How GPS Works, And How It Got Better Than The Designers Ever Imagined

[@abarratt8869]

An explanation of AGPS for mobiles would have been a good addition. To save mobile phones waiting 12 minutes to acquire and decode the GPS network data, cellular network base stations supply that data directly to the phones. Thus a phone can set up its GPS receiver for a hot start, ready to go, and start calculating accurate position straight away. If you're away from all cellular networks, it takes your phone a long time to get a position!

[@PsRohrbaugh]

Wi-Fi also plays a part in this. I recently moved to a new state where I've got no cell service inside my house. My phone sees my Wi-Fi, and thinks I'm back at my old address. This makes it impossible to get a precise location lock unless I can talk to a cell tower. I had to buy a new router. Once I did that, the phone would use it's last location to look for satellites, and wouldn't think it teleported to another state!

[@gordonrichardson2972]

Good point. The key factor is that your phone will remember that AGPS data for a long time, even when out of cellular range. Removing the phone battery will force a cold start of the GPS receiver.

[@pedromiguelsilva2077]

​@@PsRohrbaugh that issue is not from real GPS, google or some provider keep a database of WIFI acess points and their location. A phone will send a list of nearby acess points to this provider, wich will look for matches in this database and return a estimated position. This can then be used as a initial position until a real GPS fix is achieved. Then the reverse is done, a phone with a good GPS signal will send the list of visible AP to the provider to update its database.

[@FABStudios]

I used to have a handheld GPS receiver in the 90's which took forever to lock but I never thought about how today it doesn’t take long because of the dropped selective availability thing, thanks for that little gem about cell towers.

[@ThZuao]

A long time and a lot of battery, I presume?

[@jimjohnson394]

I've worked with GPS a lot while surveying. You can fix one receiver over a known point, then use another receiver to do your surveying. The fixed receiver will send the rover receiver a correction by radio and you can get extremely accurate, down to 0.01 ft, or 1/8th of an inch, which is good enough for any property lines or general construction staking. It is so much faster than using Total Stations.

[@Nickle314]

RTK if you want to google it.

[@alexlandherr]

I think SparkFun sells a complete kit for advanced embedded developers.

[@tonymcflattie2450]

Can you confirm globe earth, as a surveyor?

[@Nickle314]

@@alexlandherrThey do. I quite like their surveying pole. It measures the position at the top of the pole, but the pole may not be verticle. So it uses accelerometers to work out the tilt of the pole, and uses that to work out the position of the spike at the end of the pole.

[@jimjohnson394]

@@alexlandherr In the early 2000's, they cost around $50k, now a RTK system runs about $8k.

[@ColeDedhand]

I was part of the testing when the Marine Corps adopted GPS in the late 80s. The satellites weren't up yet so we installed transmitters on mountaintops.

[@voornaam3191]

Ha, I met somebody who was testing in 1991. The first gps satellites were up and running, already. A handful only. It was pretty special.

[@Wizardess]

They were called pseudolites. [^_^}

[@GrigoriZhukov]

SLGR then PLGR. Worked good enough.

[@nonsequitor]

😂That might be the grain of the truth in a stack of flat earth conspiracy theories 😯🤷‍♂️🤦‍♂️😂

[@notmenotme614]

Basically how aircraft navigation worked using radio beacons. Such as TACAN and VOR

[@daemn42]

Another trick the smart folks figured out was that you can directly measure the doppler shift of the carrier signals from your various satellites and use this to calculate your instantaneous velocity in 3 dimensions, rather than having to use a running average of (P2 - P1)/t style calculations which are easily confused while you're turning rapidly. Some GPS systems used on race cars and such use mostly doppler data integrated with gyro and accelerometer data to calculate very accurate position on the track at very high speeds and turn rates.

[@nicholaswjamrock]

That is using the carrier wave, that kind of equipment is very expensive though. When you used that with the coded data and a differential source you can get below 10 centimetres or 3 inches. I worked on a project in 2003 that did just that. We had things then, that made google map/earth today, look like a joke

[@namibjDerEchte]

@@nicholaswjamrocksmartphones of the past couple years have carrier tracking, on android there's an API for that so you can use custom software for sensor fusion or recording to later post-process or do differential RTK...

[@floriandaler5327]

Using Kalman filters it is also possible to combine many different measurements (GPS, IMU, Odometry,...) into a really precise and also fast (high update rate) position estimate

[@shellderp]

that would be a good video

[@mossm717]

That’s cool! So is the velocity data similar accuracy to position data? Or is it better? I’ve noticed a gopro will give very responsive velocity information

[@tianyi05]

Agriculture has really taken advantage of this. My nephew has a 120 foot wide sprayer than travels at 12 MPH and is positioned within 4 inches as it moves.

[@ryanspence5831]

There are drones for agricultural work that have RTK and can stay within 1/2 inch of their course. Really cool stuff

[@Jegorex]

My brother uses RTK GPS and that has less than 1 inch accuracy. All he has to do is turn around at the end of the field.

[@tactileslut]

Yup, from robotic lawn mowers to large scale ag. The green and yellow tractor company promises centimeter accuracy.

[@David_Lee379]

Yeah that is pretty fascinating. I have a cousin doing the same thing; the on board computer is given the soil conditions ahead of time and can turn individual spray heads on and off to match. Then he got certificated and got himself an Airtractor and does it all at 140 knots.

[@Croz89]

A lot of Ag GPS systems augment their data with ground stations too. At the sugar refinery where my dad used to work there's a base station on one of the silos that helps improve the accuracy for local farmers.

[@johankotze42]

I worked on an opencast mine which used GPS to track the huge haul trucks as part of the haulage optimisation system. At the time of SA being turned off, I was at home recuperating from a neck operations, but I was logged into the system. At that moment, I could see the truck bread crumbs resolve into the left and righthand sides of the haul roads.

[@MatteoGalet]

Wow that must have been a discussion topic next day in the office! I'm curious, how precise was the tracking before that? Did you see the bread crumbs in general proximity of the road?

[@johankotze42]

@@MatteoGalet Yes. Haul roads are made at around 20-30 m width. I called the guys at the office at the time. I must say that the supplier of the system had their own form of correction as well. Now they use GPS to steer autonomous trucks. The company is Modular Mining Systems from Tucson, Az.

[@JeffMTX]

What AVL system did you use then? Oh never mind I see it

[@jamesbanas1815]

Wow....I geeked out on this video. It was nice to see your research said the first GPS satellite was launched in 1978. I know that for a fact as I was in the Air Force as a Range Control Officer at Vandenberg AFB and participated in that launch, managing the tracking and telemetry systems during flight. We had NO idea what GPS would become. You brought back memories and gave me a little better understanding of the GPS system. Thanks.

[@uploadJ]

We at TI (in 1978) had a contract to develop both a MANPACK and HDUE (aircraft platform) GPS receivers. We had field engineers at Yuma Proving Grounds working on that ground range testing equipment and also I recall the mention of three satellites "in service" at that time too.

[@leakycheese]

When I was at university back in the late 90’s, I remember some of the geophysics students telling me about how they got around SA for gravity surveying. They set up two fixed base stations spaced apart each with it’s own GPS receiver. They then used some smart math to compare the two locations and then solve the difference to work out the introduced error. This erro was then removed from the signal and they achieved sub-metre accuracy, this being very useful in the gravity mapping surveys. Clever stuff, I’m sure they weren’t the only people to figure this out prior to SA being deactivated. Very interesting video Scott, thanks.

[@JeffMTX]

That was called differential gps. It was used well into the 2000s. I did an integrated system for PANYNJ to track trucks on the GW bridge in 97-2000 that used DGPS

[@uploadJ]

re: "They set up two fixed base stations spaced apart each with it’s own GPS receiver." The basis fir something called Differential GPS.

[@michaelhompus2475]

S\A.. could you really remove the error with some smart math?

[@uploadJ]

@@michaelhompus2475 NOT w/o averaging a single receiver's position over ... say, hours.

[@jarink1]

I remember using some of the early Magellan GPS receivers during the 1st Gulf War. They took (at least) a few minutes to give you a location measured in 10s of meters and used up AA batteries like they were going out of style. Still, it was the *only* way we were able to effectively do our mission of radio direction finding, as it required knowing exactly where our equipment was located. There's virtually no landmarks in that desert! The Magellans were also pretty scarce; we were issued one per platoon (4 teams per platoon), meaning our LT and PSG had to escort us out to our intercept sites.

[@rockmusicman21]

How was radio direction finding done in WW2 without GPS?

[@jarink1]

Most WWII DF was HF, compared to VHF in the 1960s-2000s. Otherwise, it was basically done the same way, but it's difficult to get a good fix (where 2 or more lines of bearing from your direction finders to a transmitter cross) if you have an imprecise starting location. We needed a minimum of a 6-digit grid (10m) to start from, which frankly wasn't hard to get anywhere besides a featureless desert.

[@rockmusicman21]

@@jarink1 thanks for the explanation. I remember a documentary about SOE agents dropped into France that had secret radio sets that they clandestinely used but the Germans were monitoring it and found their location

[@georgejr2640]

Better then the hand drawn maps we had in the rear. MSR, to MSR. And of course, the 7th, and 18th had different names for the same MSR. We had a lot of luck guiding us.

[@procrastinator1842]

I vaguely remember those things. Now we have a gps in our watch that is 1000x better and uses a zillionth of the battery power 😂

[@pierreabbat6157]

I'm a land surveyor. Several days ago I took my GPS pole to two nails on a job site and measured them, then returned a few days later and measured the distance between them with the total station and the prism pole. The two measurements of the horizontal distance differ by 16 mm. I can usually get within 30 or 46 mm with GPS.

[@davidschultz3585]

You missed one of the big changes which was the evolution of receiver hardware. It started with a single correlator which had to be shared between all satellites used. A PITA and a reason behind the 15 minute cold start time. If you didn't have a current almanac (rough satellite positions) and time, you didn't know which satellites to look for. So find one by chance and wait for the almanac to come in. Over time more correlators were added until you reached the "all in view" systems. These don't need almanac data because they just assign one correlator per satellite. Thirty seconds later you have ephemeris data (accurate orbital parameters) and a position. Even better is that with more correlators you can track and use all the satellites in view rather than just the minimum of four. This lets you calculate a least squared error solution which improves accuracy. (You can of course fold data from multiple GNSS systems into a least squared error solution.)

[@erikincph]

The Russians are working as crazy to jam the GPS signals in Ukraine

[@voornaam3191]

Good point. I just knew the basics of how gps works, in fact the math resembles using a sextant and the Reed Almanac tables, it's known orbits and the reveiver does observations, it's just slightly different observations, here it is distances. So, I did understand how hard it must have been these days, to get these calculations done in time. A cpu in 1991 was bloody expensive, or it was cheaper and lacking number crunch power. How did they do it? And WHO built these chips? It was a military project, after all, that precision restriction was a big deal. Well, your comment explains a lot, it was dedicated hardware, not just fast i/o and smart software.

[@Wizardess]

Increasingly cheap computer power allowed the change from sequential tracking to simultaneous tracking. One of the freely available, but not open source, SDR packages available to day allows one set of hardware that samples at a high rate to feed a generally CPU capabilities or screen size allows number of simultaneous receivers. It's as if you had 16 receivers all sharing the same antenna, as might have happened in some military installations in the 1950s to surprisingly recently. Now they share one analog to digital converter. The rest is "a simple matter of software." {^_^}

[@wtmayhew]

The first GPS receiver I had was about 1995, a Magellan with a single channel correlator. The start-up time could be up to half an hour if the geometry of the satellite constellation was unfavorable at the time. I left the receiver in the car always on because it took so long to acquire the satellites. Back then, if you wanted maps, you had to run a program on a laptop computer. I had Delorme Automap which ran on Windows 3.1. At least the computer didn’t have to remain in the car and always on. Selective availability could easily put you on the wrong street when navigating in a city. The error was up to about 1/8 mile. You could usually tell when an SA glitch was being introduced and they were usually about 30 to 60 seconds. Clinton signed an executive order in 1999 to shut off SA by May 2000. I got a Garmin 2620 automobile receiver in 2003 which included built in maps with enough internal memory to cover all major and minor roads in continental US. The 2620 also includes WAAS reception. I was pretty skeptical that WAAS would work because it is something like -140 dBm due to the distance to geostationary orbit, and the car receiver has just a very small internal patch antenna. But work WAAS does. When it has lock, which usually only takes a few minutes, I’ve seen estimated positional errors reported by the receiver as small as 2.5 feet. That is probably good enough to land a plane, but not good enough for an autonomous driving car without other sources of position input. The other restriction that has been claimed was put on early receivers is that they were not supposed to work correctly at high speed. The intent was supposedly to prevent unfriendly powers from using commercial GPS units as guidance system building blocks. I took a Garmin Nuvi on a regional jet flight between Chicago and Denver and held the Nuvi next to the window. The Nuvi seemed to be accurately updating the the indicated position with the jet flying well over 400 MPH. I didn’t have any means to tell if the reported elevation data were correct.

[@Hauketal]

@@wtmayhew The speed limit is somewhere in the supersonic range. And there is also a height limit. Both won't restrict any commercial planes, but are intended to make commercial GPS unuseable for ICBMs and other rockets.

[@MakeAMark]

Another interesting fact about GPS which you didn't mention (at least i didn't hear it) is that it requires General Relativity corrections to obtain the accuracy we are used to. I always found that very interesting and surprising.

[@twinturbostang]

Yes! I was going to mention that as well! 😀

[@tricross6]

As far as I know, it's the only system that requires both quantum mechanics (atomic clocks) and general relativity (frequency corrections) to work properly!

[@francescodepascale7188]

And even more interesting is the fact that the in the first gps satellites two clock system were onboard. One with relativity correction and another without the correction because the military where note « sure » that the Einstein relativity was true !

[@Ni999]

The satellites are gaining 45 microseconds per day as predicted by General Relativity. That's taken up by lowering the clock frequency in the satellites to synchronize with ground clocks. The satellites are moving sufficiently fast enough to lose up to 7 microseconds per day as predicted by Special Relativity. That's calculated and corrected by each receiver.

[@uploadJ]

re: "it requires General Relativity corrections to obtain the accuracy we are used to. " Adjustments are made by ground control stations on a routine basis to correct for this. No magic involved here. Per the ops plan they use. This can be verified.

[@kkpdk]

While a phone-class GPS receiver is likely to have problems when an SV drifts, aircraft models are SUPPOSED to have RAIM-functionality, which, when they do have it, can suppress one bad SV and notify the pilots even before it gets marked unusable through an overlay system. Fun extra: with a reasonable antenna, a carrier-phase receiver, and PPK, you can see your antenna move up and down due to earth tide.

[@renesoucy3444]

So cool, I’ve heard that North America raise and fall about 10 meters when the Moon passes over it , is that right?

[@kkpdk]

@@renesoucy3444 The crust is a lot stiffer than the water, you're looking for an amplitude of around 0.3m.

[@uploadJ]

No one is mentioning WAAS. Wide Area Augmentation System.

[@baylinkdashyt]

No one except Scott, who went into it at least a little bit. Was the first AS he mentioned.

[@MattAtHazmat]

Most current GPS receivers have 12+ channels- you only *need* a fix on 4 satellites, but if you track more, the potential error goes down. Since aircraft tend to have a very good view of the sky, they can track all the SVs above the horizon, with some channels searching for fresh. Receivers track but only include SVs a threshold above the horizon in the fix, since their ionospheric variation may be worse.

[@JadedArsenic]

One seemingly crazy aspect of today's GPS with larger planes is that the software on each aircraft needs to take into account where on the airframe the GPS antenna *IS*. GPS is so precise nowadays that from nose-to-tail & wingtip-to-wingtip the plane can get a different location. On small planes it's not so critical, but with larger aircraft it can become a problem.

[@triffid0hunter]

So with multiple receivers in various locations they could get heading/pitch/roll from GPS? neat

[@sarmadchaudhry6006]

​@@triffid0hunterheading (yaw) is common with multiple antenna assorted setups. Often called AHRS for attitude heading reference system. However I haven't seen anything on pitch/roll from the antennas. But I guess it should be just as possible.

[@baomao7243]

Coast guard ships use bow and stern GPS RX in combo for precision station-keeping in all weather conditions.

[@imconsequetau5275]

If you move the second asterisk after the period, you will get the desired *bold* attribute.

[@MrLampbus]

We have the same issue on trains. a 220M long train with ant&reciever at each end. I was just using them to detect station arrival for selective door operations...fairly important not to open a train door where there isn't a platform.

[@tardx]

Worth mentioning the predecessor to GPS: the US NAVY’s Transit satellite system. This used LEO satellites transmitting ephemeris and timing data at 400MHz and 150MHz. The receiver used measurements of the Doppler shift as each satellite transited the sky to calculate a position. But the position fix was not continuous so dead-reckoning was needed between fixes. Also worth mentioning is the fact that satellite XYZ coordinates need to be converted to lat/lon/height using a recognized (and typically locally chosen) spheroid and datum.

[@macwall01]

In 1994 I asked the Dean of my avionics school to substitute one day of instruction on Omega for one day on GPS. He told me GPS was a fad and we would never see it in the real world. What I said to him next almost got me expelled. It makes me smile to look back on this.

[@woutervanheusden]

Hahaha why did you write this as a clickbait article

[@MrMilarepa108]

​@@woutervanheusdenYou will never believe what happened next...

[@JackThelRipper]

Lol! Yah I did something similar back in the day, I remember I was watching some tech nerd channel on the tv (basic coax cable mind you) and they were talking about cell phones and how one day we would be watching tv on them and also be able to go online and view websites. Well I laughed and said that’s bs and that the dude was smoking crack. But here we are and I’m watching this vid on my iPhone. But mind you back when he said it most cell phones had a tiny screen that mostly was green or grey like Nokia and we could play snake. 🐍

[@coriscotupi]

This is similar to what some British politician (I forget who it was) said back in the early days of the jet engine, that those engines were "a scientific curiosity and would never seriously challenge the 'airscrew' " (propeller).

[@supergeek1418]

Scott, Back then (in 1973), there *WAS* another satellite based navigation system called "Transit", which used a series of satellites in low polar orbit. These satellites constantly transmitted their position and a time marker every 2 minutes on 250 MHz and 400 MHz. The receivers (mostly aboard nuclear missile submarines) would track the doppler shift of the received signals, and (based on that) calculate the distance from the "bird" (satellite). Then (knowing both the satellite's position and it's distance) it would calculate its position --- to within 10 meters --- all using 1960s technology! The commercial version available was called SatNav, and it was only accurate to around a 1/4 mile, or so; but the military version was accurate enough to shoot missiles. Back then, in the Navy, aboard a missile sub, that system was one of my jobs.

[@mdxtrains]

The first successful tests of the Transit system were made in 1960, and the system entered Naval service in 1964. This system was first envisioned when two physicists (William Guier and George Weiffenbach) at Johns Hopkins Applied Physics Laboratories October 4, 1957 heard the Doppler shift of the Sputnik signal. Transit was in use until retired in 1991.

[@supergeek1418]

@@mdxtrains "Retired", yes --- but not shut down. From what I understand, it's still being used to monitor Earth's gravitational anomalies.

[@w9gb]

Thanks for bringing up “Transit”, as it was a “good” system that established a baseline for the future GPS specifications.

[@paulheinrich7645]

I know a lot of people with the gift of gab, but few ever say anything worth listening to. You’re exceptional.

[@beenaplumber8379]

Notice he doesn't have edits and splices every few seconds either. He clearly reads from an outline or something (which is good, IMO), but his excellent delivery seems to be driven by sheer enthusiasm.

[@siberx4]

CDMA is a bit hard to understand at first but is one of the coolest bits of radio/signal technology out there. One interesting quirk of CDMA is that while you can have multiple users/transmitters all on the same frequency at the same time, the more active users you have the higher your _noise floor_ becomes. This effectively means that the *range* with CDMA based protocols reduces as congestion goes up, and you can sometimes see this happening in busy areas with cell phones (especially with marginal signal) where you will either lose bars at times of high congestion or will find that your cell phone only connects to the tower in the middle of the night when there's little traffic going on.

[@rgbaal]

good comment. Yes CDMA sites "breath" which can cause real problems for operators. Thats why 5G does not use CDMA.

[@Jannes54321]

@@rgbaal To my knowledge CMDA has only been used in 2G and 3G standard of mobile communication. 4G was designed completely independent of 3G. 4G uses OFDM signals with FDMA and TDMA instead of a single carrier signal with CDMA in 3G. One big advantage of OFDM in a multi-user scenario is which each user can be assigned to the part of the spectrum which is received the strongest by him. This type of resource allocation is called multi-user diversity. The same diversity concept can be nicely extended to the Multi Antenna transmission. The so-called Multi user MIMO or MU-MIMO.

[@uploadJ]

re: "This effectively means that the range with CDMA based protocols reduces as congestion goes up," Proper RF coverage, as performed by the RF Engineering department, in core areas of coverage work to assure sufficient margin exists to keep this from happening. Engineering isn't a 'crap shoot' where things just happen, there were methods and plans in place to accommodate the technology, JUST as there was for 1G (analog) FDMA cell technology. Former RF cell eng here.

[@siberx4]

@@uploadJ I wasn't implying it was some wild uncontrolled factor; obviously it's accounted for when you design cell spacing, power levels etc. Designers know very well how various numbers of active users at various distances will affect the performance of their cells and ensure they're managed accordingly to provide consistent performance. It's simply a factual property of CDMA modulation, and an unusual one compared to other modulation schemes where range is not dependent on congestion (only throughput), so I thought it was worth mentioning. It takes a pretty exceptional scenario to even notice this (edges of coverage at unusually busy times) which is a testament to the exact fact you mention, which is that a lot of design effort goes in to ensuring you have a good experience with cell networks across a broad range of conditions and locations.

[@uploadJ]

@@siberx4 Fair enough. You have a good reply there too.

[@TomHill-xh7ec]

Lots of fun to hear about this. I worked in the AF squadron that handled early orbit in the Block IIA satellites in the early 90s. Lots of ideas being tossed around for improving the signals. Great to see how it's changed the world. It will be a tough day if someone jams the signals widely.

[@Yeadon65]

An interesting quirk you may be aware of Scott. The GPS clock transmits a week number and a time offset which GPS units used to calculate the date and thus the time difference for the signal. Due to data size limits this week number transmitted cycled back from 1024 to 1 in August 1999. For many older pieces of GPS equipment this confused them and made them show a position vastly different from their actual location. I was working in the UK civil aircraft area at the time and the UK CAA actually issued a date conversion table to reset the date on those GPS units until they could be replaced to correct this issue. So the actual date may have been a day in September 1999 but to get an accurate position, you had the set the clock in your GPS unit to a corresponding day in the 1980s!

[@nitehawk86]

The y1.999k bug. But really what did they expect to happen? Did people think those devices were not going to last that long?

[@baksatibi]

There was another rollover in 2019 and according to Wikipedia it affected a lot more devices. Brace yourselves for 2038 when we're going to have both GPS and UNIX time rollovers in the same year.

[@Hauketal]

There is this quirk in German regional trains. They often have a display with the current date and time, and the name of the next station. This gets its info from GPS, and lately there where a lot of sightings with a date 1024 weeks back from 2003.

[@miroslavmilan]

That is truly unbelievable. One would have thought people would learn from the Y2K hysteria once and for all. I know you can’t use long integers for all variables just in case, but can’t we just all agree to come up with a more robust way to store date/time in general ? 😅

[@Yeadon65]

@@miroslavmilan Computers and calendars/clocks have never got on well!

[@bmanna495]

Thank you so much Scott, as someone who recently finished an avionics course and got extremely fascinated by both the mathematical and technological workings of GNSS all the added history makes it even better

[@baboutheocelot778]

Differential GPS works by correcting for the distance to each satellite from the reference station separately, then using those corrected distances in the remote receiver's calculation, not by calculating an offset position from the reference station. This way the remote receiver doesn't have to use the same satellites as the reference station for it's position calculation.

[@rowanjones3476]

Expanding on this, ‘classic’ DGPS corrects the code phase of the ranging signal, whereas RTK/PPP etc use a variety of techniques to correct the carrier phase of the ranging signal for a higher degree of precision/ and ideally, accuracy. The limit to achievable accuracy in cases where range correction terms are used is that they assume that the range error for the reference station and remote are identical. The further we get from the reference station, the less this assumption holds true. Modern state space correction techniques synthesise a wide area correction model from multiple reference station observations within it, from which locally applicable correction terms can be calculated by the receiver.

[@Wizardess]

If it is able to use the same satellites the plane can correct distances better because the signals pass through pretty much the same ionospheric effects as the reference location. You do not HAVE to use the same satellites. But, you can in principle get better results that way. {^_^}

[@airplaneian]

My mom worked on the GPS program at Rockwell and we had a few of the early generation trooper units at home. Really fun tech, and the last time I managed to power one on, still worked!

[@jdsahr]

A remarkably good and accurate description of GPS (which I know something about). A couple comments: (1) one of the reasons that SA was disabled had to do with the first Iraq war. Turns out that most of the GPS equipment that went into the field did *not* have the (ITAR controlled/classified) SA decoders. So DoD had to turn SA *off* during Iraq 1 in order to avoid hindering US troops. In other words, SA wound up working almost exactly backwards from what DoD intended. (2) GPS is so accurate that the system includes a correction for General Relativity; There is enough spacetime curvature between the 12 hour orbits and near the Earth's surface that the satellite clocks have to be slowed down a little bit so that by the time the signals "fall" to the Earth, they have "sped up" just enough to be able to ignore the GR effects on the ground. If you're using GPS far from the Earth's surface, that GR "correction" will have to be accounted for. (3) If you have some prior knowledge of your altitude, you can get your lat/lon with fewer than 4 satellites. Also, if you are not in motion, the movement of the individual satellites will effectively generate "new" satellite positions that can get you additional links. Buzzword on this topic: GDOP, geometric dilution of precision. (4) a key data output is the time. You can get GPS-based time servers for your computer that will provide the correct time within a few nanoseconds. (5) Heard on NPR today. If you're using a handheld GPS device in Moscow today, you'll notice that it doesn't work very well. This is because Russia is jamming GPS to try to confuse those wascally Ukrainian drones.

[@mnxs]

GPS not working in Moscow because of "wascally Ukrainian drones"? Thank you for that tidbit, good sir. My schadenfreude is immeasurable, and my day is perfect. Also, I'd heard about something like that with (1). Something like, you knew when some particular war started, because GPS suddenly got a whole lot more accurate. Didn't know that it was due to SA being turned off, neat!

[@joshcryer]

There is a misnomer here though that the techniques used by SA were turned off. In fact the GPS system still skews time. UTC(USNO) (GPS time) "is kept within a close but unspecified tolerance of the international atomic timescale." The military knows the tolerance in real time, the timing data is released two weeks after the fact (so atomic clocks can be sycned up). SA was just a very large interruption in the clocks which could be corrected by simply having a clock station on the ground as the video notes.

[@marcoaurelio4903]

5 is fake.

[@bevlawton1488]

We were using the NAVSTAR SYSTEM(GPS) In Gulf War 1 in 1982 - it actually played a MAJOR Part in the Victory - for once the Allies KNEW Precisely where they and everyone else where..... The First GPS Satellite was launched in 1978....... The full 24 Satellite Constellation was operational in 1993........ Selective Availability was turned off for Gulf War 1(Kuwait) - NOT IRAQ 1 - basically as you have said there were not enough Military Compliant GPS Recievers so they bought a load of Civilian Grade ones - as SA Degraded thier Accuracy they turned it off temporarily.?.. SA got switched off permanently @ 24:00 1st May 2000

[@the1spyderryder]

But the waskilly Ukrainians can do the same thing by building ground stations as in the WAAS system or the Coastguard did.

[@Cianan-vw1lb]

I worked on GPS user equipment in the 80s. Geophysical services was particularly concerned with getting the highest precision possible. Excellent summary of how things have evolved! Thanks.

[@patbrown2227]

Great overview of GPS etc accuracy. One more obscure improvement is the new much smaller Cesium Clock oscillators in the satellites. Some of the early satellites used Rubidium clocks as they could be made smaller and cheaper. All of the latest block satellites have very small and low power Cesium clocks. I believe that higher satellite transmitted power also allows smaller antennae in cell phones and other small GPS receivers. The development of better, cheaper, smaller, lower power Cesium clocks was an exciting time to be around the Hewlett Packard Precision Time and Frequency Division.

[@jeffrodeback8390]

Great video and info Scott. When I was a full time flight instructor, I used to teach an instrument rating ground course. I always looked forward to GPS day-helping the students make sense of all the acronyms (GNSS, WAAS, SBAS, RNAV, LNAV, LPV, RNP, etc). If I’d had this video then I could have assigned it as homework and then my class on GPS day could have been 15 minutes shorter! It is truly amazing how GPS has modernized air travel. Some of your Cirrus’ features are more modern than the 737 I’m on presently, but we can do full RNP(AR) approaches which make getting into places like Eagle/Vail or Guatemala City very easy. As long as the satellites are cooperating that day. Thanks for the content!

[@DNPaterson]

Hard to believe it's been 50 years since GPS was first proposed! It's something we rely on every day now, and never give it a second thought. Although, as you mentioned jamming in the last segment, it's a good idea to keep the sextant handy, and always stay in touch with the old-fashioned ways of doing things as well. Even the US Navy has started teaching celestial nav. again, just in case...

[@1dgram]

He showed a slide covering MON but didn't talk about it. MON stands for minimum operational network and is a network of VORs that the FAA will keep operational as it serves as a backup to GPS when it comes to aerial navigation.

[@justinpinto5048]

Yes. A really big solar flare could throw the whole thing out of whack, I guess.

[@beenaplumber8379]

I've always imagined that the Defense Dept. has kept a switching mechanism available so they can shut the whole system down if needed for national security, like, if they believed an attack was imminent. Maybe shut down or introduce an error again.

[@michaelgoetz7506]

A few corrections: You missed one of the major drivers in the demise of SA. The purpose of SA was not to keep civilians from having accurate navigation. It was to stop an adversary from using our GPS system to accurately guide their weapons towards American targets. The plan was to have SA turned off during peacetime, so everyone could get the benefits of the system, and to turn SA on during wartime so that GPS could not be used against us. Well, you know what they say about plans not surviving the first contact with the enemy. During the Iraq war, as, you noted, because of the lack of available military GPS receivers, our ground forces used privately purchased *civilian* GPS receivers. SA had been turned on before the war, but, ironically, the DoD had to turn SA off during the war so that our forces could use civilian receivers. This was the exact opposite of the intended usage of SA! I'm pretty sure SA was never turned back on after the war. One other thing that you missed: Even at the beginning, it was possible to get centimeter-level resolution with GPS. I'd have to look up some of my old, old documentation, but I think it was some form of early differential GPS. One of the uses was to measure continental drift. It's worth noting that there's also something called AGPS -- Assisted GPS. This is what phones use. Cell phone towers send (I believe) the almanac and the position of the cell tower to the phone. This lets a phone's GPS initialize from a cold start in seconds rather than tens of minutes. Finally, you mentioned that the data frames give the almanac data for one other satellite. It's possible the system has changed in the 30 years since I last looked at the details, but I'm pretty sure each satellite transmits the entire almanac. This was important in the early days before the full constellation was in orbit. Aside from those little errata,, that was an excellent explanation of a very complex system. Well done!

[@RoccondilRinon]

They did turn it back on after the war, but that was one of the main reasons for turning it off permanently in 2000. One of my favourite pieces of historical irony.

[@Wizardess]

Search terms might well include Magnavox. For increased accuracy look up Kalman filters. That is a trick for deriving better data by grading multiple sources for goodness and factoring that into the final solution. That was another lunch time discussion that resulted in table linen leaving with the customers. {^_-} {o.o}

[@brianorca]

I think another factor in disabling SA was the launch of the other systems such as Glonass and Galileo. The purpose of SA was so an adversary could not use GPS to guide a bomb to an accurate target. But if those adversaries have their own system, then there's no reason to degrade our own civilian usage. But they have tested a new type of SA which can be turned on or off for small regions and specific times in response to a specific threat.

[@wtmayhew]

SA was definitely turned on through about May of 2000 in the area where I live. It was easy to tell when the satellite clock was being messed with because it would suddenly make you appear to be up to 1/8 mile from where you thought you were a few seconds ago. SA made using GPS based road maps for navigating in cities annoying to say the least. Clinton signed the executive order to ‘permanently’ turn off SA in 1999, and it was done in 2000. It would be interesting to see what the economic boost has been to have a reasonably accurate civilian positioning system. I am sure the benefit is huge and outweighs the cost of the system many times over.

[@Wizardess]

@@brianorca Do not confuse Selective Availability with Denial of Accuracy. They were two different animals. I touched on both. On SA I guessed which way NSA was going to jump for encrypting P-Code, disabling military precision. But, that was not enough. So they cooked up Denial of Accuracy, fuss with the base satellite frequency source in a manner that should be undetectable. They settled for "very hard to detect". The betraying signal artifacts were nominally buried in oscillator native noise by about 20 dB, a factor of 100. I designed that synthesizer concept. Another fellow did a bang up job designing the actual circuit more or less while I was designing and developing the RF half of the pre-launch test set with those 60-70 hour work weeks due to (dirty word alert) politics. We came in on time. ITT had their Nav Data Unit, NDU, ready about when the unpaid overtinme I had accrued proved the OT was not needed. SIGH. Basically the design theoretically permitted enabling either the SA or the DA only for partial orbits. I never heard anything about whether it could do that or some problem they''d missed popped up. And don't forget that the synthesizer allowed Phase 2B to run with an accuracy greater than the standards aboard, which is nice. Rb standards are wonderful for most casual uses but really not that good for something like GPS. And the GPS standard design flown (memory says 2 Rb and 1 Cs) was not as good as what exists on the ground. So the synthesizer gives a nice accuracy boost by fine tuning the onboard standards. Those birds are nice beasts. Um, and Rockwell International provided me with some nearly magical tools to work with. And Sam Costanza, our department manager, built the best group of people I ever worked with. Sadly he passed die to pancreatic cancer and the department slowly decayed. That's why I ended up "across town" at Magnavox in Torrance. {^_^}

[@pleasestaysafe2787]

Just to clarify one point for Scott, the United States Coast Guard was not part of the DOD. That is the department of defense until after 9/11. Before 2001 they were part of the DOT, the department of transportation. That’s why the military being part of the DOD would not have shared technology with the DOT being the Coast Guard. Does that make any sense at all? No, of course not. But it’s why the Coast Guard seemed to be excluded. Thanks Scott, great video!

[@scottmanley]

My point was they were all part of the US Government and they were wasting money and effort to circumvent SA

[@shawnmiller4781]

The DOT had the coast guard for years decades because the US Lighthouse Service was merged into the USCG aim 1939 and that is when the maintenance of marine navigational aids such as bouys, lighthouses and Lightships. That is why the USCG would also later take on the mission of maintaining electronic navigation systems. They operated the Loran A and Loran C systems and eventually took over the GPS and WAAS system operation from DOD

[@shawnmiller4781]

The USCG was one of the agencies that was merged into the Department Of Homeland Security as a result of the panic that created DHS post 9/11 that also resulted in their “law enforcement” mission. Keeping the USCG out of DOD also elevated posse comentatus (yeah I know my Latin sucks)issues because the military is prohibited from civilian law enforcement activities

[@pleasestaysafe2787]

@@scottmanley I one hundred percent agree. I was not trying to correct anything that you said. Just trying to explain why in the 90s the Coast Guard was not part of the military.

[@solandri69]

@@shawnmiller4781 USCG was part of the Dept of the Treasury until 1967. Presumably because it was involved with immigration and customs (which also used to be part of the Treasury). It then got moved to Transportation, until the Dept of Homeland Security was created. It gets transferred to the DOD during times of war. Used to require a Presidential executive order. But they passed a law in 2006 making it automatic after a declaration of war.

[@chrisglen-smith7662]

Back in the 80ies (may have been early 90ies, not sure)I fitted an early Garmin GPS on a ship for a special project, it had a monochrome CRT screen (only displaying Lat, Long and Speed IIRC, no maps!) and was so heavy it took two of us to hold it up while a 3rd put the bolts in ! The satellite constellation was not complete so it only worked about half the time when there were enough satellites in line of site. And it took at least 1/2 hour from turn on to getting a fix even if there satellites available because it had to receive the orbit elements for the satellites every time, no saving it for next time. But it was still like magic to us at the time.

[@csongorvarga]

My major in the uni was surveying. I remember when we used GPS in classes it still had CA. And we would stay ona a point for several minutes that the random error would average out from the readings and we can an accurate coordinate. And it was turned off once I finished my studies. And I started using GPS car navigation on a Windows CE Compaq iPaq with a handheld GPS receiver connected with 9pin serial cable and some adapters. Boy it was messy, but it worked.

[@SHAMIRMCLEISH]

I’m an avionics technician and started working on airplanes with GPS as the norm for long-range navigation. But it’s wild to consider pilots are doing synthetic ILS approaches now to land commercial flights and coupled with transponder data, planes can be their own ATC with ADS-B Out and subsequent ADS-B In capabilities. Awesome video!

[@CW-0114]

I really like the part on vertical variance vs horizontal variance, it made so much sense!

[@markfisher7962]

Per 17:00, in the 1990's the vertical error was a source of humor when using GPS on my boat. We regularly got altitudes of negative tens of feet when we demonstrably were AT sea level.

[@gordonrichardson2972]

Part of the problem is that the geoid used by GPS is a theoretical model, while sea-level depends on gravitational fluctuations in the earth's crustal rocks.

[@markfisher7962]

@@gordonrichardson2972 I know about that, but it doesn't explain from 70 feet above sea level to 50 feet under water in one mile. When all your triangulation points are above you, accuracy suffers.

[@phuzz00]

With more precision you run into the problem of how do you define what 'sea level' actually is. You can take an average of the sea's height over many tides, but that will vary according to location, some places have very large tides, in some places the difference between high and low tide is relatively small.

[@markfisher7962]

@@phuzz00 Well, my rule of thumb is that if fish are swimming in it and it's tidal, it's below sea level. I do have concerns about depth soundings, but as my boat draws 3 1/2 feet, if I run aground, I can jump over the side and push it off.

[@gerrywalsh5766]

The KC-135A that I was with in the mid 80's had 3 separate navigation modes. TACAN1, TACAN2, GPS, and the old reliable sextant. Lol

[@leepshin]

Never mind GPS..."!!Finally!! KZbin notifications worked for once letting me know right away there was new content. 😊

[@jakeroper1096]

Thanks Obama!

[@richb313]

I have lived and worked through the entire evolution of navigation from my time in Sea Explorers in the mid to late 60's through my time in the Navy on subs though the early to late 70's to working off shore from the early 80's through 2010 so I can relate to and understand almost everything you touched on as when doing pipeline surveys before 24 hr coverage reacquiring the GPS signal was money. I don't have the expertise to explain it as well as you do thanks Scott

[@MillionFoul]

I never knew what exactly WAAS was despite knowing what Differential GPS was and starting my own IFR training. I like that you used an approach into KJAC as a demonstration of a GPS/RNAV, as I live there and worked at the airport for a while. Speaking of, that runway actually does have cat III ILS, but nobody uses it anymore. Also, we just got the runway replaced like two years back and it has cool centerline lighting and all that. Fun airport for sure!

[@nathanel1313]

Great video as always, but I was really waiting for mention of the need for relativistic corrections, as it blew my mind the first time I heard about it. I guess that's so common knowledge to Scott that me did not feel the need to mention it.

[@gumbykevbo]

When I graduated with an Engineering Degree in 1985, I interviewed with TI. One of the things they showed me that they were working on was an early GPS receiver. It was literally larger than a foot locker. Just small enough to be mounted on a tank. They were working on a next generation, trying to make it small enough to mount on a Jeep. These had multiple circuit boards with semi-rigid coax running between them. Also no mapping at all in them…you got lat. long. Alt, and time out of the box, and that was it. Now all this is one tiny chip that takes up part of one corner of a wristwatch, and the accuracy is better.

[@uploadJ]

We (at TI) had pretty much finished up the HDUE (High Dynamic User Equipment) and MANPACK (portable) GPS receivers at the end of 1978 ... I would have expected smaller units were being made by 1985, like the size of their LORAN receivers and the marine radios of that same era.

[@erikstreck2916]

Thought you might enjoy this. I'm an AVI tech on CRJ-900's. The FMS (flight management system) on the fleet I work on use DME and VOR for navigation. There are two DME systems on the plane and each system can handle 3 channels at once. 2 of these channels are for the FMS and one for pilot control. So when the FMS is programmed it listens to up to 4 DME stations using trilateration to plot against a map database in conjunction with VOR stations. The CRJ-900 does have GPS capabilities as well but as far as I'm aware they use the FMS as well as iPad with foreflight for most navigation.

[@FlyNAA]

Normally we run with autotune on, so all 6 DME's are jumping around controlled by the FMS. And regardless of whether the first 2 are auto or manually tuned, they are used in the multilateration, if receiving a station. Also the GPS'es are running full time and mixed into the FMS position (as are the IRS'es, if installed). The FMS position uses all* of these sources simultaneously, but weighs the GPS info far more heavily. You can check all this stuff out by hitting MFD DATA, then MFD MENU and then line select NAV STATUS. It shows you on the MFD all the nav sources, each with yes/no being currently used, and the difference from the FMS position. GPS is always 0.0 or 0.1 NM, while the others are normally half a mile or more off. Since you're on the ground there's a good chance it won't be catching more than 1 (or even zero) DME's, so it might not be using that at all. For even more detail on that, line select VOR STATUS. * when I say "all," actually it only uses 1 GPS and 1 IRS at a time, I guess ready to switch to the other one if the first fails. Why not both simultaneously, I don't know.

[@hartvanmeter214]

Thank Scott for one of the most thorough and cogent explanations of GPS beyond just saying it’s a couple of dozen satellites and based on time of signal propogation. Had no idea how the code timing and interleaving worked until you drew the picture for me. Again, thanks.

[@MikanYT]

Hey Scott, nice video! Especially alot of the political things were very new to me. Im working on the clocks for the next generation of GNSS satellites and its pretty crazy how stable the new optical ones can get (and damn are they annoying to fit onto a spacecraft :D). Maybe a follow up on clock technologies could make for an interesting video? :)

[@clayel1]

woah, what company/organization?

[@MikanYT]

im talking about optical clocks based on doppler free spectroscopy, those havent launched yet but are being prepared to fly on the ISS @@mandellorian790

[@MikanYT]

some national space agency in europe@@clayel1

[@UpLateGeek]

I remember learning about CDMA in my uni wireless communications subject, it was one of the more fun subjects. We actually did a paper example of decoding the CDMA signal, recovering multiple data streams from a raw signal. Took a while to find the right alignment of the code and the signal, but was like figuring out how magic works when you found it. So it's fun to learn how it's used by GPS.

[@HappyBeezerStudios]

CDMA is easy compared to QAM, that is still some kind of sorcery to me.

[@UpLateGeek]

@@HappyBeezerStudios they didn't cover QAM in as much detail as CDMA in my uni subject. I learnt a lot more about it when I was a broadcast engineering trainee, since it's used in DVB-T transmission and satellite up/downlinks.

[@JeffMTX]

Ya gotta know your trig and complex math to get QAM! Digital trellis modulation works the same way :)

[@davidford694]

One of the earliest adopters of GPS was the BC Wildfire Service. I remember sitting in the office, which was near the harbour, of their manager of technology development, watching a computer tracking a receiver in a pickup parked outside. Suddenly Selective Availability kicked in, and the truck went SPOOSH right into the drink. Well, the computer thought so.

[@jimparr01Utube]

A super intense information narrative Scott. As a retired engineer, I appreciated it. What has always astounded me is that GPS receivers can fish out accurate data-sentences from satellite signals that are typically 12dB beneath the thermal noise floor. Mostly arcane math processing to accomplish this.

[@KentHenry8]

As a fairly new amateur radio operator, seeing how they shifted the spectrum lobes for around for different lobes to keep compatibility with older systems is pretty cool.

[@fsodn]

A quick correction: aircraft "surveilance" (reporting your position) uses BOTH secondary transponder radar AND ADS-B. To fly in controlled airspace in the US you're required to have BOTH. Presumably so that the FAA can smoothly transition between the two. But even for those with 978 ADS-B are still required to have the old-school 1090 transponders. Go figure.

[@SeanBZA]

Redundancy, so that you still have one if the other fails. No doubt they eventually will have a single system that is a tenth the size, with both integrated into it, and with a quarter the power consumption. then will demand 2 as well.

[@rgbaal]

Secondary radar does not require GPS - it will work just fine without so it will be kept (like ILS/VOR) do deal with situation of GPS failure. Some reason that the powers that be are restarting LORAN and eDME research.

[@petesheppard1709]

I had to step away from general aviation in '07, as the GPS revolution was taking off, and I found my little Garmin portable unit to be an incredible boon! I hope to get back into flying, and expect the biggest challenge to be catching up with the new nav systems. That said, I will keep charts of some sort with me; besides the unlikely event of some sort of equipment failure, the challenge of navigating by following features on the ground is FUN!

[@JonathanMickelson]

Outstanding video Scott! Just brilliant! Thanks for putting out such stellar content. Dang son… that section on the frequencies was an awesome deep dive into GPS… even the layman can follow and understand why it works even if they don’t understand the actual math 😀

[@philipoakley5498]

That code shifting for the extra signals in the same band is very similar to the UK analog Colour TV signal called Phase Alternating Line (PAL) which had a 25 Hz offset relative to the black & white signal. It was very similar to the French SECAM system. The magic of comb signals.!

[@deepthought3316]

Always in amazement with GPS and the ease it will give your position. I remember when in the Merchant Navy you needed a time signal to correct the onboard chronometer, a noon day sight and a lot of calculation you might be accurate within 5 miles..... An amazing system, much underrated.....

[@calvinkielas-jensen6665]

Thank you for saying "1 meter in the 95th percentile". As someone who works with probabilistic methods a lot, I am always bothered by people arbitrarily stating things like it's accurate to 1 cm without at least the percentile!

[@kirkwagner461]

Shortly after I got my PPL I also got a hand held GPS unit. That was in the days of Selective Availability GPS. Although I never got lost during a flight using the old paper maps and eyeballs, it was reassuring to have that unit in my bag. By the time Selective Availability was turned off I was no longer using that first, clunky old GPS unit. And now my phone has far better accuracy than I had available in flight way back when.

[@baboutheocelot778]

If it was an old clunky Apollo GPS, I programmed them when I worked for II Morrow.

[@kirkwagner461]

@@baboutheocelot778It was a Lowrance Airmap handheld unit. Literally about the size of a brick.

[@xxwookey]

That was great. I took a huge interest in GPS in 1996 when I first got one (for locating cave entrances) until after SA was turned off and I could stop worrying about the details so much because it just got a lot better. And I was amazed at how much better the clients got from 8 channels and 2W in 1996 to lots of channels and multiple systems (glonass etc) using only mW with _much_ faster startup just in an average phone. (The better accuracy is nice but it's actually the the much lower power consumption that really makes the most difference from the early days INHO. But I was always rather vague about the various DGPS and WAAS and EGNOS systems, so it's nice to have the details and timelines clarified. Some info on what you get with modern paid-for DGPS would be nice one day too.

[@kmlammto]

A very good summary of GPS. I was on the Dept. of Transportation (the lead Civil Agency on GPS) GPS staff and for a short period of time (due to maternity leave of tha actual lead), I lead the effort to obtain the second Civil frequency on the GPS system. We came up with a way to jointly use the L2 with the military. I then left the program to work in Congress. I have many stories to fill in some spaces in Scott’s presentation, but let me quickly summarize. The sideband signal was developed by a very smart young engineer to add a civil signal to L2. We could not get any Agency to admit the superior accuracy of having 2 frequencies was beneficial until the USDA stepped forward and said they expected autonomous farm equipment would experience benefits in the range of $1 billion per year. So, if you love your GPS being able to get you somewhere new or your pizza to be delivered directly to you anywhere you are, thank a farmer.

[@flat-earther]

hi kmlammto, have you become a flat earther yet?

[@keithrange4457]

It is crazy how much much GPS has entrenched itself in our lives. For simply the aspect of civilian daily navigation - how dud we ever get anywhere before GPS. Even my industry uses GPS. We make aircraft obstruction lights (wind farm, cell tower, skyscraper, chimney, etc lights). There are many times where lights separated by large distance must flash in sync for FAA regulation compliance. We do that with GPS. We only need 1 satellite for sufficient time accuracy of 1/60th of a second

[@toonarmycaptain]

GPS is awesome

[@JohnMuz1]

I was playing geostationary bingo too. Double fun.

[@ashchap]

Don't forget RTK GPS which uses carrier phase information to get cm level accuracy relative to a nearby reference station (and can be used to calculate orientation if the reference station is on the same vehicle), or the more exotic PPP, which is a private (expensive subscription required) Space Based Augmentation System that provides

[@HughWilliams1]

Would love a follow-up dealing with RTK (real-time kinematics) enhancements that get you centimeter precision. Can you imagine if every cell tower was required to act as an RTK base station?

[@jasonspall2157]

I have been involved with CORS (Continually Operating Reference Station) and Network RTK for many years. With multiple constellations and improved processing algorithms, you don't need a base every few km's. Generally speaking, the accuracy of modern Survey RTK GNSS equipment is 8mm + 1ppm in the horizontal with vertical being 1.5 to 2 times that. So at 5km, the Hz accuracy is 13mm. Network RTK allows the modelling of Ionospheric, Tropospheric, Clock and Orbit errors over a wide area such that you don't need to be really close to the base to get the accuracy. Most CORS networks work on a 40km baseline between CORS. Improvements in algorithms has seen 80km, 100km baselines provide great results. Saying all of that, the GNSS Reciver and Antenna are critical to the accuracy. The Rover Antenna's ability to filter out unwanted signals, the ability to detect and reject multipath (reflected signals) and the algorithms used to compute Network RTK, are vital to accurate and reliable results.

[@mnxs]

​@@jasonspall2157 13mm accuracy... Based off signals coming from satellites in medium earth orbit. That's absolutely amazing, bonkers, wild, and a whole host of other superlatives. Mind = blown.

[@peterjenkins3922]

Thanks for the info. You may also like the following story. I worked in the North Sea oil business in 1973 - 1976. Exploration drilling rigs needed to accurately determine the location of the oil/gas well you were drilling. The old Decca system was very iff'y. So we switch to using the US sat nav system. A container sized computer would be shipped out to us. Once the rig was achored we would wait for a satellite to pass over. Then wait hours for the next one and so it went. After acouple od days we would roughtly (better than counting Decca lines) know where we were. Happy days. PJ

[@treborg777]

This was a great description of GPS with lots of details I haven’t heard before. Great job! I’ll add two details (1) GPS uses General Relativistic corrections to achieve its accuracy, since the signals fall down into Earth’s gravity (2) The actual GPS signal is too weak to see on an oscilloscope. The signal is below noise.

[@Rob2]

Maybe a nice followup video could cover the problem that we have complete dependence on these satellites (not only for location, but also for accurate time and frequency in many systems, e.g. cell networks, broadast transmitters etc) and it could all be gone with a big solar storm... Not fun when on a single day there is no more navigation, no more cell phone network, and no TV and radio either.

[@pseudotasuki]

The first time I ever saw GPS used was during a Boy Scout orienteering competition. One of the other kids was using a handheld GPS receiver to essentially cheat, but because it was such a primitive device it didn't store any of the orbital elements. As a result, he had to wait a few minutes every time he tried to use it, resulting in him coming in last place.

[@wesmcdonald7624]

Of course associated with the deployment of the satellite system was the development of small, low power receivers. In the early days we needed a gps receiver at a site - hauled a rack mounted unit on a commercial flight. The receivers shrunk over time a specialized correlation chips were made. I think Rockwell made the first receiver on a PC Card about the size of a pack of cigarettes, could be wrong there. Anyhow now of course the receivers are tiny, thank you Moore’s Law.

[@SeanBZA]

Biggest impovement was the patch antenna, which placed most of the receiver filtering at baseband into a single precise ceramic puck, that not only was the antenna with a wide acceptance angle, but also the front tuned circuits and selectivity, and with it being a very high Q filter to reject out of band noise as well. Got rid of half a rack of cavity filters and amplifiers, that you needed to keep at a constant temperature, and also a good chunk of power needed for all the amplifiers and temperature control.

[@rkan2]

Nowadays GPS antennas can be like 1/10th of your fingernail!

[@connorrules6093]

As a flight instructor, I still learn a lot everyday, and I am very familiar with GPS, but I stilled learned a lot going through this video. This was a very well put together video, and now I have more knowledge on it, and I really appreciate that, this is definitely going to be a video I recommend to my instrument students, as it made understanding the GPS system very clear, as well as very interesting. I knew the GPS systems were cool, but now I have a whole new appreciation for the system

[@MarkLanett]

What a tremendous amount of reading you had to do for this, and how much much further it could have gone!

[@hamilde]

Caterpillar has a system called acugrade (or something like that) that uses some kind of known location. They have GPS receivers on each side of a dozer or grader blade. It is my understanding that they can cut a grade within a half inch.

[@Octa9on]

I don't know how Caterpillar's system works, but if the two gps receivers can exchange phase data from the satellite signals they can calculate their relative orientation much more accurately than their absolute position

[@hamilde]

@@Octa9on I never worked with the system during my career there but I worked around people that did. I think you are probably right but I never learned the details.

[@SeanBZA]

Also your mobile phone absolutely depends on GPS, just to provide an accurate clock that allows for the time slots per device to be as tight as possible, allowing for maximum data throughput, but also to provide an accurate clock that allows the base stations to be able to use QAM256 to get as much data through per signal transition, using the GPS clocks to generate a very precise, as in down to single parts per billion accurate, clocks for the base stations, to do this. No GPS, no clocks, and very much degraded phone service, as the towers need to fall back to a very coarse clock provided by a heated crystal, good enough for one part per million, but now resulting in a much reduced data rate per device.

[@paulhaynes8045]

I've often wondered how apps like Google maps know where I am so accurately. Is it GPS (and if so, have I got it for free?!), or is my phone just triangulating from the nearest masts? Turns out (if I understand your post correctly) that it's both - my phone triangulates, but the transmitters know where they are via GPS. Have I got that right?

[@salerio61]

that isn't true. Most base stations use an internal time base and will synchronise across the network to ensure clocks are synchronised.

[@abarratt8869]

That's not quite right. The cellular networks use GPS for time and frequency references, to help keep all the base stations in sync (which is necessary for modulations such as CDMA (3G) and OFDM (4G upwards). The phone itself takes its timing from the cellular network, and doesn't need GPS at all.

[@SeanBZA]

@@paulhaynes8045 The phone has GPS yes, but also the towers, because they have a fixed nearby location, also will send the ephemeris data for all the locally visible satellites to the phone on request, so the phone GPS can lock really fast, and it also uses the time difference between local towers to gat a very fast and somewhat accurate location if indoors, so that it can at least give a position to within 30m almost immediately, even if there is no sight of the sky for GPS signals. The GPS signal is used outside in your car, to track better, but rough location is done using the signal strength and timing of local towers. Did once turn on a phone GPS with no coverage from the network, and it took around 20 minutes to finally position me, within a 100m sphere, while another phone, with service and data, took under a minute to position, using data and the 2 visible towers to get a rough area to gain coarse position fast. GPS in your phone needs data, stand alone GPS can do without it, though your accuracy improves with time on, as it refines errors out long term, though you also get the map used providing some sort of sanity check, as there is an assumption if you are moving you are on a road, allowing it to remove parallel paths easily.

[@SeanBZA]

@@salerio61 Yes, but the local master stations will also use a Cesium or Rubidium clock, which is slaved with the GPS to provide an accurate local time and frequency, which will be shared by the local cells. But the local stations all will ultimately have a clock derived from a GPS receiver somewhere local, which is at a precisely mapped point, used as a reference marker for all timing in the area. Depends on cell density just how big an area is covered by each one, at least large enough that the coverage to the other GPS equipped stations is under the error budget, so that you can hand data traffic over to cells with different master clocks and not loose too many packets in the handshake.

[@MCsCreations]

Scott, we use GPS a LOT with rc stuff as well. "Drones", but they can be multirotors, airplanes, boats, rovers... And now you can find on the market what they call M10 GPS chip based antennas. The difference is that you can connect with multiple satellite networks, so you get a much better fix, faster and with more satellites. You usually need at least 8 satellites locked before you take off, but with those chips you get 16 satellites easily. Some times even more! And people use to fly to pretty distant places, in the middle of nowhere, to record waterfalls on the mountains, for example. So what it gives you is "return to home". Which means that, if you lose contact with the craft for some reason, it's going to fly back and land by its own. (And those things are expensive, unfortunately it's not a cheap hobby, so you need it.) Anyway, stay safe there with your family! 🖖😊

[@ninehundreddollarluxuryyac5958]

Last time I flew a cessna was in the late 1980's. Your nav display is amazing. You're right, you could land with the GPS alone. A few years ago, I had to drive my boat at night (running from a hurricane) and was amazed at the accuracy of my GPS chartplotter as well as its hyper-accurate maps. That thing knew where I was and where channel markers were within no more than a couple of feet. I could drive all night with no moon, by GPS.

[@bozhijak]

I was in the Coast Guard and served at a LORAN-C secondary station on the gulf coast. To think that i was involved in this precise location technology in the early years gives me a sense of awe. Thanks Scott!!!

[@flat-earther]

hi bozhijak, what do you think about all gubments drawing a line around you(Antarctic Treaty) and saying you are not allowed to leave? If you don't know what I mean, read my about tab.

[@CompleteAnimation]

Now that you've introduced the topic of jamming, now I want to know how jamming actually works!

[@PsRohrbaugh]

GPS signals are incredibly weak at ground level. So you either block them completely (can be done with a $200 box) or transmit more powerful signals with spoofed signals (can be done with a $20,000 box). This only works in a local rang of you transmitter, so hundreds of meters to 10s of kms depending on transmitter power. But it's very easy on a local level. Like say around an airport. Jamming on a large scale requires much more power or number of transmitters, since GPS signals are coming from space.

[@SeanBZA]

@@PsRohrbaugh Plus possession of such a jammer, or at least using one near an airport, will have a few government agencies find you, and have a very long and pointed talk to you. Ask the truck driver who did that near a US airport, because he wanted to jam the GPS tracker on his truck, so he could have some off time on the clock.

[@JohnShalamskas]

@@PsRohrbaugh Or put your jammer on an aircraft.

[@repairengineer]

I always thought the reason you need 4 satellites was the following: 1 satellite puts you somewhere on the surface of a sphere (distance from Sat) 2 satellites put the reciever on the intersection of the 2 spheres which is a circle 3 satellites intersects a new sphere and the previous circle which gives 2 points The 4th satellite is required to resolve which of the two points is correct. Good video as usual!

[@aquilux-vids]

If the time function wasn't involved you could in fact use 3 minimum for all non-aerospace applications because if the receiver is ground based it's fairly safe to assume that the point at a higher altitude than the satellites themselves is probably not the correct one.

[@beenaplumber8379]

I could swear only 2 were necessary for my ancient Magellan, though much less precise. The 2 spheres alone will give 2 points on the surface (the only part of the circle that matters), one of which will be stationary while the other is moving as the satellites fly their different orbits.

[@SimonLanghof]

I was in Switzerland last week and came across a multilingual "GPS data not correct!" sign in a valley. When I checked my recorded hiking path later, it was quite OK until I stopped for a little break next to a building and suddenly the position jumped to the mountain wall. After I was moving again it was off by a few meters.

[@bobshowrocks]

That's probably multipath! The signals from the satellites can bounce off things and take a longer path to the receiver, which makes the receiver think it's in a slightly different location than it actually is. That or there's someone spoofing signals nearby, though that's usually against the radio transmission rules in most countries. I'm not sure why they would put up a sign if they were illegally transmitting.

[@Keldor314]

I suppose what must have been happening was that the mountains were not only blocking direct signals from satellites, but the detector could still pick up reflections bouncing off other mountainsides (and thus traveling a longer distance due to not following a straight line). This is very much analogous to echos.

[@SeanBZA]

@@Keldor314 Yes multipath reflections and diffraction off of the sharp edges of the mountain peak, which means the signal appears to be shifted in time with the reflection off the other side, and also takes longer with the diffraction, so the GPS receiver calculated position is wrong, but to the GPS it is still a valid position on the internal map it has of the area.

[@MatteoGalet]

Where did you find this sign?

[@SimonLanghof]

Lauterbrunnen Valley, if my memory is correct 46°35'18.8"N 7°54'29.5"E

[@RalphEllis]

Regards VIR and ILS transmitters, I understand that they will be kept in operation. The problem being that if we have a Carrington Event, all of the GPS satellite will disappear, and aviation will be grounded. But the old technology of VOR and ILS is more likely to survive. R

[@SKYGUY1]

As a 40 year pilot I used to think the the first aviation versions of the Loran receivers were the best tool a pilot could have since cabin heaters in January. Then I played w/ my first Garmin portable receiver... GPS-95 XL about 1993ish. That was slicker than A/C on a car in Southern Arizona. It is a wonderful tool for navigation. I wonder if anyone could calculate the fuel savings by aircraft since using GPS. And of course, fuel savings means less poison spewing into the atmosphere, and less money spent on the hundred dollar hamburgers (which are up to around $225 now). And the peace of mind the GPS receiver lends to a pilot having trouble w/ recognizing pilotage landmarks, or receiving a VOR between mountain ridges, is worth a great deal too. Thanx for the insights Scott. Terry - CFI-I

[@cseguin]

I worked in forestry in Canada back in the mid-90s and we used GPS for various purposes - the position data we gathered was uploaded to a computer that had special software that took the uploaded data and compared it with a base station file from the same time frame to correct for errors . . . the base station was a stationary GPS receiver/recorder that ran 24/7 and you had to dial into their server to get access to the files.

[@happyhome41]

I remember when there were three satellites up - our Air Force office had a lunch-box sized GPS receiver, and it could only be used certain times a day. I got permission to take it home one weekend, and invited a lady friend to go out in the woods. As we got deeper in the woods, she got agitated, convinced I was lying about the box, and was going to take advantage. The receiver came through, and we arrived at the car. Those were the days. Most excellent topic and coverage Mr Manley.

[@mattmatt516]

I'm a corporate Jet Pilot and we use GPS LPV approaches all the time. It's pretty amazing that there are GPS approaches that can get us down to the same 200-foot minimums on an ILS will!

[@seanspoerhase3878]

Do you know of a specific airport where that’s the case? I’ve yet to find one personally. I believe you lol I just haven’t seen it.

[@glujaz]

@@seanspoerhase3878LFPN has one for the 25R. 07L is 250ft above ground, which is also impressive. Both are DA. To be more specific, it's 200ft for cat A aircraft, otherwise, it's 220ft. For the 07L, it's 250ft for every aircraft.

[@mattmatt516]

@@seanspoerhase3878 Off the top of my head, KARG RNAV RWY 18 and 36 have 200ft mins. Same with several of the RNAV's into KOKC and KOUN

[@LK-jo8xt]

@@seanspoerhase3878very common in the US. KDLH RNAV 09 is the first example that comes to mind.

[@ivanscottw]

One thing that isn't mentionned I believe is that the GPS (When 4 sats are visible) also provides an extremely accurate TOD (Time Of Day) clock - and is extensively used by - amongst other applications - financial institutions that trade, usually against each other, stocks - automatically sending pull/push orders to some clearinghouse that are then granted or denied on a 1st come 1st serve basis based on a GPS clock timestamp with a, I believe, better than microsecond accuracy (The L1 sampling frequency - And GPS doesn't use the peak to trough time of the signal - it uses the signal slope to get a better estimate). Oh and last thing - I'm pretty sure modern GPS receivers (like a smartphone) don't have to listen for the WAAS information (the iono/stratophere atmospheric induced variations) - they just get it off from the Internet !

[@Octa9on]

interestingly enough, GPS time is not equivalent to UTC. for example, it doesn't include leap seconds, so it's like 17 seconds off of UTC at present

[@ivanscottw]

@@Octa9on Very true ! But IIRC the clock tick message sent by the GPS system has a "leap seconds" (as well as a century field just in case) field to indicate the offset ! It's probably very important for the thousands if not millions of events that occur between 23:59:59, 23:60:60 (the leap second) and 00:00:00 next day

[@declan8649]

Yes, I thought Scott might mention the derivative uses and the impact GPS has had in global cache coherence. The Spanner system is a brilliant example of this. For every OPS in the GPS constellation there are now m(/b)illions of correlated IOPS.

[@LcavalcanteR]

I work with satellite positioning for marine usage. We can offer 4 centimetre precision. That DGNSS explanation you did was very accurate, and we can offer that correction thru a geostationary satellite or even thru internet connection. We cover pretty much all the globe, with hundreds of reference stations. About the height, you can see a better precision under good HDOP values, HDOP indicates the concentration of SVs in the sky. When they are more spread, the better is.

[@tomfromtacomas7249]

Thanks Scott it was fun geeking out with you. I’ve been around GPS from the early 80s and it is truly amazing how dramatically it’s improved overtime for both Marine and aviation and of course, along with automobiles

[@markmcculfor6113]

Very good video Scott! I wish you would have covered some of the uses of GPS in agriculture as well, many farmers are getting sub inch accuracy on their equipment using ground stations and GPS.

[@biggles1852]

Awesome video as usual! GPS has been my livelihood since I first encountered it in college in the late 90’s. I remember taking land use sampling points and seeing all the SA error, then diffGPS those points against CG buoys and seeing the magic when the points suddenly all made sense.

[@cathedrow]

ESA have a really good manual on this - “TM-23 GNSS Data Processing”. Surprisingly readable, and very comprehensive.

[@jonathanamundsen6086]

This is a great video. And I'd love to have it include how Geologists squeeze 3 more orders of magnitude out of the system so that they can track geological plates and individual mountains moving much less than a mm per day.

[@mattcwatkins]

As a guy who had a GPS Receiver (yes I know my receiver isn't a "GPS" itself) in 2001 and an early geocache, I remember when Selective Availability was turned off in 2003 and received my first GPS with WAAS enabled. Instead of being in a football field of a cache it was like like 15 feet or so depending on canopy. I can't believe it's been two decades and how much I and the world take GPS for granted! Thanks for the retro view with the right level of detail to understand. Didn't know the maths was so layered.

[@billy-go9kx]

In the mid eighties our submarine bought a commercial GPS because the military version was not coming soon. Well a few months later we got the USN version. There was not much difference in results. Of course being in the middle of the ocean who cares about accuracy of a few meters.

[@RamsesYT]

This was thoroughly educational, Scott, and entertaining as always. Many thanks! 😊

[@ke6gwf]

You left out one key part at the end. Many GPS receivers, for surveying, grading, farming, etc especially, use ALL of the constellations at the same time. This way, even without RTK (Real-time kinematic positioning, the system used for surveyors, heavy earth moving equipment, and precision farming to get millimeter accuracy, similar to the differential GPS systems you talked about), you can get greatly improved accuracy, because each system has different error modes, different frequencies so you can calculate ionosphere conditions better, plus you get a lot more satellites, so you can get more accurate positioning data. And combine that with RTK with a base station at a known location, and you can do a GPS controlled CNC milling machine with sub millimeter accuracy! Lol Or, you know, get accurate survey data with no Total Station and lasers, or have a Grade Control equipped road grader or paving machine make perfect roads, or have an auto steer equipped tractor fertilize a field and keep the drill lined up perfectly between the rows of seeds, while the millennial farmer makes silly KZbin videos and complains about the times the system crashes ;)

[@sentinel76]

When you say "GPS controlled CNC milling machine", are you talking about the ITAR-controlled machines that lock out if they're moved? Or something else?

[@ke6gwf]

@@sentinel76 no, I am talking about a CNC milling machine that uses GPS receivers instead of positioning transducers or shaft encoders ;) Picture a GPS antenna on the base for the RTK signal, one on each end of the head, etc. Totally impractical, but it would be a fun project for Make Stuff Here or someone to make! Lol

[@TaylorPeterson913]

When I was a teenager, an adult tried to convince me we need four sats for gps and it was not clear why. You cleared that up in mere seconds 15 years later!

[@lon3don]

At those frequencies, there will also be tropospheric anomalies caused by temperature inversions. Colder air and warmer air will have different refractive indexes. Thus, there will be a lens effect that will bend the signal paths. A different effect to the ionosphere but significant nonetheless.

[@offdagrid877]

My Garmin Epix pro can use multiple gps systems at the same time depending on which setting I use, last weekend I navigated around the maze at Leeds castle using my watch, it could tell which side of the hedge I was on that’s less than 1 meter accuracy.

[@PsRohrbaugh]

This ability is incredibly important given geopolitical issues. For example, I was using GPS + GLONAS for years but when Ukraine stuff started, my accuracy went WAY down. Like, hundreds of meters off. Turning off GLONAS fixed everything.

[@TealJosh]

@@PsRohrbaugh Related to the Ukraine stuff starting. GPS was jammed in the Eastern Finland and couple airports had to cancel some passenger flights because instrument landings were impossible. VOR had been disabled years ago so the approach was GPS only. Weather and jammed gps ended up being a major problem.

[@thekamakaji]

This sounds like you were cheating!

[@hairychris444]

Yeah, I can easily tell which side of a road I was on when logging runs on my Garmin Endura. Clever kit which seems to be getting more accurate as time goes on!

[@InconsistentManner]

Hedy Lamar an Hungarian born actress created CDMA during WW2. We use this in Wi-Fi, Cell Networks, and as Scott Described In GPS. As well as many, many other radio communication technologies.

[@paulhaynes8045]

She was Austrian born. Her mother was Hungarian, but Hedy was born in Vienna and never lived in Hungary nor (as far as I know) spoke Hungarian (her mother, like most middle class Hungarians of the time, probably spoke German). Her father grew up in what is now Lviv, which was then in the Austrian part of the Austro-Hungarian Empire, but it was never part of Hungary.

[@geirmyrvagnes8718]

@@paulhaynes8045 Yeah, but they didn't have GPS back then, so who knows? 😅

[@paulhaynes8045]

@@geirmyrvagnes8718 I'm not sure it would have helped! The political geography of that time was a nightmare. It's indicative that almost every city in that area (modern day Slovakia, Ukraine, Poland, Romania, Hungary, even parts of Austria) has at least three names in different languages, sometimes four! They might have known where they were, but they would all have had different opinions as to who or what they were!

[@geirmyrvagnes8718]

@@paulhaynes8045 Yes, I was in the Balkans this summer talking to various people about what nationality Nikola Tesla was, and I figure the most correct is probably to call him Austrian, because that whole region is more of a mess than even the stereotype says... Beautiful area, lovely people, but... I wouldn't want to be a kid learning the last few hundred years of history. It is a lot to take in.

[@paulhaynes8045]

@@geirmyrvagnes8718 especially as everyone has their own version of that history - often nothing like that taught in the country next door! The current situation with Crimea is a classic example of this. My youngest is studying a poem at school and I was helping him with it the other day, when I suddenly realised that it was set in the Crimean War (as we in the UK call it). The Turks, the French and the Brits fighting the Russians over a place that didn't really belong to any of them. Try explaining that to a 14 year-old in 2023! And the lad in question (my son) is half Hungarian, so he is currently utterly confused as to why his mother's and father's views on the Russians invasion of Ukraine differ so much (his mother, although Hungarian, was actually born in what is now Ukraine but is pro-Russian!).

[@LG-ct8tw]

I remember sailing across the Atlantic starting in France with DECCA and switching to LORAN half way to the US. Early GPS was downgraded for civilian use to about the same accuracy as LORAN. At one point we had to have all 3 on board + paper charts!

[@northerncaptain855]

I’m an old commercial sea captain, back in the mid 70’s we were still using the old LORAN A on the U S side of the Atlantic and Decca if fitted in European waters, both of these systems were very range limited and were dependent on shore transmitters which usually only covered highly trafficked areas. In fact in international trade we lived and died by celestial navigation. Around 1979, I saw my first Transit Satellite Nav setup fitted which gave occasional fixes. I believe LORAN C came into fairly common use in the very early 80’s. In the later years when the full modern GPS systems came into common commercial use, the everyday sextant and celestial navigation skill set largely faded although commercial deep sea navigation officers are still required to be trained in its use.

[@fallinginthed33p]

​@@northerncaptain855A bit of a shame really, since it was celestial navigation and reading the ocean that allowed early explorers like the Polynesians to cross huge distances. I'd wager most people nowadays don't know how to find the north star or what the Southern Cross looks like.

[@KyleRhoades7]

Hi Scott, I just saw you at the airport on the East Coast right after watching this video! I decided not to bother you, but I want you to know how much I appreciate what you do. You are a modern day renaissance man. Fly safe!

[@bazem]

15:00 So basically "It knows this because it knows where it isn't. By subtracting where it is from where it isn't, or where it isn't from where it is (whichever is greater), it obtains a difference, or deviation."