GPS – a quick guide

Hey all, just found this on & it is a great read, hope you find it useful!

GPS Accuracy and Limitations

GPS is proven to be a very valuable tool for the purposes of Surveying and Navigation however its users must be aware of its characteristics and cautious of its limitations.

Common Factors affecting the accuracy of GPS:

  • GPS Technique employed (i.e.: Autonomous, WADGPS, DGPS, RTK, etc.)
  • Surrounding conditions (satellite visibility and multipath)
  • Number of satellites in view
  • Satellite Geometry (HDOP, GDOP, PDOP etc.)
  • Distance from Reference Receiver(s) (non-autonomous GPS ie: WADGPS, DGPS, RTK)
  • Ionospheric conditions
  • Quality of GPS receiver

Precision and Accuracy

Although precision and accuracy are often assumed to be the same thing, technically they are slightly different. Precision refers to the closeness to the mean of observations and accuracy refers to the closeness to truth.

Care must be taken particularly when using differential GPS to the accuracy of the results (closeness to truth) as reference points used can and often are inconsistent with truth.

The precision or accuracy quoted by many GPS manufacturers is often done using a statistic known as CEP (Circular Error Probable) and are usually tested under ideal conditions.

sp = standard deviation of latitude
sl = standard deviation of longitude

CEP = 0.59[sp + sl]

CEP is the radius of the circle that will contain approximately 50 percent of the horizontal position measurements reported by the GPS receiver. This also means that 50% of the positions reported by your GPS will be outside of this circle.

Another common measure of accuracy is 2DRMS (Distance Root Mean Squared).

2DRMS = 2*sqrt(sp*sp + sl*sl)

2DRMS is the 95-98% probability that the position will be within the stated 2 dimensional accuracy. The probability varies between 95-98% because the standard deviation of latitude and longitude may not always match.

Two plots are shown below (data courtesy Satloc)

Each has been created using 24 hours of data taken at 20 second intervals in the south western USA.

Figure 1:

GPS Corrected with WAAS

Figure 2:

Autonomous GPS

There are 4 techniques commonly used for GPS Navigation: Autonomous, WADGPS and RTK. Surveying applications usually require the use of RTK or Post Processing.

When used properly under ideal conditions, the CEP precisions for each method will depend on the quality of the GPS equipment in use and is approximated below:

Autonomous <10m
WADGPS 0.3-2m
RTK 0.05 – 0.5 m
Post Processed 0.02 – 0.25 m

Accuracy (closeness to truth) of differential systems is relative to the accuracy of the reference points used.

When used in less than ideal conditions, the accuracy and precision of any GPS system can be degraded significantly.

Ideal Conditions

Ideal conditions for GPS Surveying or Navigation are a clear view of the sky with no obstructions from about 5 degrees elevation and up.

Any obstructions in the area of the GPS antenna can cause a very significant reduction in accuracy. Examples of interfering obstructions include: buildings, trees, fences, cables etc. Obstructions may have the following effects thereby reducing accuracy:

  • Reduced number of satellites seen by the receiver
  • Reduced strength of satellite geometry (Dilution of Precision (DOP) values)
  • Satellite signal multipath
  • Corruption of GPS measurements

Multipath is caused by GPS signals being reflected from surfaces near the GPS antenna that can either interfere with or be mistaken for the signal that follows the straight line path from the satellite. In order to get an accurate measurement from a GPS satellite, it is necessary that the signal from the GPS satellite travels directly from the satellite to the GPS antenna. If the signal has been reflected off of another surface prior to being received at the antenna, its length will be greater than was anticipated and will result in positioning error. Multipath is difficult to detect and sometimes hard to avoid.

Figure 3

Other Sources of Error in GPS

  • Signal Delay caused by the Ionosphere
  • Signal Delay caused by the Troposphere
  • Orbit Errors (GPS satellite position inaccuracy)
  • Receiver Noise

Common GPS Surveying and Navigation Techniques and Associated Errors

Autonomous or Stand Alone

The method involves using a GPS on its own with no additional correction information other that what is broadcast by the GPS system. Prior to May 2, 2000 accuracies obtained using this method weren’t usually much better than 100m due to a US Department of Defense induced error called Selective Availability (SA). On May 2 SA was turned off and now accuracies are usually better than 10m.

Autonomous receivers will attempt to correct the Ionospheric and Tropospheric errors bases on mathematical models which are very limited in their accuracy. They have no way of correcting for orbit errors, multipath or receiver noise.

Wide Area Differential GPS (WADGPS)

Examples of Systems that use WADGPS include:

These systems receive an additional satellite signal that contains more accurate information about GPS Ionosphere and Orbit errors allowing the GPS receiver to determine a more accurate position. These systems have no way of correcting for multipath or receiver noise. Accuracies of WADGPS are often better than 2-3 meters. Although multipath can cause very large errors as is the case in the Autonomous positioning. A solar maximum of a an 11 year solar cycle occurred near the year 2000 which can also have dramatic and unpredictable effects on the accuracy of WADGPS systems.

Real-Time Kinematic (RTK)

Many GPS receiver manufacturers provide a system that employs a technique known as RTK. RTK implements the use of much more complex GPS data processing than other techniques, although RTK can eliminate many errors characteristic of other systems. RTK has additional limitations.

When using RTK, a reference receiver (Base) must be placed on a known reference point. This reference receiver then transmits measurement or correction information over a radio link to the roving receiver (Remote) that will be used for positioning or navigation.

This technique can result in accuracies as good as 0.05 m – 0.10 m if used properly and in ideal conditions.

The limitations of an RTK system include the following:

• Initialization – The receiver must be initialized in good GPS conditions for up to 15 minutes before achieving sub-meter accuracy. If the receiver sees less than 4 satellites at any given time after being initialized, the receiver must re-initialize before again achieving sub-meter accuracy.

• Baseline Length – As the distance between the Base and Remote receivers grows larger, the errors observed between the GPS receivers becomes less and less common degrading accuracy at the remote. Good accuracies can normally be achieved with baselines (line between base and remote) in the order of 10 – 15 km. Baseline lengths can be reduced considerably when strong ionospheric conditions exist.

• Radio Transmission – The base and remote must maintain communications at all times in order to maintain good accuracy. Terrain, distance and interference all have effects on the distance in which the base and remote are able to maintain communications.

• Visibility and Multipath – Usually at least 5 satellites must be available in order to achieve good results. Although less susceptible to multipath after initialization compared to other techniques, RTK results can seriously be degraded by obstructions such as trees, fences and buildings.

• Accuracy of Reference Point – The absolute accuracy of the position reported by the Remote receiver is only as accurate in an absolute sense as is the position of the base station coordinates.

Post Processing

This technique of GPS is used mostly used for Surveying and is not used for navigation. It is similar to RTK in that a base station must be placed at a known reference point and a rover is used for gathering new positions. Instead of obtaining accurate results in real-time, accurate coordinates are generated by taking data stored from the receivers and processing them using special software on a computer. Extremely accurate results in the order of a few centimeters can be obtained if done properly and the conditions are good but post-processing is subject to many of the same limitations as RTK.


All GPS navigation and surveying techniques have limitations that may not permit desired accuracies in a given environment. The cause for poor accuracy is not always obvious but is usually attributable to one of the following source of error:

  • Multipath / signal corruption
  • Low number of satellites / poor satellite geometry
  • Erratic Ionospheric activity

These errors can lead to position errors as large as several of meters or more.

How do smartphones know where North is?…..

Conversation on Twitter, where someone pulls a statement on you and you KNOW it’s true but at the same time your brain is wondering why, you hadn’t thought about it before….yesterday was one of those days. I made a statement about whether the compass was dead and whether this is the rise of the GPS (mobile) compass, taken from an article written by the Ordnance Survey- Here

Obviously someone pointed out that GPS needs at least 3 points to calculate direction (an m, movement) value….couldn’t argue, completely right but something was telling me that my phone isn’t that stupid to tell me where North is with only 2 points.

I’m not stupid (well……)

A typical smart phone has three magnetic field sensors, fixed perpendicular to each other, which are used to work out the local direction of Magnetic North.

In addition, they have three accelerometers which sense gravity to give tilt information and to help work out which way is down.

Image courtesy of British Geological Survey

Further to this, my phone is even more clever than I realised, when it can’t get a GPS fix it will calculate where I am using its angle of approach to the cell tower, how long it takes the signal to travel to multiple tower & the strength of your signal when it reaches the towers. Since obstacles like trees and buildings can affect how long it takes your signal to travel to a tower, this method is often less accurate than a GPS measurement but it is still pretty clever!


Unfortunately mobile GPS is not that accurate compared to a dedicated GPS unit, this is usually due to;

Chipsets. Dedicated GPS, and Smart Phones use a chipset to actually perform the GPS calculation. The accuracy and capabilities vary. Some chipset’s are optimized for urban canyons and are particularly suitable for smart phones (SIRFstar III). Others can track a greater number of satellites concurrently and feature SBAS augmentation. Some cell phone chipsets may outperform those used in dedicated devices. Even some professional grade GPS (Nomad 900G) units ship with less accurate consumer grade chipsets. Antenna. The placement and quality of the antenna effects GPS accuracy. Some dedicated GPS units have external antenna’s instead of the internal antenna typical in smartphones. They may also feature an antenna socket to allow an aftermarket antenna to be added.

How bad?

The difference between smartphone GPS and dedicated GPS is shortening, in 2009 the difference was around 20m, today a mobile GPS will get a fix within 5m of a point, or if ESRI are to be believed, 3m. In this blog ESRI discusses mobile GPS in phones & tablets v dedicated units.

Nick D


Back before you’d ever realised I’d gone….

Those of you who read my blog may have realised that I’ve not written for a while, in fact for a month or so….without getting too personal, a couple of months ago I dislocated my left leg. Being a bit of a bloke, I popped it back in and thought it was okay, went up A&E the next day to find that they wanted to put me in a wheelchair for the rest of my life. Yup, it’s been a little hectic for the last few months but be rest assured, I am walking and (more importantly) back to doing the blog!

Goodbye to #GISChat

Whilst I have been pre-occupied, a lovely lady called Emily has started up a Wednesday Twitter chat called #GeoTribe. Held every Wednesday at 8pm GMT, it is a great chance to catch up on what people are talking about and to meet like-minded geopeople.

Unfortunately, it does mean the end of #GISchat as the two coincide. I am sure people won’t be too devastated and I look forward to talking to you all on the #geotribe.

Hello to Geoserver…

Those of you who I’ve met at the recent round of conferences will know that I’ve got into Geoserver and selling Boundless’ OpenGeo solution. Now, lets get this straight, I’m not selling the software but selling the skills to install it, get the plugins (like the print module and QGIS) working & also provide a little customisation at an extremely CHEAP price….only problem, you’ll have to put up with me chatting about Geoserver A LOT more….

Okay, I’ve rattled on enough and said nothing GIS, so I will leave you with this; Google Earth displays in WGS84 but calculates in Web Mercator……