# What is the maximum Theoretical accuracy of GPS?

I was talking with a potential client, and they requested that we plot some points with GPS, with a maximum (or should that be minimum?) accuracy of 2 m.

This is an area with no WAAS, and I was under the impression that even in the best of conditions, a single gps point can be accurate up-to only 15 meters(Horizontal field). Is this correct?

What is the maximum theoretical accuracy of GPS without using WAAS or differential GPS?

• What type of receivers will you be using? – R.K. Dec 10 '12 at 13:05
• It is receiver dependent of course, see what a low end gps provides in this thread gis.stackexchange.com/questions/12011/… – user681 Dec 10 '12 at 13:27
• I'm not interested in any specific instrument. I'm concerned with the Theoretical Accuracy in Typical Environmental conditions. – Devdatta Tengshe Dec 10 '12 at 13:42
• @DanPatterson I think that particular instrument and readings are using WAAS, and are not pure Satellite based GPS positioning. – Devdatta Tengshe Dec 10 '12 at 13:45
• Not WAAS enabled @Devdatta, 7 year old Garmin Etrexs. Sometimes stationarity at a site yields better results for any GPS. The link was just food for thought. – user681 Dec 10 '12 at 14:28

The United States government currently claims 4 meter RMS (7.8 meter 95% Confidence Interval) horizontal accuracy for civilian (SPS) GPS. Vertical accuracy is worse. Mind you, that's the minimum. Some devices/locations reliably (95% of the time or better) can get 3 meter accuracy. For a technical document on that specification you can go here.

For more general GPS accuracy information, head to GPS.gov's website. That website also includes data and information on WAAS-enabled systems and accuracy levels depending on location. It's a great resource.

Basically, you can't get 2 meter accuracy reliably without some form of correction.

Edit: Something else to contemplate is using a device that can communicate with both GPS and GLONASS satellites. I'm not aware of any accuracy articles or studies that combine both systems to improve accuracy, but at the very least, it increases the potential satellites that may be available at one particular location/time, especially near the poles.

• I had looked at the 2008, SPS standard before, but it is very dense and a difficult read. I can see that they have mentioned 4m as the 95 percentile for the error; But I was rather hoping for some kind of formula or equation that proved it. – Devdatta Tengshe Dec 11 '12 at 3:21
• why do you insist on "theoretical" best performance, or want to see the math behind the answer above (esp after you say that the standard that lays it out is "dense and a difficult read")? Your potential client has a budget in mind which will determine what grade of GPS units you can take to the field. Your readings will not be any better than the vendor claims for the unit, and probably will be worse. As several answers/comments have noted, a consumer grade GPS will get you 10-15M. Without WAAS or differential GPS, you won't achieve your goal, regardless of the equations. – Llaves Dec 11 '12 at 5:01
• For clarity, the advertised 7.8m (95% ci) is of the pseudorange (satellite-to-receiver range), not necessarily user location accuracy. Though they are strongly related, they are not the same thing. The horizontal accuracy of a receiver is dependent on where the receiver is to all of the satellites, the angles to each satellite, etc. – r2evans Mar 31 '16 at 17:17
• Does the precision depend on a crystal inside the mobile phone? – prosti Jun 10 '17 at 22:20
• @prosti - on handheld device, the major limitations are a) antenna length/quality and b) battery consumption tradeoffs in the circuitry. – ToolmakerSteve Apr 7 at 14:55

Ionospheric delay effects are the largest source of error in a single-frequency GPS receiver. WAAS and CORS are able to correct for this better than a receiver's almanac, so the best you can do with uncorrected GPS is typically about 15 meters. Survey-grade GPS using RTK is able to achieve centimeter accuracy.

in European countries, out in the field (not inside a city with buildings), the best accuracy without any aid is 5 meters. I have also witnessed a 2 meter accuracy but that is extremely rare and I would not take it into account. The average best would be 15 meters and the average worse close to 30-40 meters.

The results stated above are from my own field work and come from using various types of smartphones. GPS accuracy greatly varies depending on surroundings, devices used, weather and many other factors. The accuracy results are derived from compairing my actual position with the GPS position.

I hope this helps.

Cheers, A

• And how do you define accuracy? – johanvdw Dec 10 '12 at 20:59
• I use Arcmap to set my fixed test positions and then measure the distance from the GPS positions. The results that I stated above are average and rounded accuracy values for European countries (mostly UK, Spain and Germany ) – Antony Dec 11 '12 at 11:14

These are results of tests carried out in October and December 2014. Tested were the following GPS telemetry, GPS data loggers and a hand held GPS.

GPSFlight STX900e telemetry, embedded GPS unit ublox LEA - S5

BRB 900 MHz GPS Telemetry, Lassen IQ series GPS receiver

TLA 900 MHz GPS Telemetry, ublox 6 series GPS receiver

Holux 1200e GPS data logger

Tripmate 852 GPS data logger, GPS chipset, MTK

Garmin etrex H

The benchmarks were 2 state survey marks (SSM) 35 km apart in Sydney, Australia. The data sheets for these marks were provided by NSW Spatial Exchange. Devices were left on the marks for 15 minutes at mark 1 and 30 minutes at mark 2.

All devices are DGPS capable.

All use WGS 84 height datum

Mark 1 (60% sky view approx)
Mark 2 (80% sky view approx) Deviation from position, altitude

GPSFlight 1= 9.5 ft, 18.4 ft (7 sats)
2= 3.5 ft, 4.6 ft (10 sats)

BRB 1= 11.9 ft, 16.2 ft (7 sats)
2= 13.6 ft, 3.7 ft (10 sats)

TLA 1= 12.6 ft, 14.8 ft (7 sats)
2= 14.2 ft, 3.6 ft (10 sats)

Holux 1= 29.3 ft, 40.3 ft (7 sats)
2= 9.8 ft, 20.4 ft (10 sats)

Tripmate 1= 18.5 ft, 27.3 ft (7 sats)
2= 7.1 ft, 1.6 ft (10 sats)

Garmin 1= 45.6 ft, 10.8 ft (6 sats)

2= 13.5, ft, 23.0 ft (10 sats)

Looking at the data from these readings, all these GPS unit's positional data "drift" around the static mark but on average give very acceptable accuracy for my purpose IMO. I would think that these devices would need to be left for 1 hour to stabilise properly. I use these GPS devices on World kite altitude record flights and in the remote zone I use and at altitudes up to 16,000 ft, there are 12 or even 13 satellites acquired. That's way more than most people seem to get in urban or forested areas. Any of these units will give me better than 0.08% accuracy over my height profile. If you want better accuracy then you need something like a Trimble Total Station or a similar survey grade GPS with additional augmentation. It depends on how much you pay and your understanding of the usage and limitations. People sometimes have unrealistic expectations of what is wonderful technology at a dirt cheap price. I originally paid $600 for the GPSFlight telemetry 9 years ago. The BRB was$300, TLA around $250, Garmin etrex$99, Holux $100, Tripmate$50. I can remember handheld readily available GPS units were close to \$800 about 20 years ago. They are getting cheaper and cheaper but I cannot vouch for the quality or accuracy of all cheap units. Manufacturers of consumer grade GPS units usually make claims of sub 3 meter for position and sub 5 meters for altitude under good conditions. It seems it's still tremendous bang for buck seeing these devices can place you within a 10 meter square on the surface of the earth with 40 billion square metres.

• Really interesting data - just to confirm, the 'deviation' readings are the actual deviations measured compared to the state survey marks? Did you take multiple readings or a single reading at the end of 15 / 30 min? – Simbamangu Aug 6 '15 at 5:54

Realistically with a modern survey grade GPS receiver you can easily get down to a few centimeters via RTK within seconds or down to a few millimeters with static post processing and a several hours of post processed data.

## protected by whuber♦Dec 30 '15 at 15:58

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).