SpaceX Satellite Signals Used Like GPS To Pinpoint Location On Earth (gpsdaily.com) 27
schwit1 shares a report from GPS Daily: Engineering researchers have developed a method to use signals broadcast by Starlink internet service satellites to accurately locate a position here on Earth, much like GPS does. It is the first time the Starlink system has been harnessed by researchers outside SpaceX for navigation. The researchers used signals from six Starlink satellites to pinpoint a location on Earth within 8 meters of accuracy. The researchers did not need assistance from SpaceX to use the satellite signals, and they emphasized that they had no access to the actual data being sent through the satellites -- only to information related to the satellite's location and movement.
For this research, [Zak Kassas, director of the Center for Automated Vehicles Research with Multimodal Assured Navigation (CARMEN), a multi-institution transportation center housed at The Ohio State University] and his research team studied the Starlink system and analyzed signals being sent by the satellites. They developed an algorithm that could use the signals of multiple satellites to locate a position on Earth. Then, they set up an antenna on the campus of UCI and tried to use the network to pinpoint the antenna's location. Using Starlink, they identified the antenna's location within about 7.7 meters. GPS, by comparison, generally identifies a device's location within 0.3 and 5 meters. The team has used similar techniques with other low Earth orbit satellite constellations, but with less accuracy, pinpointing locations within about 23 meters, Kassas said. The team has also been working with the U.S. Air Force to pinpoint locations of high-altitude aircraft; they were able to come within 5 meters using land-based cellular signals, Kassas said. The research paper has been published in the journal IEEE Xplore.
For this research, [Zak Kassas, director of the Center for Automated Vehicles Research with Multimodal Assured Navigation (CARMEN), a multi-institution transportation center housed at The Ohio State University] and his research team studied the Starlink system and analyzed signals being sent by the satellites. They developed an algorithm that could use the signals of multiple satellites to locate a position on Earth. Then, they set up an antenna on the campus of UCI and tried to use the network to pinpoint the antenna's location. Using Starlink, they identified the antenna's location within about 7.7 meters. GPS, by comparison, generally identifies a device's location within 0.3 and 5 meters. The team has used similar techniques with other low Earth orbit satellite constellations, but with less accuracy, pinpointing locations within about 23 meters, Kassas said. The team has also been working with the U.S. Air Force to pinpoint locations of high-altitude aircraft; they were able to come within 5 meters using land-based cellular signals, Kassas said. The research paper has been published in the journal IEEE Xplore.
Nothing new here (Score:4, Interesting)
The beacons on all of these low altitude satellites can be used exactly in the same way as GPS or Galileo. Someone (of course) has to write the software and build the hardware for this to be of any use. Not that difficult nowadays provided you do not need a small low power embedded version. A lot of the prerequisites are readily available.
Re: Nothing new here (Score:4, Informative)
Re: (Score:2)
Re: (Score:2)
I would have sworn there were a small number of geostationary satellites in addition to the polar satellites, but apparently they never actually rolled that out. (It was part of an early plan to improve accuracy. citation, p. 1-20 [uscg.gov])
Re: Nothing new here (Score:4, Informative)
Current Satellite Based Augmentation System (SBAS) implementations, such as the US WAAS and the European EGNOS, use GEOs to augment GPS solutions. WAAS also provides a ranging function, so its GEOs can contribute to positioning directly.
The next major version of SBAS will add Dual Frequency Multi-Constellation (DFMC) support, which does what it says on the tin: It will allow augmenting multiple constellations at once, rather than only GPS or GLONASS, and use two frequencies per GNSS satellite. Using two frequencies usually almost eliminates ionospheric errors in ranting estimates, leading to much more precise position estimates and freeing up the bandwidth that allows the multi-constellation capability. DFMC SBAS will also allow SBAS satellites to use non-GEO orbits.
Re: Nothing new here (Score:5, Informative)
No GPS satellites are in medium Earth orbit (MEO) with 55 degree inclination. That means they never go above 55 degrees latitude (north or south), much less over the poles. Galileo, GLONASS and BeiDou are mostly similar, although BeiDou adds geostationary (GEO) and inclined geosynchronous (IGSO) satellites. India's IRNSS uses a mix of GEOs and IGSO satellites without MEOs. Japan's QZSS uses three satellites in a novel IGSO that keeps at least one visible at a high elevation above Tokyo's horizon to help provide solutions in its urban canyons.
This means that GNSS geometry is worse for polar users, giving more uncertainty (higher dilution of precision) about their position than users can get at lower latitudes.
Re: (Score:3)
The satellites orbit at 20,000km ~12k miles.
So, GPS resolution around the poles is just fine.
You might need to adjust your GPS receiving system regarding the "flatness" of the earth around the moles, but most systems do that automatically. Be prepared to make sure if it does or does not, if you are sailing in the Baltics, though :P
Re: (Score:2)
What I said is right, although I skipped some of the details for simplicity: https://mycoordinates.org/chal... [mycoordinates.org] . (Lower HDOP does not usually help that much because it is already lower than VDOP; satellites are often in multiple compass directions, but never below the horizon.)
Ellipticity is handled through the ephemeris and the equations used throughout a GNSS receiver. If you have to adjust a receiver to handle polar flattening, it probably doesn't really work anywhere else either.
Re: (Score:2)
If you have to adjust a receiver to handle polar flattening, it probably doesn't really work anywhere else either.
Oh, you have. Basically all "cheap pleasure boat GPS" systems do not adjust automatically. Probably they are to old. No idea about models on commercial ocean going ships.
In the Baltics already, which is not that far north, you can have displacements of about 100m. Not really sure (never needed to do that), if it is simply fixed by adding the offset into the GPS or if the problem is indeed alrea
Re:Nothing new here (Score:5, Informative)
GPS, and in fact all the GNSS systems (Galileo, Baidu, Glonass), don't use geostationary orbits.
The main reason why it takes a while for your sat-nav to lock on when it's been powered off for a while is that it has to receive data about the orbits before it can calculate position. Your phone downloads that data from the internet but non-connected devices have to get it from the satellites, which takes at least 30 seconds.
Re: (Score:2)
Baidu [wikipedia.org] is a Chinese company. BeiDou [wikipedia.org], China's GNSS constellation, now includes satellites in both GEOs and IGSOs.
Re: (Score:2)
I recently turned one on after ~4 years off being powered off and relocating it over 5000 miles. It took closer to 30 minutes to "find itself" again (with a view of only ~1/3 - 1/2 the sky).
Re: (Score:2)
IIRC 35 seconds is the theoretical minimum, but in practice it usually takes closer to a minute in ideal conditions and much longer when the signal isn't so good.
That's for GPS, I don't know about the other systems.
Re: (Score:3)
Satellites in low Earth orbit like One Webb and Starlink are not great for GNSS applications.
First their movement is less predictable since they are more affected by atmospheric drag and the Earth's uneven gravitational pull. GNSS relies on knowing the precise location of each satellite and while it is possible to compensate for the extra wobble of satellites in lower orbits, it means that more data is needed and overall accuracy is lower.
Secondly you need a lot more of them, since each one is visible from
Re: (Score:2)
I don't think this is true, because unless a satellite is designed with GPS in mind, it will not have an on-board atomic clock. Such clocks are necessary for GPS to function at the accuracy it does.
Re: (Score:1)
Re: (Score:2)
The beacons on all of these low altitude satellites can be used exactly in the same way as GPS or Galileo.
GPS/Galileo broadcasts a timestamp, which is central to how those protocols work to measure travel time and determine position.
Starlink is not broadcasting timestamps, the method of measuring the signals being used here is nothing like GPS.
I'm not familiar with the OneWeb beacons, but they can't possibly be both transmitting timestamps and not transmitting timestamps at the same time.
If they do, they are like GPS and nothing like what is being done here.
If they don't, then you're right this isn't new, but i
LOL (Score:1)
within about 7.7 meters
So you're saying it was exactly 7.8 meters off the target? Or only 7.75 meters off?
Re: (Score:3)
Re: (Score:2)
I didn't even suspect that was the use case, it's true.
What if you're building a fence, then does it matter?
Re: (Score:2)
Without seeing the full paper, it's hard to say whether that is a circular wrote probable (CEP) number, meaning a 50% chance of being within that distance, or a one- or two-sigma bound. Or maybe it's the actual error for a single position fix, with no known statistical distribution.
Getting to 7.7 m does require a separate altimeter. Without knowing the receiver altitude, the horizontal error was 25.9 m (with 3-D error of 33.5 M, meaning about 21 m vertical error).
Re: (Score:2)
Goddammit, phone, circular ERROR probable.
I thought these things were supposed to be smart.
Re: (Score:2)
I thought these things were supposed to be smart.
You mean the Humaans, or the jeejah? If it's the jeejah, the humaan has to be smart enough to operate the settings interface to enable an advisory correction interface. Otherwise the jeejah does much of the talking. Don't worry. It's Smart(TM).
Re: (Score:2)
If it was CEP or something, we'd know exactly the calculated number.
I'm sure they used something sensible in their paper, and then it was written up poorly. But the summary takes whatever they really said, and instead says "about." But then doesn't use a round number. Which is what it would be if they just glanced at one of the measurements, and never calculated anything.
If they want to use the word "about," they should choose between 7 and 8. "About 7.7" is silliness, it will be all over the place within "