Explained - GPS/GNSS Corrections

Positioning and correction services improve default results — from meters to centimeters — enabling growing markets for high precision applications. Which should you use for your specific needs?

Consulting Editor | AEC

GPS (and now GNSS) are already cool and have been since their inception. It has become a global utility for positioning, navigation, and timing (PNT). And now, with an entire industry devoted to advanced positioning and correction services, high precision GPS has become democratized. Centimeter positions can be derived, in real-time, nearly anywhere in the world. This new reality was decades in the making. Fast forward to the present day, where we have so many to choose from.

The “glass floor” of high-precision satellite-based positioning and measurement has been broken. With the progressive advances in the hardware, software, terrestrial infrastructure, and multiple constellations of navigation satellites themselves have come advances in methods to augment the signals you receive with additional data. Such data takes raw, uncorrected navigation satellite capabilities from default precisions in the range of several meters to sub-meter, and even centimeters — in real-time.

Understanding more about how these various technologies, methods, and services work, what precisions to expect, and respective hardware, software, and communications prerequisites for each, users can make informed decisions as to which to choose for their field applications. From agriculture to marine, construction, surveying, mapping, and asset management, positioning methods and services are suitable for each.

What are positioning and correction services?

When using any GPS (or GNSS) receiver, or chip in our phones, we get standard observations from each satellite, multiple frequencies, clock corrections, and orbit data. This basic data, by default, enables such devices to yield rough positions in the range of meters. The data received from the satellites alone is not sufficient to mitigate sources of error that limit precision. Several meters may be good enough for many applications, like consumer navigation and location-based services, but not for others.

The error sources include delays in signals passing through the ionosphere, troposphere, rough predicted (broadcast) orbits, and resolving the time component. Remember that the satellite carries an atomic clock, but your receiver does not.

Corrections and augmentations add more and higher accuracy data to help mitigate multiple error sources that otherwise limit standalone GPS and GNSS results. These are derived from terrestrial tracking networks, to improve clock and orbit “products”, and may also include data from global, regional, or localized ionospheric and tropospheric models. Various augmented data can be delivered via radio, the Internet, or communications satellites. Delivery of augmentations by public or commercial generators of this add-on data is broadly referred to as “positioning services”. These augmentation methods and services fall into two key categories, and with some hybrids of both:

Differential: Observation Space Representation (OSR) leverages differentials between the observations of the user’s field receiver (usually referred to as a “rover”) and fixed base receivers (or networks of bases). Differential solutions from single bases include DGPS and real-time kinematic (RTK). Networks of connected bases have been established spanning entire regions, states, and countries, to create and deliver network RTK (NRTK) corrections. These networks are more commonly referred to as real-time networks (RTN). We will examine each in more detail later.

In each of these OSR methods and approaches, the differential between base/network and rover helps determine how much sources of error, like ionospheric and tropospheric, can be modelled and (for lack of better term) “corrected”.

Precise point positioning: State Space Representation (SSR) implementations and methods use an improved clock and orbit data and delivers this to the rover, sometimes adding in data from global or regional ionospheric models. So, instead of differential data derived from localized observations, SSR data represents modelled “states” of resource data at the moment (or valid for sliding windows of time). Precise point positioning (PPP) is a common method for applying SSR data. PPP began, as did differential methods, via post-processing, but now is more commonly applied in real-time.

“The differential solutions had been pretty well matured when the development of real-time PPP began,” says Tom Mackie, an engineer whose career began in pre-GPS maritime navigation but has since been deploying and supporting each progression of GNSS solutions and services for Trimble. “The biggest error or the bigger pieces in the error budget, are the iono and tropo. The localized augmentation systems like DGPS, RTK, and RTN, take that into account by setting up a localized correction and infrastructure, which ultimately minimizes the errors. This brings the spatial decorrelation (parts per million error) down based upon the separation of a base to the rover.”

“So now, with the concept of the wide area, or the PPP solutions, you’re not having that localized base to give you definitive information,” says Mackie “PPP has to send out global models, plus the better clock and orbit in the broadcast data package, and the rover uses those to resolve the best solution given its position on the globe. I’d say it’s minimizing systematic errors based on some messages in the data, but they’re global messages.”

How we got here

The original charter of the inaugural global positioning systems, namely the TRANSIT (Doppler-Based) system, and then Navstar (commonly called GPS, the Global Positioning System) of the United States military, was to serve global navigation. This was a quantum leap forward from centuries of compass and sextant. The precisions envisioned initially were in the order of 10 meters, with 95% global coverage, and the system would also provide precise time for applications like telecommunications and network computing.

“GPS was wonderful, all of a sudden you could get a position anywhere on the earth within a few meters; great for navigation,” says Tom Mackie “And then people got clever about how to post-process, and differentially you could get geodetic grade results — but with a quarter of a million-dollar set of receivers.” And while this was a boon to geodetic surveying and maritime navigation, there was a “glass floor” for other applications in terms of precision, costs for equipment, the time needed to derive such precise positions, and the expertise to do so. This changed rapidly, said Mackie, “There soon a lot of lower cost, but less precise consumer and mapping grade units. These early recreational, mapping, and navigation units brought global positioning to the masses — if only in ranges of meters to tens of meters at the time.”

Despite there being initially a deliberate degradation of signals, GPS revolutionized civilian applications like marine navigation and geodetic surveying. For instance: pre-GPS, establishing geodetic control via legacy terrestrial methods could take days, weeks, or months. The global geodetic infrastructure consisted mainly of passive control marks set in the ground, and propagation could involve time-consuming celestial observations, long traverses with analog instruments, long spirit-level loops, and wide regional or national arrays of triangulation stations.

Before the advent of real-time correction methods and services, post-processing was the only way to derive better than default precisions. “Using three or more receivers, usually over a period of days in an observation ‘campaign’, data was collected simultaneously,” explains Mackie. “Some are set up on known points, and together with the data from receivers on unknown points, this “static” data is post-processed together to yield positions that could be as good as a few centimeters. This method is still used for some precise applications, but observation times can now be much shorter with more satellites.” GPS also enabled the refining of the global model of the earth, tightening the reference frameworks.

A huge step forward was when the US government, formally recognizing the economic and public safety value of high precision for civilian uses, disabled Selective Availability (SA). This time-varying obfuscation of the signals (codes) could add 50 meters in horizontal error and 100 meter in vertical. When permanently lifted in 2000, uncorrected GPS leaped to the 3-5-meter range. This also enabled further gains for the developers of more advanced high precision solutions and techniques.

Further, soon to come was another revolution in real-time positioning — sub-meter or even centimeters in seconds — using both L1 and L2 signals (code and carrier) from GPS. In the mid-1990s, the German firm Terrasat while working with Spectra — now both are part of Trimble — launched what is widely viewed as the first commercial real-time kinematic (RTK) base-rover system. This and subsequent methods and services require external augmentations, or “corrections”, provided by local or global ground infrastructure, radio or Internet for the broadcast of the corrections, and later some solutions utilizing dedicated satellites to broadcast augmentations.

As the Beatles song goes: “things are getting so much better all the time”. So here we are in the present day with many positioning services — a rapidly growing global industry — bringing higher precisions to marine navigation, precision agriculture, surveying, mapping, asset management, machine guidance for construction, robotics, improved positions for consumer applications, and the latest boom autonomous vehicles.

Vertical precision

It is important to note that there is a universal Achilles Heel for any GPS/GNSS solutions, in that the vertical imprecision may be as much as double that of the horizontal results — it is simply the nature of the beast. It has to do with the strength of the geometry: all of the satellites are above us and not below, so the strength is wide to the horizontal, not so much so for vertical. So, we must keep in mind that if the application requires high vertical precision, we may need to step up a level or two in methods or solutions chosen.

Better antennas

It should be noted that one of the other keys to higher precisions is better antennas. There is always speculation that some-day a phone will be able to get centimeter positions. But not without a better antenna. The chip and included antenna are tiny and buried in the phone. Where would you measure the centimeter to on the phone? There may be many advances ahead in chip design, adding more constellations and signals, connecting to correction sources, etc., but without a better and more practically an external antenna, there will be some hard limits on just how precise a phone can get.

Devices that use more precise methods and solutions like RTK/NRTK and PPP come with high-quality antennas, built-in or connected by an external cable. These are not inexpensive components and may have features like ground planes, good for working in high multipath environments. A cheap and tiny antenna can defeat the benefits of corrected GNSS.

GNSS – GPS is no longer alone

Before we go into the various types of augmentations and correction approaches, it should be noted that the term “GPS” may need to be dropped as the generic shorthand for satellite-based navigation, positioning, and timing (PNT). GPS (Navstar), the first of the systems, and in many ways the gold standard, has now been joined by other global constellations of satellites. Europe’s Galileo, Glonass of the Russian Federation, China’s Beidou, India’s NavIC, Japan’s QZSS, and many regional augmentation system satellites (SBAS). GPS is being modernized with more signals, and several of the newer constellations started out of the gate with 3-5 signals or more each.

Combined, these systems comprise a system of systems: Global Navigation Satellite Systems, or GNSS. Get used to the term… It is commonly used in the high-precision end-user communities, a shorthand for anything that implies multi-constellation, even if it is just GPS plus one more.

The skies are crowded, almost four times as many navigation satellites as in the legacy GPS-only era, and many more signals. One of the biggest benefits of this is that we have more satellites in view than ever before, which especially comes in handy in sky-view challenged environments like under light tree canopy and in urban canyons. The new signal structures and the ability to mix-and-match from multiple constellations/signals has been a boon to receiver and applications R&D. New players in the GNSS industry are building off decades of R&D, lowering costs, making operations simpler and more reliable. Now, it is “centimeter GNSS for everyone”. Provided of course, you employ an external augmentation or connect to a positioning service.

GNSS positioning methods and solutions

Uncorrected: Uncorrected, GPS/GNSS is quite imprecise and can take time to come up with a position. GNSS does ranging using the satellite signals. The simplest and least expensive GNSS receivers, like those in phones, early recreational handheld units (without any augmentation turned on), can get the basic information directly from the GNSS satellites, like a broadcast ephemeris (orbit data). Such orbit data indicates where each satellite should be in relation to the earth. The satellite is, therefore, a known location, at a predicted point in time along its orbit path. And the time it takes for the signal to reach the receiver is used to gauge the distance, a vector, so to speak, from the satellite (around 20,000 km afar). The receiver does this with four or more satellites in view. The convergence of those vectors gives us a location. An oversimplification, but that is the basics.

But there are a lot of sources of error in that equation. Time is the key. The GNSS satellite has an atomic clock onboard, and it is synced up with the respective GNSS ground control network, which in turn is synced up with a global precise time service. The clock in our receiver is much more primitive and relatively imprecise. GNSS signals provide time, clock corrections, and orbit data. But other influences cause delays in the signals, namely the ionosphere (strata of ionized particles) and troposphere (lowest region of the earth’s atmosphere, weather). The ionosphere in the upper atmosphere ranges from about 60 km to 1,000 km; it can vary quite a lot between day and night. Most augmentation approaches concentrate on addressing these four key sources of error: clock, orbit, iono, and tropo.

Even with the most rudimentary clock and orbit data, a simple GNSS receiver can eventually derive a position in the order of a few meters. There can be a delay in getting to the best position a receiver could be expected to achieve. This is known as “time to first fix” TTFF. Once a receiver “knows” its rough location, subsequent positions can be derived much more quickly. But if you walk out of the building with the tiny GPS chip in your phone, and if there is no augmentation, you may notice that your position is wildly off for a short while.

Pluses: Free
Negatives: Imprecise, slow TTFF
Expected Precision: 1m – 20m

Cellular Carrier Augmentations: There is great value in at least having positions for phones being within a few meters, not only for personal navigation, commercial applications, and especially public safety but also in getting to those precisions rapidly. For example, the Federal Communications Commission (FCC) of the US has mandated that all cell phones sold must be capable of using a simple augmentation that is known as “AGPS” (assisted or augmented GPS). This to overcome the TTFF delay so that semi-precise locations are available more rapidly for public safety responses. There are similar systems in many countries.

If the AGPS is turned on, the phone can grab some data from the cellular network to speed up the TTFF. Cell towers have GNSS receivers running that know their location precisely at any point in time, and they have up to date clock and orbit data that is transmitted to the phone’s built-in GNSS receiver. There are some implementations where certain wide area modeled iono/tropo data are added, and with these, the overall quality and reliability of the phone’s position can improve. But still, any results are in the range of meters. So, AGPS does not qualify as high precision GNSS for any applications where we would expect results of a meter or less. But, AGPS has a lot in common with other augmentations that can.

Pluses: May be included with your cellular service, improves TTFF
Negatives: Imprecise, requires a cellular service
Expected Precision: 1m – 10m

SBAS: Satellite Based Augmentation Systems (SBAS) gather observation data from regional, nationwide, or even continental networks of continuously operating GNSS stations. They can also pull data from other sources, like clock and orbit data from scientific entities or cooperatives like NASA/JPL or the IGS (International GNSS Service). Together with models of iono and tropo data (depending on the system), the “correction” data is broadcast to end-users from dedicated GEO satellites (that will maintain a location above a specific region as the Earth rotates).

The primary goal for many of the national and regional SBAS services is for positional integrity assurance. Public safety and civil aviation are drivers for providing such services. For instance, in North America, the Wide Area Augmentation System (WAAS) was chartered by the Federal Aviation Administration (FAA). In Europe, the equivalent is EGNOS, GAGAN in India, MSAS and QZSS in Japan, SDCM of the Russian Federation, SNAS/BDSAS in development in China, SPAN for Australia and New Zealand, and others in development for South America and the Caribbean (SACCSA), KASS (Korea), and an initiative for SBAS for Africa and the Indian Ocean (ASECNA).

In most cases, GNSS manufacturers will enable the respective civilian SBAS system for where a unit is sold or provide firmware upgrade options to enable in other areas. In some cases, the high-precision positioning engines in surveying or mapping receivers have solutions that can use the SBAS satellites for additional ranging. This is very much the case for Japan’s regional QZSS satellites, where they are processed alongside other observed GNSS satellites. Civilian SBAS like WAAS is even an option in some low-cost consumer receivers, like those used for outdoor recreation purposes. And SBAS is also widely used for coarse resource and asset mapping applications. One drawback is that you must have at least one of the GEO satellites in view to receive the corrections. Depending on the system, you may be able to maintain precisions for a period while the GEO satellite is briefly out of view.

Pluses: Civilian SBAS is free
Negatives: SBAS satellites must be in view, hardware must have such systems enabled
Expected Precision: 1m-5m (sometimes better depending on the respective system)

Broadcast DGPS: Early applications, like marine navigation, took advantage of GPS for applications like harbor approaches. Having previously relied on sextants, compass, and certain radio beacon solutions, the initial 10-meter precision of early uncorrected GPS was a game-changer. Soon after the first differential solutions, like those using the GPS L1 signal C/A (coarse acquisition) code, were delivered from fixed GPS bases and companion dedicated coastal radio towers to ships. Commonly referred to as DGPS, networks of such towers were also sometimes established inland to support river navigation and low-precision terrestrial applications like mapping.

“DGPS was the first real-time correction,” says Mackie. “When it first got implemented, if you were working ashore, you went from having to observe static data and post-process to where you could now go out in the field and have continuous motion and know where you were second by second. This is versus determining where you were hours ago, after all the processing you did to figure it out. One of the early markets was to meet a requirement for harbor approaches for ships as well as positioning of aids to navigation in the harbors. Old methods could add many hours, even days. And when a lot of countries put in coastal DGPS beacons, as the Coast Guard did in the U.S., we had positions for the ship on demand and precise positions of the physical aids to navigation that those ships could use to stay in the channel.”

“As soon as the SA was turned off, GPS suddenly got magic, or much better,” says Mackie. “Default precision jumped up to nearly that of DGPS”. And with publicly accessible SBAS systems being established, DGPS in some ways became moot except for the fact that DGPS had a very robust integrity management aspect aimed at the maritime community. “In the U.S., the national DGPS systems have been decommissioned,” says Mackie. “There are still valid applications for DGPS, and it does still exist for specific regions and needs. But in regions where maritime DGPS was never adopted, SBAS, RTN, PPP, and some new navigation aids are the go-to technologies.”

Pluses: Typically, free broadcast services
Negatives: Requires compatible radio receiver. Many DGPS systems being phased out
Expected Precision: 1m-3m (sometimes better depending on the respective system)

Real-Time Kinematic – RTK: This method, and its network-style cousin NRTK/RTN, represent the top of the food chain for GNSS precision. In the case of RTK, a base receiver is required, typically set on a known point (unless used purely for relative positioning), and it needs to have a way to deliver “corrections” to a “rover” receiver that is seeking to determine a new position.

It is resolving sources of error by differentially examining the observations received at both the base and rover. The corrections are often transmitted by the base and rover radios (e.g., UHF, spread spectrum, long-range Bluetooth, WiFi) or via I.P. (cellular). The Achilles Heel of RTK is that the precision degrades over distance from the base. 10 km is the simple rule of thumb, though longer distances can be achieved depending on conditions like space weather (ionospheric). There is an advantage over any other solution in precision, particularly if the base is quite close to the rover, like on a construction site. Under such very-short baseline conditions, the horizontal and vertical may be as good as it can get. But you have to correctly establish a base or have access to one nearby whose quality you trust. It might get stolen, and you also have to resolve the geodetic values for the base location (unless working on a purely relative mode).

Most manufacturers of high precision, RTK capable bases and rovers accommodate international standard, non-proprietary, correction formats (e.g., RTCM) in addition to any proprietary formats they may offer. RTK is improved with the implementation of multiple constellations mainly. This enables users to work in more locations, like under light canopy or in urban canyons, where previously it was difficult with only one constellation. This benefit is the same for any of the methods and solutions — the more satellites, the merrier.

Pluses: High precision. Fast TTFF
Negatives: Need to set up or have access to a base receiver. Base can be stolen. Base setup and local conditions like multipath can impact the quality of solution sent to the rover. Power and stability of the base, no automated integrity tools unless PPP also running o the base. Communications between base and rover must be worked out technically (coverage area) and through radio licensing. Need to resolve geodetic reference of the base (if applicable). Degradation of solution over long baselines.
Expected Precision: 2cm-5cm

Network RTK – NRTK\RTN: Networks of base stations have been set up encompassing wide regions, whole states, cities, and countries — to form real-time networks (RTN). These can provide both single-base solutions from individual stations and often multiple styles of “network solutions”. While there are some variations that employ rover-side solutions (e.g., master-auxiliary), the server-side solutions (e.g. VRS), quite common globally, work like this — data from multiple bases is modeled to create corrections tailored for the location of the rover within the network, and the corrections are transmitted to the rovers via I.P. (e.g., cellular) and mostly in standard formats. The standard protocol for such transmissions is called NTRIP, and nearly all manufacturers have included NTRIP in their field software.

Some of the advantages of RTN are that you do not necessarily need to own a base (though many survey firms keep a base-rover pair for certain situations). Provided you have cellular coverage, you fire-up the over, connect, and off you go with high-precision results on an established reference framework.

Like RTK, there are options for international standard correction formats, and almost without exception, RTN can provide these even if it also provides proprietary corrections. RTN is operated by public and private entities, and public/private cooperatives, so there may be a need for subscriptions. These networks have become de facto positioning utilities and increasingly have become core geodetic reference infrastructure for their respective coverage areas. Some manufacturers have built broad footprints of their own networks to provide their customers with full solution portfolios. But a point to make is that essentially there is little difference in results and precisions between the different types and styles of RTN. What can make more of a difference is how well an RTN is operated: how well the geodetic integrity is maintained, uptime, modernized for multiple constellations and signals, the spacing of stations, etc.

Pluses: High precision. No base needed. Work is done on a common geodetic reference. Most RTN do automated integrity checking. Solution good over longer baselines. Fast TTFF. Better orbits and tectonic loading models are often applied.
Negatives: Need cellular comms. May need a subscription
Expected Precision: 2cm-5cm

Global Positioning Services – PPP (Standard): There is a technique for deriving high precisions that is similar to SBAS in that it does not necessarily require a dense network of local reference stations like RTN, and the corrections can be delivered by communication satellites. But unlike most SBAS, some of these services can yield precisions in the range of a few centimeters.

PPP was originally pioneered by Natural Resources Canada to support surveying and mapping in the expansive Canadian wilderness; it began as a post-processed solution. Public and commercial entities have also established many real-time and post-processed PPP services in the years. By applying high-quality clock and orbit data to the processing of GNSS observations, one can now derive positions in the order of 2cm-5cm. But there is a long convergence time for standard services, which can range from 15-30 minutes. Once the solution has converged, you can go about at the high precision work. But faster convergence systems now exist. More on that later…

Early real-time PPP services were set up mainly to serve marine applications, harbor approaches, docking, precise navigation, offshore construction, etc. Systems were developed by Fugro, Veripos, and others. Precision agriculture was another early implementer, like Starfire and Omnistar. Marine applications were not necessarily looking for centimeter precision, and for agriculture such systems provide excellent results, though for certain high precision needs, some in agriculture also tap RTK and RTN.

In the past ten years, a new wave of global real-time PPP services has emerged. By either acquiring and then augmenting existing systems or building them from scratch, a number of high-precision GNSS manufacturers have established global networks of 100 or more tracking stations each to create the PPP “products”. They have also worked out agreements with commsat providers to broadcast these via GEO sats. Examples of such systems are Trimble RTX, Leica SmartLink, and Hemisphere’s Atlas. It is not uncommon for such PPP services to offer different flavors, and different precisions and price points for different end-users.

Rovers need to have compatible receivers, and with most manufacturers providing positioning services that are limited to their customer base with compatible hardware (and those with gear from certain OEM manufacturers).

These services are yielding what is essentially survey-grade precisions. With a few caveats… One problem is that you need to keep one of the broadcasting GEO satellites in view; this, in addition to any challenges you might have to see enough GNSS satellites. The GEO sats are typically up over the equator, and if you lose the PPP signal for more than a few minutes, you need to re-converge. There are some applications of these manufacturer PPP services that assist when RTK loses cellular connections. Both RTK and PPP are running at the same time on the rover, and if the cell signal is lost, the PPP can take over seamlessly, for a number of minutes at least.

But a solution for the GEO view limitation may be on the near horizon. It was recently announced that one of the providers (Trimble) is entering into an arrangement with a satellite radio provider (Sirius XM) to broadcast PPP from one or more of its HEO (highly elliptical orbit) satellites. Like Japan’s QZSS quasi-zenith flavor of HEO, the beauty of such an orbit, that it is kind of like a lopsided figure-8 that covers a specific region of the Earth, is that the sats will be in high view. The news is early and few details of what level of PPP services this may serve, but an interesting development all the same. It looks though, like it will be focused on autonomous vehicle applications.

Pluses: High precision. No base needed. No need to resolve geodetic reference. No baseline limits. No cell comms needed.
Negatives: Long convergence time. May need a subscription.
Expected Precision: 2cm-5cm

Global Positioning Services – PPP (Fast): The key drawback of standard real-time PPP is, as noted, convergence time. But there have been great improvements in the past few years. Commercial positioning services, in particular, have augmented their own global tracking networks with data from local and regional semi-dense (100-300km spacing) networks of ground stations. The global networks generate high-quality clock and orbit data, and the local/regional networks add in elements like iono model data, and even (for some) tropo models. This has not only reduced the initial convergence time to under a minute, even faster in some cases but can also improve results. These semi-dense networks, for fast PPP, are rapidly propagating. For instance, the RTX Fast service now covers most of Europe and the US, and Canada.

Autonomous Applications PPP and Hybrids: It is the wild west days right now in the development of RTK, PPP, and Hybrid solutions focused on the growing autonomy market. Dozens of developers are creating new networks of ground infrastructure or purchasing data from existing RTN to create real-time positioning services to feed integrated systems being designed and built for autonomy. Sapcorda, HERE, Swift Navigation, and many more, including several of the manufacturers of high precision GNSS gear, have already stepped into the PPP-for-autonomy ring and Trimble for example is on the road with the GM Super Cruise system.

The requirements for autonomy tilt more towards resiliency and reliability than to precision. Autonomous vehicle solutions involve multiple sensors and do not rely on GNSS alone, but it is important for the GNSS solutions to be consistent, with a minimum of interruptions, have fast TTFF, and be at a price point suitable for mass-market applications. It will take another long article to examine real-time positioning for autonomous applications.

Which one to choose?

There are a number of issues to consider when choosing a positioning solution — price point, hardware requirements, precision requirements, personnel training considerations, sky-view limitations, the need for relative or absolute accuracy, availability of services in a given location, etc.

“There’s still a lack of understanding of why you use one type, what the limitations are, and where you have to start thinking about the other,” says Mackie. That’s the biggest education piece we run across. It’s not so much talking about the technology or how it works, it’s talking about when to use which.”

There are resources, though, both formal and informal to help you decide, and this article only scratches the surface. The manufacturers often post case studies, and while those might be motivated by marketing, at least you get to see real-world examples, and maybe some like your planned application. And the proliferation of social media, like LinkedIn and work-type-related pages, also help users find examples.

Choosing the right solution is application by application. For example, why do construction companies often put up a base at construction sites for RTK instead of RTN or PPP? “Many construction companies are working with fine vertical. They are always going to try and use the solutions that give that absolute best vertical solution,” says Mackie. “If they’re roughing something in, there’s no worry at all for using a network or even some kind of PPP. But once again, for much finer grading, short-baseline RTK is extremely reliable. Assuming the base doesn’t get stolen. And it’s a known technology that provides the accuracy they need.”

Then there are the numerous choices of PPP positioning services providers and multiple flavors from each. Which to choose for which application? For an insider’s view, I asked Michael Bruno, an engineer who has worked in both the realms of differential and PPP solutions and is now a program manager in Trimble’s Advanced Positioning group. Like other PPP providers, they offer various tiers of PPP solutions. Theirs are all under the blanket term “Trimble RTX”, for different end uses, differentiated by precision ranges and price points.

“We have one flavor that is the same kind of precision users were used to with DGPS, for sub-meter, we call it ViewPoint RTX,” says Bruno. “This is good for the kind of field mapping folks do for certain GIS layers, asset inventory, rough mapping. And it’s at a low price point.” For farmers looking for half-meter pass-to-pass precision, there is RangePoint RTX, “This is widely used in Agriculture; there are broad areas of the US and Europe for instance where the majority of our ag users are using RangePoint RTX to achieve better performance than SBAS in the region. It is very affordable for the benefits they receive.”

“For some asset inventory and mapping, users are looking for tighter than sub-meter, but do not need survey grade, for this, there is FieldPoint RTX,” says Bruno. Again, the precision goes up, the price goes up. And then there is the tier that is essentially survey grade, centimeters. “CenterPoint RTX is our top tier for real-time, and we also have an online portal for automated post-processing for situations where you might not have a view of the L-Band sats (those GEO sats deliver the RTK PPP data),” adds Bruno. “CenterPoint RTX gets you 2cm once it is converged, and this is where the newer ‘fast’ service comes into play. All of the RTX tiers come from data we get from our global base network, and the old convergence times were measured in 10s of minutes. Years ago, we started adding data from more stations, like in the U.S. Mid-West where we have a lot of ag users, and these semi-dense networks did not have to be as close together as RTN stations but provided enough additional data to get the convergence times down to well under a minute.” The fast service now covers most of Europe and the U.S.

If PPP is now capable of survey-grade precisions, are providers urging users to simply move to PPP instead of differential solutions? Like some other providers, Trimble has PPP services, operates their own RTN (e.g., VRS Now), and also sells software for public and private RTN. Are these services at odds, or do they complement each other depending on user needs? I asked Bruno how they would advise different users on which to use. “CenterPoint RTX, especially now that it converges fast, can do many of things that users are used to with RTN, but there are limitations, like keeping the (GEO) sats in view,” says Bruno. “RTX has big advantages for areas where there is no cell coverage, as RTN rely on a cell to send corrections. But where there is cell coverage, and an RTN, surveyors for instance should simply use the RTN, and the PPP can fill in seamlessly where there are cell coverage holes.” This type of top tier, fast PPP are growing in emerging markets, says Bruno. “Our biggest user base is from ag, especially for RangePoint RTX, but now we are seeing a whole wave of different users for CenterPoint RTX. You can imagine uses cases like asset inventory and management some autonomous delivery services, and automotive autonomy development. The sky is the limit.” An example of the latter that Mackie mentioned was the Super Cruise system, making use of RTX and installed on many new Cadillacs, has just reached the milestone of over 6 million miles of hands-free driving.

Example application scenarios

A broadacre crop farmer might choose a low-cost, lower precision PPP as they are most interested in pass-to-pass consistency only during plowing/seeding. While a specialty crop farmer may be interested not only in row-to-row repeatability but even year-to-year, so might need to go full RTK.
A rural city has an asset inventory program for putting fire hydrants on a GIS map in wide open sky desert town and can do this with a standard PPP service. They have bought some high-precision rovers to operate with phones and the Esri Collector and Survey 123 apps. They chose receivers that support both PPP and can be used as RTN rovers for flexibility.
Another mapping team, putting dozens of interns out to inventory every street sign might not want to deal with the long convergence times of standard PPP and the mapping manager opts for a fast PPP service. But soon they realize that many of the street signs are along east-west tree-lined streets, the PPP GEO sats might be hard to see, so they choose rovers that can also connect to an RTN. And to also help with the tree canopy obstructions, they choose rovers that are multi-constellation capable.
A mapping crew that only does field work during the summer decides to rent rovers for those few months that come with a PPP subscription, and with pay-as you-go token options.
One construction company uses an RTN for rough layout, while another that needs to do precise grade checking sets up a base on the work site for RTK.
A surveyor that typically uses and RTN for in-city work gets a contract for a job in a remote area with no cell service. They use PPP to establish a control marks, and then set up a base for RTK.
An industrial robotics developer is designing bots for the materials yard of a large manufacturing facility. The bots have IMU and LiDAR to assist navigation and collision avoidance, but the GNSS is essential for keeping the IMU calibrated. The PPP signal is lost when the bots travel under gantries, so they opt for a fast PPP service to keep the convergence time brief.

What is your scenario?

One size does not fit all, so fortunately there is a whole range of options. And we can be sure that many more are in the works. “It’s interesting to see how technology progresses,” says Mackie. “Today, we might be saying that we don’t see PPP fully supplanting RTK. But not too long ago, we were saying, ‘Wow! We are getting better than a meter of accuracy. How could it ever get better than that?’ I think as processing gets faster, and with the science that’s being put into understanding what can be done, we may someday look back on the current solutions with that kind of nostalgic perspective as well.”

Note: Geospatial Media and Communications does not endorse or recommend any specific commercial positioning services. Examples given are for explanatory purposes only.