okay well here we are at the Computer History Museum I'm Chuck Reno docent at the Museum and we're here to record the oral history of Franklin Nicollet Frank is the I think one of the architects of what we'll call a GPS and we'll get into that as we go along but you want to just identify yourself Frank sure I'm Frank van dillen and I've worked on GPS almost forever I actually have worked in navigation for ever on account of I started off as a teenager in the South African Navy and I went to navigation school and learned navigation I didn't know back then I was gonna end up in the GPS industry and that's where I've been since I graduated college and I work at Google just down the road and so you said that you were in the in the Navy that would have been the South Africa yeah yeah midshipman eyes yeah that's right now I was a midshipman and then an ensign which is a junior officer after high school so that was like in the early 80s mm-hmm and what was your duties as a midshipman I did like I said I went to navigation school and I was on a small ship and I was responsible for the navigation of that ship so that was an early introduction to navigation or yeah it was I've been sailing since I was in racing sailing since about five years old my dad was a sailor and he okay taught me and so racing sailings got his own form of navigation and that you're trying to go faster than the other boat so you got to be aware of you know the surroundings the wind and all that so it was very natural for me to get into that in the Navy and then kind of led to a whole career of GPS so how did you get then transitioning into technical you were interested in math there science or things in the early part of your life yeah so then the Navy was that it was the draft in South Africa so white boys men young got drafted out of high school into the military and so I was there for two years and then then went to college which was always the plan to get to go to college and I studied Electrical Engineering and then went on to PhD in Cambridge in England and that's when I met someone who had a GPS company her name's Allison Brown yes and maybe you'll interview her for this serious - yeah yeah so she recruited me to a GPS company in Colorado and that brought me out to the US okay there's some big jumps there can you tell us a little bit about how you get from South Africa to Cambridge all right well yes I did my undergraduate degree in South African and I wanted to go to carry on and do graduate studies and Cambridge seemed like the best place to me because of the scientific heritage so I applied there and I got in got a scholarship and and I went yeah that's great yeah so that's a yeah the legacy Isaac Newton - yes Charles Babbage I want oh yeah just a lot of names and then you mentioned now you've got you got recruited when you were at Cambridge as a when you were doing your PhD finishing your PhD yeah but that got you all the way to Colorado that that seems to be another big jump yeah that's right oh that was a big jump so Allison Brown who ran this company called nab sis in in Colorado Springs and she still does it's still there they do government contracting for on GPS mm-hm to do some research in GPS and she needed someone to do that work and she was visiting came she went to Cambridge - and her professor was my professor she was visiting him so she just happened to be there at the time and I was looking for a job at the time we my professor knew I was looking at a job he just called me up in the lab said hey do you want to come talk to someone who's he was looking for a researcher in in GPS he knew I did of course so so I went up and lucky for me I was busy writing up my PhD right then so I wasn't sleeping very much I don't think I changed clothes in about three days I hadn't shaved so I looked like the perfect candidate to her she was like this is exactly the kind of guy that won so that's I got the job and so what was your major in Cambridge it was control theory yes so that was kind of a lot in the math department yeah very much math and that's kind of what took me and that's why I Alice and wanted to be on that project and so that's what brought me out to America and then I stayed yeah okay so from so we're always from South Africa to Colorado Springs now and that would have been in the early 1900s I know heart tight it just ended about that so that was yes I was 19 my son would say it's the early nineteen hundred's I was kidding me about being too old 1990s 1992 came to to Colorado Springs yeah and actually there was two years before the first democratic elections in South Africa okay yeah that was 1991 I think for our tide officially and I think I think it was 94 before they actually had the elections okay yeah I remember my parents came and visited us and it was all in the news and so they and I don't know if you remember but it was on the news in the US every night and that was the first time in America they were like other South Africa's really famous are they the same parents are they still like my mother's still alive yeah and where is she in Johannesburg they're still there yeah yeah it's not the area you were mainly raised in uh well like I said we moved around a bit I was in Mozambique for a while Durban I was born in Cape Town which is really the place to be that's and that's why I ended up in the Navy again yeah all right fascinating okay so we are at Colorado Springs now you had a number of associations then with companies at that time that really got you introduced to GPS mm-hmm tell us about that in in Colorado Springs yeah well you started with what was it the naps his corporation yeah so as an abscess in Colorado Springs and that was small company and I was doing this work with the Coast Guard and ended some work with the US Air Force so got to do a lot of things like we built a receiver from scratch there was kind of spread it was a prototype to spread out in the lab front end and processing system and so I got to learn a lot from Allison Brown and then got to work with met new some people from Trimble and a shtick which were to survey companies ok Trimble was already an operation then yes it was and it's very interesting I feel like because that was really the first commercial use of GPS was very high accuracy survey used actually in oil fields for geologists in the oil industry they were prepared to pay thousands of dollars for an instrument that could tell him exactly where they were and so Trimble navigation developed systems that did that by measuring the carrier phase of GPS and a shtick was a competitor to theirs and and and what's and so I worked for ashtec and I worked on a high precision GPS in the mid 90s and what's really interesting is that GPS is kind of that's where it began the commercial GPS was was very high accuracy GPS and they're slowly evolved into consumer used in phones at medium levels of accuracy as you used to you can navigate your car but the loop is gonna close we're gonna have very high accuracy and phones again yeah well a very high accuracy in the hands of consumers in phones because those kind of measurements those carrier phase measurements are now now available in phones and it's just a matter of time before people write the software to make use of those measurements and get the kind of accuracy out of a phone that these companies like Trimble pioneered in big boxes like this that sold for $50,000 well we're the early systems then just using ranging I mean they didn't use the carrier phase at all or well now the early system the the early military systems just used the from the the code phase on the signal but the the the early Trimble systems just used the carrier phase yeah it's an interesting thing to explain when you explain the different levels of accuracy to Jeep to people with GPS and most people who used to the GPS and their phone are surprised to hear that GPS is capable of centimeter accuracy and an even more surprised to find out that that was actually the very first use of you know people say oh so when did this centimeter gps happen and the answer is that was the very first thing in the commercial world the first commercial use of gps was the very highest accuracy funnily enough and then we we sort of learned to do low accuracy but that's a really a joke we learned to do low-cost less expense yeah yeah yeah and so so there was a yeah and then you work with the company Magellan to at that time yeah so that company ashtec merged with Magellan they were there was a whole period of consolidation of GPS so a lot of little companies the medium-sized companies I consolidated and yeah so so a schtick and Magellan merge so I didn't really the company name changed my job didn't change yeah and so yeah so at that time then you were really getting into commercial applications of GPS that had to keep cost down and processing right what kind of signal processors were you using at that time and these were these were not dedicated chips that early wear day yeah those companies did dedicated Asics to do the signal processing they'd they they'd prototype and they'd sometimes yeah they'd prototype on DSPs but they would make their own Asics to do the GPS signal processing is really a matched filter process where you've got a signal coming in you know the structure of the signal just trying to match it and get that so it's it's simply processing than what you have in LTE or something like that were you trying to communicate some with the main purposes data communication GPS does have a component but the main goal is just to match this incoming signal because the the signal at the satellite has got a spreading code on it that's a known spreading code and then you you you align yourself with that spreading code and then you know how much delay there was from the satellite to you because if you move further away from the satellite that code appears to shift in time yeah it's really measuring a change in position from that point yeah and so that's really it so in a way it's it's a lot simpler than in communication systems yeah alright and so that's I like to say that's why I like it well you still kind of modulate the signal but at such a low rate yeah yeah so okay so from there yeah the let's see did we jump past the first we haven't come to the first cell phone application no so in the 90s now yeah okay a schtick and Magellan Magellan what's interesting about them they made the first handheld GPS there was a thing called the Magellan pioneer I think all the GS 104 yeah it looked like a big old phone like a big brick and there was about that long and I just told you latitude long as you don't know it's like you know a number you know like 37 points and nobody should do 20 to it yeah was what it was for but for people like sailors especially sailors there was it was terrific you had this thing in your hand that did for you what used to be a big box to use something like the transit system and and in people like hikers emergency workers used that so that would there was a Magellan so I in the 90s and then cell phones so GPS got in cell phones around about 2000 and the thing that spurred that was the 9-1-1 mandate from the FCC they and so maybe I should yeah playing that yeah yeah so so with with 9-1-1 but you know everyone in America knows downline when when you get through to the emergency services and in other countries there's similar thing like one one two in Europe and so on and what was happening when cellphones came out people so it's just well before cellphones your phone numbers registered to an address so when you dialed from a landline phone they knew where you were calling from from the from the phone number then with cell phones what would happen is the cell phone itself would be registered to some region like you got your area code and what they were finding was people calling on cell phones I had no idea where they were they did the put the emergency services answered the call would maybe contact the you know Sheriff's Department or something in the home area of that person the person's on vacation somewhere else they didn't know they have way of knowing where the person was and more and more 9-1-1 calls were going out over cell phones so the FCC mandated something called 911 911 and said that when a call was made from a cell phone it the carrier that placed the call so verizon or AT&T or whoever had to provide location with the when they connected the kola they had to also provide the location of that device and so that was like an in late 1990s that that mandate came in was was made and there was a period of about ten years which was phased in and it had to be certain number of phones with a certain accuracy by a certain time so is there more than one way people were thinking about doing it or exactly yeah and gps was sort of the was not number one choice at all yeah it's really interesting now it seems obvious you do it with GPS but it was not like at all they were there was several a lot of choices and gps was considered you know like not the favorite to win and so yeah the choices they were the favorite to do this with systems or location systems based off of the signal from the cell towers because the carrier's already knew where the cell towers were any any cellular system knows where all those towers are and that's how it looks as you roll them around they keep track of you so that they know which cell tower to connect you to next or if someone calls you they no they don't they don't ring you from every cell tower on earth and hoping to find you they write I would never work they they ring instead they ring you from the cell tower in the region that you are the cell towers already know where you are roughly and then they developed systems to measure the time of flight from the cell tower and so that was you know something you could do quite easily off the existing system so that was considered the the the logical way to satisfy the e-911 mandate but the trouble was you don't get such high accuracy like that because the cell towers don't have the precise clocks in them that the gps satellites had so with GPS you can get much better accuracy as so with the cell tower you get accuracy of just by knowing which tells how he talking to its accuracy of a few miles by measuring the time of flight of the signal from the cell tower you get accuracy of something like a hundred meters with GPS as everyone knows you get accuracy of about five meters but the trouble was the GPS didn't run in the phones the signal was too weak just took too long to get the signal and then because it took so long the gps chip was running at full speed trying to acquire the signal as match filter I was talking about before you use a lot of battery and so you'd run your battery down really fast so GPS was considered signal was too weak to go into a phone and the battery use was too much and so it was just a few entrepreneurs who kind of had the foresight to say well that's going to change because you got Moore's law working in your favor so the powers going to come down and a few companies started up to specifically to design chips to go in cell phones for GPS to satisfy this this manner with the hope that GPS would become the way to do location in phones for all phones and and that proved to be the case and so that that's kind of how GPS got in phones and then and and now it's it's there's so many use cases came out of that right like you know it just but everybody uses being a lot of traffic and you take a side road because Google Maps told you to and you I've had to saw up and off and you see the six cars in front of you we'll take that same side road that's a little d2 and they've all got each got their phone up and their phones are all telling booth same thing for the same way around the blockage yeah okay and now the company's worse we're starting to look even heal it packard at some point was trying to get into that business or well yes so he look in the cell phone business so let's go back to like how what companies they were so one of the first companies to do this was a company called snap track that developed some some chips for cell phones and then one of the other early ones was company called global okayed which the founders that were friends of mine and and I joined that company the day they got funding so I was one of the first four employees and and then we developed chips and okay so that's how you got from Agilent to yes so global locate so yeah so one of the founders charlie Abraham had been at Magellan with me so he left and was one of three founders of his company Global okayed and he brought me in and yeah so we worked at getting making chips that would have the sensitivity that they would work in phones and acquire the signal much quicker than traditional GPS and and we did that and one of the first so you mentioned Sheila Packard one of the actually the first smartphone to have GPS and it was a Compaq furnaces I was a it was called the iPAQ IP aq and I was I guess it was before I became the property of therefore and it was a it was a evolved from a compact handheld computer it was a this iPad so compact and Hewlett Packard had and this iPAQ 6500 was actually the first smartphone to have GPS in it and it actually had it had all of the function that almost all the functionality that modern cell phones use this was in 2005 so this is now 13 years ago and that phone if you look at it now it had a keep keypad or keyboard had a keyboards kind of look like some of the later blackberries had a keyboard had a screen it ran TomTom navigation so Google Maps in phones wasn't a thing yet didn't exist TomTom was a company from the Netherlands and Britain that did car navigation systems and so their software ran in that phone so you could you'd you could do navigation didn't know about traffic like like the maps of today but it could rat you and tell you estimated time of arrival based on on speed limits and so you could do that you could you could send someone your location if you were diligent about it there wasn't like now you could just go Shey location you'd bring up an app and we'd show you your position you could cut and pasted you got another app you paste it in you say Here I am you send that it was also it was a kind it was a real prototype of current things that if you if you had a someone sent you an email with an address you could copy the address and him go to the TomTom app and paste it in long you're all on your device on your device and so it was like the primitive it did everything that the current phones do but in a very primitive way and it kind of laborious sometimes you'd have to you have to get a piece of paper and write down an address and then type it in again and yet in a different app but you could kind of do the same things you can do that yeah that's interesting yeah we take a lot for granted now obviously as we look back so yeah I mean those days we used to talk about things like oh wouldn't it be nice if you if you had a calendar entry yeah that said you know bb-8 Computer History Museum at 2:00 p.m. for if it could know how far you are away and it could warn you up time to leave because the traffic and that's that's exactly the kind of things that happened now and so I'm caught up with it yeah so okay so that's uh oh yeah one question another was back there and at that time these companies that were designing these devices you say they were building their chips no that meant they were designing those chips and sending them out to be yeah brocaded and then yeah so in I think in all cases these companies were fabless semiconductor company meaning that they didn't have the fabs like Intel and there were mostly small companies doing this even the some of the established companies like Trimble and a shtick would have sort of fairly small to medium sized companies a few hundred employees back then so they would design the chips and then they'd get manufactured by big manufactures like TSM see that's Taiwan Semiconductor similar companies like that so innovators with ideas had the wherewithal to put together what they wanted and then move it forward yeah yeah okay say the the I we got to the magic transition 2005 mm-hmm now that's when they upgrade to GPS started is that well so you're talking about the when they switched off a selective availability no well when they made the plan for upgrading the entire GPS right yeah I forget what date that actually was I mean in the meantime there was this I'm sure some other people in this series to talk about this in in 2000 they switched off the the deliberate degradation of the GPS system and that that's because they did that made it possible for consumer products to have good enough accuracy to navigate you around the streets and so and so companies like TomTom came into being then and so that was 2000 so yeah and then then the GPS system they refreshed the set and this is sort of a continual renewal going on of GPS satellites last about 20 years and so as they as they start to get old they'll replace them with newer satellites and so in the mid 2000 they were going for what was called block to - block to our satellites and block one was the original ones and I never really think about that as a specific moment because it because there's 32 satellites and they get old so kind of thing that year there would be a natural progression of them yeah so at the moment for example is they they've already designed what they call GPS three so it's the third generation of GPS I think that's more what I was talking about yeah changing the signal structure so you could get more accuracy yeah yeah but even even going back now to it was quite remarkable and I guess so there's been a lot of controversy about it at GPS was started as a military system more or less like the Internet mm-hmm I think was always understood that there would be commercial civilian applications but they had never anticipated that the civilian applications probably Drive GPS now yeah well it's true it was always meant to be a it was developed by the Air Force right so it was meant to be a military system borrow from the original design that they did intended to be used for civilian use it wasn't like civilian use so sometimes it gets talked about you got GPS was a military system that blade about used for civilian use so it's meant as a dual use system from the very beginning and you can see that because the signal structure has this just this open code the so called c/a code which is which is public and then I had a proprietary code precision code which is military only so they had the stools you'll use from the very beginning but I think nobody anticipated you have billions of users by an I think you go look at some of the early studies and they talk about you know millions of users by two thousand and something and and now just in cellphones there's over two billion GPS receivers out there like Brad Park and sits talks when he lays that all out of how many GPS receivers there are and yeah every cell phone and yeah every cell phone many watches it's about every car yeah so we start out saying that you were the architect or one of the architects of a GPS so tell us what a GPS is all about now okay so a GPS in a nutshell is what makes GPS work in phones and the a stands for assisted and GPS on its own takes you you need a fairly big antenna like meaning about this size you know that something that's much much bigger than the kind of thing that you could stick on to a cell phone do you need that size intended to get a strong enough signal to be able to acquire the signal in the first place and observe this range delay that you're measuring from the satellite and so people who had some of the early hiking receivers might remember that they they either were fairly thick and I had this patch antenna in size about that size about the size of a square of chocolate hidden in there or it was a helical antenna and it would stick up so it's like thicker than your thumb would stick up out of the device and when smartphones came along that kind of thing just there's no place for it in a smartphone and the antennas inside a smartphone we really Tonya there's lots of them because there's all the cell phone antennas it's Wi-Fi as Bluetooth so the GPS antenna is a tiny little piece of metal it's hidden away in there and as a result the signal is very weak and a regular GPS receiver just won't get the signal that needs to to be able to measure the distance to the satellite so you needed something else to help the receiver pick up this weak signal and make measurements wrong and that's something else was this assisted GPS technology so that's what it is and how it works is you have reference stations that are fixed GPS receivers scattered around the world so you set up some network of reference stations and those things observe the satellite signal and then they send a hint over the wireless data system over the over the Internet to phone and say okay if you're at this location these are the satellites you can see and these are the signals that they sending and so it's it's completely analogous to like suppose you get off a plane and you pick up a rental car and you're driving in someplace you don't know and you want to find a particular station on the radio right so you just you just start turning the dial alright though nowadays maybe just press a button but when you had it if you had a radio station where you turn the value you'd start searching every now and again you might hear some crackle it'll be noise and then crackling you have to spend some time there and you search some more and spend some time and if the signals are kind of weak maybe you're driving through the countryside take you a long time to find that signal but now if somebody could tell you oh by the way I know where you are and I know what radio stations are there go to 97.3 and there's the radio station you looking for well then you'd go there even if as you got there the signal is weak because maybe you're driving through a tunnel or something if you knew in advance to go there you'll pick up the radio station and you'll do it quickly because you know exactly and then so assisted GPS is the analogous thing in GPS it's this network that watching where the satellites are moving and can tell you in advance which satellites overhead where you are and then technical details like what's the Doppler frequency that you should be observing of that satellite at that moment and so then your receiver doesn't search over all possible satellites and all possible frequencies it just looks where it knows the satellite is going to be in terms of frequency search and so it's quicker and then because it doesn't have to spend time searching around it can search deeper in terms of signal processing at that particular frequency for each particular satellite and so you get extra sensitivity and you get extra sensitivity in big numbers like you GPS receivers before assisted GPS had they could pick up a signal up to maybe 10 times weaker than the nominal signal that's normally where it is in the noise yeah so then where normally where it is they could get that and maybe 10 times weaker so then so if you walked under a tree but often they would stop working and if you look the people who had hiking receivers probably remember this when you first switched it on you couldn't stand and retreat I have to move in the open and assisted GPS and cell phones allows you to get signals there are a thousand times weaker than the nominal signal so the dynamic range increased by 20 DB over what was nominal layer which means it increased by a hundred times so you went from being able to get ten times weaker than the nominal down to what a thousand times and waited more yeah well once you compress the signal you've got 30 DB and you're still integrating on top of that yeah so I'm saying beyond that extra 30 DB so not it's I'm talking about beyond the nominal so this is on top a few yeah so ordinary GPS has some processing gain and and assisted GPS when I talk about this extra 30 DB that's extra on top of what you already get so the total gain from from when you first pick up the signal to when you observe it in a assisted GPS system is something like a million times you get you get that you you make that signal a million times stronger than what it was when it came into the antenna and that's done with more integration time yes it is more integration time at the particular frequency of interest and so these integration times is something like a millisecond but it used to be without assisted GPS you spent a millisecond at a whole lot of different search locations and if you can know in advance that 99.9% of those are not useful and you should spend all your time at one spot well then that's it's a it's a simple idea and then there's a question of design and implementation and and that's that's what was going on in the early 2000s okay so so for all of that then the there have been so many applications that and it's a that have come from that kind of technology night we wrote the textbook on assisted GPS applications big one being for airplane navigation you can we now make a GPS landing with the systems that are available and approved oh yes and in fact I think most aircraft have got order landing systems that that work off GPS they don't necessarily use them the whole time because they've got things like microwave landing systems in airports like if you fly into San Francisco Airport for example and a lot of other things you'll see those like at San Francisco is very obvious because there's a jetty going out into the and there's all these things put on there and that's that's that's part of the microwave planning system so there's a lot of basically sending up a beam into the sky that the that the plane flies dance there's a lot of infrastructure that they make use of but order landing from GPS was demonstrated quite some quite some time ago and up the road here at Stanford a group there and the pair Inga worked on that in that area and now then the continue to work to work with the FAA and demonstrated order landing quite some time ago yeah yeah I know he passed away earlier this year I know yeah it's a wonderful individual okay so so we've got you know the the expanding technology the the application saw from every conceivable thing you can think of to put a GPS system onto and and everything so how is the evolution of that impacting the design of the systems and integration augmentation minister that must be what you're heavily engaged in right now yeah so so you're saying how does how does what impact the design well the the different applications there so how did the education by a GPS chip that'll do just about anything you want now yeah roughly at the power you want so as you start integrating these things together though there's still a lot of challenges in terms of I would think how you how you could build devices and make you yeah well I should explain how big these GPS chips are that in the phone if you have a discreet chip it's that's just a GPS chip it's the the die size is some something like two millimeters by two millimeters okay yeah and that's that's our endless thanks to Moore's laws got so small and they used to be much bigger used to be 20 millions by 20 millimeters and so and then that's just to you know just Moore's law and if you work it out you think well over 20 years it would have got about a thousand times smaller because every two years it's cutting the you know your chip size is going in half for the same functionality but what's what's been happening in GPS is over the last 20 years is the chips have been getting smaller and more capable so twenty years ago the gps chip would be a big thing like this and it was capable of tracking 12 gps satellites typically that's what it was is a single frequency down to that then it was write a typical consumer chip back then there were two frequencies available but the typical consumer chip was single frequency now you'll have a two millimeter by two millimeter chip in a phone and it'll it'll be capable of tracking over 100 satellites and you think well why do you need 100 is because you've got you've got GPS now there's the russian system called GLONASS they have 24 satellites up there there's the european system called galileo and they have 18 to 22 satellites they just launched four more a couple of weeks ago so they have 18 operational and 22 in space is the chinese system called bado which has 30 satellites and in his regional systems such as qz SS Japanese system for satellites and IRNSS a Indian system which has about six satellites yes way so you so you have well over 100 satellites available and you now have dual frequencies signals available to civilians there's signal called l5 which is on GPS and Galileo and bado and QC SS so in four of those system basically all except flow Nass you have this l5 signal and they are now chips available for cell phones that include all those satellites I mentioned plus the dual frequency so you can just track a whole other more signals and you and then the benefit of that is you can start getting better accuracy in cities and people probably use do this as well if you and and so so back to your question like how is the use case driving the design so when did not possibly the number one use case of GPS that people are used to is navigating around in new cars and nowadays calling uber or lyft calling a rideshare service right that's that's two of the things that almost everybody with a cell phone does at a certain stage and anyone who lives in a city is probably had this experience that they will call up uber or lyft and the GPS will think they're on the other side of the street and that's because the signals are blocked by the buildings and and with these extra signals and with the extra satellites you can solve that problem and so that to solve the problems you know driving the design and is just now drive that just the first phones are coming out with dual frequency GPS in it this year and so that's a big that's a big step and that's part of what I was talking about before the kind of technology that used to be in these big receivers that was sold for industrial commercial use and is now making its way into into cell phones and and in a way even even more capability than we used to have because of these new satellites and the new signals yeah I know I mean I know from my own experience the multipath has always been the biggest problem that we've ever with a yeah EPS receiver cuz we put one up on the roof at a company I'm just working in with all of the performance was terrible yeah and in the head and in the city with one of the before before we had I I before we had assisted GPS you just wouldn't get a position in a in an urban Canyon in like in in the in wall street or you know in any way in Manhattan in any kind of narrow Street in Manhattan or the financial district of San Francisco for example if you use GPS in one of those narrow streets just you wouldn't get anything because they were they weren't enough satellites in view directly overhead and the satellites were blocked by the building the the signal that bounced off other buildings was too weak so then with assisted GPS we added all of this processing gain that I was talking about you have all this extra sensitivity so now you saw tracking the satellites in the urban Canyon but what you're tracking is something that is being reflected or for another building and so now the accuracy gets bad do you know where the buildings are I mean do you so we make use of the calculating where those signals though so Google knows where all the buildings are so this is something that that's if you if you look on Google Maps for example you can see you can see 3d models Serling's yeah no I've seen that technology so yeah so that that information is there and that together with these new signals means that this problem will will be solved in in the future this problem of being on the wrong side of the street yeah and and that's yeah it does sound like a more direct approach than trying to unravel multiple signals on a single in a single channel but yeah increasing the number of delays you look for for example yeah well when you get these reflections you you you might get many reflections from one satellite yeah and so you just can't kind of figure it out just by measuring the channel yeah you you have to have some knowledge like of where the buildings are and so that's that will that's part of how you would solve this problem it's still still not a solved problem right now if you take any phone and you go stand in the financial district of San Francisco was some similar similar dense area and switch on the GPS the very first blue dot you get is probably gonna be on the wrong side of the street anytime I talk to a large audience about this I was people who's had blue dot on the wrong side of the street everyone puts up an yeah yeah in fact I had the experience with GPS telling you on the wrong side the road and so yeah okay so that that technology it's it's coming along I know the other was one other question that you hadn't talked about this a little outside but one of the major impacts of GPS to is goal blindness for iguana during for example that very signal that little ten meters that you're correcting for the honest very care can be tracked in and it can now get maps of the entire ionosphere over time scales of inside of ten to twenty minutes or so rough dated yeah maybe I should explain that a little bit we kind of skipped over that I haven't talked about that explicitly but when yeah the whole system of GPS is working by you have an atomic clock in a satellites it sends a signal with a with a time tag on it so she has here's a signal and this is the time and then sometime later about 70 milliseconds later by the way it's this long that's 70 milliseconds if I tap on the mic you hear that really but in that short amount of time the signal travels to you on earth and your clock inside your receiver your phone for example will will time tag the signal when it sees it so your GPS receivers got its own clock and in the difference in those two times tells you about the range you just multiply the time difference by a speed of light and that would be the range if it was all vacuum between you and the satellite well it's not all vacuum it's mostly vacuum the signals traveling from the satellite through vacuum of space then it hits the ionosphere which is a area outside our atmosphere where there's a lot of free ions and free electrons and that change that changes the propagation characteristic of the signals and slows it down and then the signal hits the air yeah the troposphere which is where we live that's the air so it's down even more and so you get extra delay that is not accounted for in just saying multiplied by the speed of light because when we say the speed of light we really mean the speed of light in a vacuum speed of light it's different in a in a medium such as air so you can in in a typical application like in a phone will just compensate for that with fairly well-known models of how much that delay is and but to do more precise work what you can do is is if you have a if you have a device at a known location the satellites at a known location then you can actually measure the exact delay caused by the ionosphere in the troposphere and you can infer properties of the ionosphere from that so GPS instead of you instead of trying to illuminate the ionosphere area you can use the GPS as something to observe the silence very that tells you how many what the characteristic of the ionosphere is at any particular moment and it changes during the day as at night the ionosphere is very quiet there's not many free ions and then as the Sun rises over a particular area it'll the energy from the Sun will cause more ions to be freed up and you get a peak of an aspherical every day at about 2 p.m. local time and then it goes off and then this changes as well with solar activity so people know about sunspot cycles and the Sun being more active and it's a big deal because when you have a tremendous amount of solar activity you get the lot of every now and again I'm meaning like every few years you can get enough activity in the atmosphere that will disrupt satellite communications so your TV might not work so well any other form of communications through satellites and you can measure this and then actually predict it with GPS but I mean you're using the for the ionosphere you're using the dispersive nature I mean the frequency dependence would allow you to correct and back that error out on a single channel yeah routinely in fact that's why they had two frequencies for them yeah that's that's true that's one of the that's the original reason for having two frequencies was to so this ionosphere delay as you mentioned it's a function of frequency yeah so if you can observe the relative delay on two different frequencies you can figure out the ionospheric delay that's that's not such a that's not really the benefit of two frequencies these days because you've got internet connections to just about every device especially your phones so you can just find out the ionosphere ik delay if you care about it that much if you really care about high accuracy you can get Corrections for it from some other devices like a GPS reference station that's measuring it the more interesting thing I think is is the science that comes out of it where instead of trying to measure the ionospheric delay to remove it which is what people do for high accuracy is use a station receiver that's at a known point to measure their as fire delay and that's it's work that at the University of Colorado at Boulder they do there and there's a professor there called Jade Morton who's kind of one of the leaders in this area and she actually showed that you could measure the effects of so when they were doing some nuclear testing in North Korea the effect of the nuclear blast disturbed the atmosphere you could actually measure that with GPS stations nearby like in Japan and South Korea could measure the disturbance in our sphere linking out through the air and tell that a nuclear blast had happened nearby so so that's like who would think we would have thought that even though you know we were saying when they first built GPS they were intending that a never anticipated 2 billion devices that we will carry around I'm sure also doesn't anticipate you could make such sensitive measurements of things like the atmosphere and and infer things like nuclear blast that's very interesting at view you would think of that as the way to measure it would be to make comparisons to known distances rather than to use the frequency dispersion the first yes EPS system why well well the first GPS system was but the military use of the GPS system was for the airforce they imagine some F was plane flying far away from from its home base it would need a stand-alone method of always a jackhole Bashar was the architect of those first le models that up with the ionosphere into it so yeah we I work with him in the in the nineteen around 1980 I used the GPS receivers and weighed a hundred pounds and cost $100,000 really make to frequency measurements so we could check his model right and now so that model is now in use and in every cell phone I think that's the standard models of it's kind of nice thought that that's like at any moment there's millions of cell phones of running that model to model the ionosphere and remove it to the at least the kind of accuracy you get from itself yeah it's probably five meters or so yeah I'd get the actors say that model but right and the technology's interesting okay so yeah more about then you're as you look at this this GPS now in your own interest in where it's taking you is how do you see it what are the new technologies that are really going to drive it augmentation seems to be like in marrying GPS with uh national inertial navigation as a as a means of really changing the dynamic in the near term if you had a good enough inertial system GPS would become an augmentation rather than the other way around or how do you see that right well that's so yeah that's something that's happening array is the read it's in in cell phones for example you you you have an inertial system in in and all cellphones have an accelerometer in there and all high-end cell phones have rate gyroscopes it's alright yeah so any any flagship any top-of-the-line Samsung or iPhone or pixel or and any one of the big name phones that the other high-end phones all have a rate will rate gyros actually have three they've one for each axis they have rate gyroscopes so the phone knows when it's rotating so for example people experience this all the time when you when you take a panorama picture and you rotate the phone it knows how it's rotating yeah so just to establish its orientation is the primary user that well that's where that's where the accelerometers come in so so it's it's kind of a good Trivial Pursuit question what was the first navigation sensor in a smartphone and the answer is an accelerometer and accelerometers were put in every phone there are in every smartphone is accelerometers that they were put in in the first place that when you put your phone down on the table it could sense that it was still because if you hold the phone in your hand no matter how little coffee you've had to drink you actually shaking at at a level at the accelerometer can measure when you put it down in the table it's very steady and so to save battery as soon as the phone's put down in the table it goes into a different mode and that's and so that so the accelerometer was the first motion sensor to go in any phone and all phones have them as for that reason and then they put in three axes accelerometers so that your smartphone could switch from portrait to landscape when you flip it you know so you can see that then they add a gyroscopes for things like sensing orientation and camera use and those two things together being combined with GPS so when you're driving in your car they will integrate the GPS and the accelerometers and the gyro sensor so one thing that it's if you know is watching this actually want to learn something useful one of the things you can do if you're navigating with your phone in a car and you put it in the cup holder and leave it alone or put it on a - mount so that's being held still with respect to the car the software in the phone will sense that it'll sense that when the car stops the accelerometer is stopping a little and it'll realize that it's it's being that it's somehow mounted so you mounted means either on a something holding it on the dash or even just sitting still in the coupled as long as it's not flopping around the software in the phone will sense that and will and then they will combine the GPS with the accelerometer and the gyroscope the thing up on the dashboard you kind of have a better user experience and if you hold it in your hand or if it's like laying around and sliding you're implying that the positioning and the cell phone itself is being tracked on the on the map or is it using a scheme where it knows you're on roads in the first place and take some of that jitter out by well well there's layers of it those if you're using the GPS it'll use the GPS so if you bring up Google Maps for example it'll use the GPS then the software in the GPS chip will will integrate the GPS position with information from these inertial sensors if the software determines that the phone is not being held in that in hand if it if it's if it's mounted back to the car and then that position will will then be combined with the map information okay yeah right sorry you really don't want to see yeah the thing jumping around on the road the way the positions are coming yeah so back to your question about what what do I see in the future will this use of inertial technology is is moving into handheld use cases now so and when you hold your phone in your hand you can see it's a it's a more difficult problem your hands moving around all over the place right as you walk and so now you can't do the kind of integration of the accelerometers and the gyros that you can do in a car with the phones nice and stationary and I think in the car the car behaves very well you turn the corner that the gyro changes very steadily if it's in your hand you turn a corner will you your hand might be swinging as you turn so it's much harder to use inertial information for navigation of a pedestrian but that that is something that's where companies that do location like like Google and they are bringing that kind of feature into phones and might have seen in recent Google i/o they they did a brief preview of walking navigation and they show what that might look like that used augmented reality lifts up the phone and sees through the camera so it can sense what buildings are around it and you can actually navigate as a pedestrian the way you would in a car tells you turn right here turn left there and actually show you and on the video screen like a little arrow right at the building go around this building yeah I saw a beautiful application of that was a person that was impaired mentally by using they usually have to have somebody walk along with them right a GPS substitute for that actually improved their mental acuity right just by virtue of the fact they were doing it themselves you know it felt like they were doing it themselves so well I suppose the question this way you're still on the market there are commercial GPS receivers for diagnostic applications is very monitoring like novotel and you know sub 10 trio right and those things cost a few thousand dollars no yeah and so how what's the differential why are they so expensive for doing what they do when you can put it on a chip and get meter accuracy in a cell phone somebody not closed the gap yet I might I'd like to buy one for a few hundred dollars yeah right well so what's so the companies that make these kind of scientific grade gps's they also make their own chips but they make them it's very small volumes like a few thousands so they have to pay a lot more to get those chips made and then those chips are designed to give the very high fidelity carrier phase signal and the carrier wave of GPS is only 19 centimeters long so only that long and in those chips are designed so they can measure the phase in that carrier wave to a fraction of the wave itself so that means a fraction of a centimeter now the the GPS in your phone is optimized to have low power and high sensitivity and so there's the gap right now that the GPS new phone has not been designed to measure this carrier phase very very accurately bad as the capability of the chips increases and that that just keeps happening because you a lot of this is digital signal processing and so with each iteration of Moore's law you in the same size chip you can do more processing so even the carrier phase measurement is done with digital signal processing so the chips and the phones are getting closer and closer yeah so you see that gap closing yes so so in a way it's a it's already happened and they used to be so that there's these high-end receivers for measuring things like ionosphere and then there's things for measure for doing surveying and in the middle they used to be a class called GIS receivers GIS is geographic information systems and it was a class of receivers that cost also a few thousand dollars but single thousands not tens of thousands and they were kind of big receivers they would measure the carrier phase and get submitted accuracy maybe but not centimeter accuracy and they would use those four for mapping things people go out so like a facility wants to map all the fire hydrants for example like a map each fire hydrant to 20 centimeter accuracy for example well the phones are good enough now that a lot of organizations that do this mapping just use the phone now even though maybe it's not 20 centimeter accuracy but it's good enough and it's it will get to that 20 centimeter accuracy as these carrier phase measurements become available and and they are available from phones on and the Android operating system makes the raw measurements available so there's a public API on Android that you can get these measurements out of a phone so people can take that and it's off those apps out there that take that take the measurements from the phone process them and give you an accuracy at the sub meter level from your phone today mm-hmm so you could yeah but that's still single frequency measurements yes as I said dual frequency incoming yeah and so it's just gonna get better it's good it's just gonna get more and more accurate yeah okay so ice-covered uh that territory very nicely in terms of they say think about GPS in general I guess the the miracle that was a system that was never intended for the use that it came to be and and it's free the users don't pay a fee for using GPS and so far we're putting up with that yeah so I yeah it's it is an amazing thing I think it's it's it's um it's it's up there with with the Internet as one of you know America's great gifts to the world maybe maybe America's last great gift to the world yes you know it's really so now there are many other systems as I mentioned there's but they're all copies of GPS and yeah it's certainly as I mentioned earlier it was meant as a dual use military civilian system from the beginning but the number of uses as far exceeded anybody's early predictions so is there anything in any of the systems outside of GPS that are functional improvements over things they're in GPS are they all pretty much interchangeable and you say there's a sign no signal architecture that actually with him proven era on your ability to acquire the signal or integrator right yeah beyond like l5 for yeah so what's interesting is the GPS GPS system was designed in the 70s right so the signal architectures relatively simple by modern standards of what kind of signals people use in communication systems now and so the gap so so you know so GPS has of the different systems that are out there the the the GPS system itself the u.s. system actually has the simplest signal architecture and the Galileo system from Europe has the most complicated and in theory you can get more accuracy out of the more complicated system but in practice the simple one often works better because the the signal is especially it for consumer users like cell phones because the signal is so weak as we discussed earlier that some of the stuff that works in theory it's a more complicated signal that would give you slightly high accuracy in terms of measurement of time of arrival just doesn't it just takes so long to acquire that signal because it's it's a much finer signal that the the more simple signal actually often works better in practice so I'm a I'm a big fan of the simple GPS signal structure that was designed in the 70s and you can look at it in theory and see why a more complicated signal structure would be better for something like these industrial use cases this scientific use case we have a big antenna receiver that's outside and gets the full signal but for the use cases for the masses you know those those industrial receivers represent point one percent of all gps is or less for the use cases of the ninety-nine point nine percent you're dealing with weak signals because of small antennas and I don't think you actually benefit from a more complicated signal so so sorry wait say you what you know the question is like the other are the things outside of GPS it would help GPS the answers hmm yes for scientific use but maybe not for for your typical consumer use case yeah that's a powerful perspective something to keep in mind so when we get to a point to that you've had just listening to you talk you have a passion for teaching and and especially working with students and stuff so um how do you bring the GPS and how are you teaching it and are you having success and getting interesting okay so yeah well I teach a class at Stanford on GPS is a one quarter graduate class on GPS I've been doing that for six years so that's that's you know that's how you teach grad students about this and we go through all the the main parts of how the system works so signal structure so it's a signal processing orbit mechanics so how you work out where the satellites are receiver designs a little bit of hardware so it's it's a course that covers sort of broad not very deep each of these areas you can dig really deep if you want to and that this is a broad course and so that's how we teach it in terms of introductory system and then there's there's quite a lot of fairly well-established places if people want to learn GPS and you're not a student at Stanford there's a Coursera course that pair anga who we mentioned before and I did we developed it and it's available from Coursera ins on YouTube so I've just got a YouTube and look up GPS introduction to GPS claw you'll find several I think it's 53 or something I remember it was the same number of lectures as they are episodes and Breaking Bad so I was very happy that that turned out like that and there's that there's that many lectures introducing GPS from and it's that's meant for people who really don't have any background and want to learn quite a lot so we go through that so that's that's another place and then this Institute of navigation which is the u.s. so if you like to and I do that's the the US Institute for navigation that that is right the premier Institute for for GPS in the world I think they have conferences each year and they have tutorial sessions before the conference's so that's a great place to learn it's one of that's when I first came to the US and joined abscess back in the 90s that's actually where I although started learning about GPS so there's that and then in Europe is a similar thing through eisa European Space Agency they sponsor a summer school for grad students so anyone in Europe that's a really any students in Europe want to learn about GPS that's a great way to do it there's every summer they have a two-week so does it stimulate interest in science mathematics and you know if you if you get into this you can yeah GPS are these these things are all targeted at University students the graduate students or well on their way you know they've already really really in the STEM program so yeah it's it's it's a good point I don't know so anyway those are the big areas where you can learn about this is that there's every year at the the Coursera class is up there on YouTube so it can find out anytime and in these these conference these things that go along with the conferences happen every year so the the IO n conference coming up next month in Miami so so and it's always in September and so they have these classes before the conference some of them you have to pay for and some of them are free so that's a great way to learn about it and then for student that's so anyone can attend that conference but and for students in Europe these summer schools are very good but for to get young student you like middle school students involved in white damn this this would be a every middle school student knows about GPS that's and it's not I've done a few like classes for students but on an ad-hoc basis just try to get them interested in that well it was yeah in my my own kids were in middle school and my brother's kids at various stages I went and taught the kids about navigation and GPS and it was great fun actually in here in San Jose we got middle school students to learn how to work out their lat/long just from the Sun and just using a stick and a shadow and a clock and a watch they actually worked out their latitude by it by understanding how the how the Sun moves so and they loved it but it's I don't know being like an established program for that and I think would would be a great thing that one day once we've solved all the problems in the cell phones yeah there's one one thought that triggered on that on the orbit the people probably don't appreciate it but I mean it's the GPS satellite itself has to know where it is very accurately yeah and that has to be updated it's good for about two hours right now they actually had to modify I think I'm correct and this the orbital prediction codes that were used just for keeping track of where satellites are like NORAD to put extra terms in those pseudo pseudo moment equations that keep track of the position so that the GPS had an impact now in fact now I think that probably if you pull out a piece of code to put in a satellite number and say where is it that's probably it's probably calculating it with an upgraded set of codes that came about because GPS yeah well that's right what you're saying is that for the GPS and it's it's interesting it's complete backwards of how somebody that doesn't have any knowledge of this they talk about it in a way they talk about it backwards people say GPS tracks you and it's actually the other way around that the GPS satellite has no idea where you are it's just up there and it's not broadcasting a signal and actually technically you're your receiver your phone for example it learns where the GPS satellite is you so to make it work to make the blue dots show up you track this the GPS okay and yeah and you have to know that so you have to have that satellite orbit accurate to whatever accuracy you're hoping to get to better than what if you're hoping to get five meter accuracy and we've discussed these ionospheric and tropospheric errors and there's multipath errors from buildings on the ground so whatever error there is in the satellite orbit that's going to get in there as well so typically you want that error to be down less than a meter and it is and it's it is remarkable and at all starts at for GPS it starts with the US Air Force and in the similar organizations for Galileo and so on each system has its own control station and there they have people who do the modeling and they predict to within a fraction of a meter where the satellites gonna be at every moment and I put that in normal model and when you think of so satellites moving at something like 3 miles per second or you know 5 kilometers per second so to know where and it's way out there in space 20,000 kilometres out in space so to know where that thing is to whose list to better than a meters a remarkable feat it's every violating the uncertainty principle no but it's it's it's it's really something and then added apart from the orbit what's really cool and a lot of people don't know about but they should know about is that because you got an atomic clock up there that's so precise and sending the signal and you measuring you use you're subtracting two times and multiplying by the speed of light even an era of a nanosecond shows up as one-foot era in your position so you have to have those times right to order of a nanosecond and because of the satellites moving so fast and it's far enough away that the gravity is much less both special relativity and general relativity from Einstein are big effects on GPS and so when you when you're trying to when you're sitting in traffic and you're looking at the blue dot on your screen your phone is busy running calculations to take out the effect of general and special relativity on the clock that's on the satellite and so I think most people don't realize that that they're running a little every time they use a GPS it's like they're running a little experiment proving or using Einstein's equations and those equations actually running in the software in the phone every single time they get a position so that's a very cool thing yeah I think we should turn this interview around let you say that at the beginning yeah that's a very very powerful thought so okay we feel ya we've covered the ground pretty well it's been a it been a pleasure to go through with you and it's been very informative so everything we've left out thinking want to add to the historical document yeah that's well I mean you you took us through very methodically thank you I I guess I guess the interesting thing is they're kind of looking to the future right sure like what's what's gonna happen next and and I suppose we you know I'm I think in the short term we're gonna see just more of the same with is more signals people who use GPS should know that already today it is you it's almost impossible to to actually use GPS that the u.s. system alone like when you're using an phone the chips on the phone are all tracking multiple satellite systems so yeah so you you're always using multiple satellite systems already today it's actually you actually don't have the capability to go and say just use the GPS system or just use the Galileo system in most devices so it's already a multiple constellation system which makes the thing more accurate and as we talked about there's a second civilian frequency now available on many of these satellite systems including GPS and Galileo and that's making it more accurate so we're gonna see more accuracy will solve this problem of being on the wrong side of the street in urban Canyon and I think you'll you'll start to see things like like mapping survey level accuracy coming out of phones and what that'll mean for use cases is kind of it's hard to say right now like what if your phone could measure your location to ten centimeter accuracy what would you do with that and it's hard to say right now because it can't do it so people don't write apps to make use of that what but once it becomes available I can see people starting to go oh well if one like things like golf come to mind you know you could very precisely measure distances and then some other things might come to mind and people might map out you know if you're doing some landscaping you know and instead of getting a surveyor in to map out your yard you do it yourself and things like that so these kind of things are gonna happen full for your average consumer just based on the evolution of the technology in California you could check your neighbors position very accurately and see if they're moving relative to one yeah.well thing that you can see today from the scientific use of GPS if you go to Google you nav Co tectonic plate shift or something like that similar woods you can they you can bring up a map that shows you how anywhere that's a California for example but anywhere else in the world you can see how the the earth is moving and we here in Mountain View we're moving at about three centimeters per year significant amounts if you bought a house 10 years ago it is now about one foot further north west move we're moving up towards Alaska so your entire house was moved a foot since you bought it if it's ten years old now the thing is everything else around you move the same so you didn't notice but compared to the they're the competitive Nevada for example the whole of Northern California the whole of California coast is is moving Northwest at that speed and we know it you can actually go look on a daily basis if you feel like it at this site you have co University navigation consortium so if you google you know if Co and GPS and crustal motion or some similar words you you can discover this yeah okay I think again that's been a pleasure and a very informative discussion great well thanks very much
