at a particular time slot. The TADV value is the time that an uplink signal coming from the UE needs to reach the base station. Hence, the TADV is a ToA (time of arrival) measurement and it can be useful to locate the UE within a circumference centered at the base station position.

Thank you.

Many GPS mapping applications for Android already do this. For example the Topo Maps series from Atlogis. But from what I have seen using it, the accuracy is pretty poor. Typically it seems to give error circles around 100-500 m in diameter. When the system finally gets a GPS lock, that falls to around 10-20m. I guess if you have nothing else, this is still useful. How exactly its done, I have no idea, but I presume it uses TOA to a few cell towers to do it. Here ( in Australia ) we have far less cell coverage than GPS, so it pretty useless. You are much more likely to get a GPS fix than a cell fix. My strategy when walking is to climb to the nearest high point, and get a good GPS fix from there. That works pretty much anywhere.

 

[emexes]

Typically it seems to give error circles around 100-500 m in diameter.

You might find that error circle is centered on the nearest cell tower.

When the system finally gets a GPS lock, that falls to around 10-20m.

I can't remember the last time I was more than 5 metres off when outside and stationary. If I am in a car park, then it regularly locates me to the correct parking bay or the one left or right.

https://www.b4x.com/android/forum/goto/post?id=908190

Lol using iPhone > 10 years old. I still haven't confirmed whether it uses GLONASS and GPS simultaneously. Not that it matters: Lebara/Optus are discontinuing their 3G network soon, and I have to move to a new phone anyway.

 

[Adrian Jansen]

No, the error circle is centered on the current position estimate, at least in the app I quoted.

Those 5m GPS error estimates are very optimistic, and based only on the current satellite geometry. The real errors due to ionospheric delays and possible multipath reflections from things like rock faces in canyon walls are much worse. Turn on a logger for an hour or so with the gps in a fixed position and look at the actual position plot to see what happens.

Certainly the error improves dramatically if you can get the other sat constellations like Glonass, Beidou, etc. The more sats the better. Many modern phones use all available. There is a nice app GPS Test by Chartcross which allows switching on and off the sat constellations used in a fix, as well as many other settings, to see what effect it has on accuracy etc.