CHOICE membership

Automotive GPS Speed Accuracy vs Speedo


They are a great check tool assuming at the same time you are the only car or truck on the road for some distance and can keep a fixed 100kph or what ever speed you choose to sit on for the check. 180 seconds for 5km at 100kph you still need a stop watch function and passenger to have a reliable enough result to be within a 1kph range or approx +/- 1%.

Also note that the effective rolling radius of a tyre changes with inflation pressure and car load. It’s easy to get slightly different results. The more variables the more the little variances add up to one big error.

Good luck finding a set of mileage (kilometer distance traveled) markers to check your speed against around here, SE Qld. I used to drive in other parts of QLD and there were several set out in convenient locations.


Once again we have people telling us that in theory speedos are more accurate than GPS and the reverse. Both measures are bedeviled with many sources of error and without any solid numbers it is a guess which is better or worse.

Perhaps some measurement of real world examples would shed light on the question. If the measurements were taken over a sufficient range of circumstances by a sufficiently accurate method then the matter might even be settled. Until then it is arguing how many angels can dance on a pin head.


Fact - neither the GPS in a mobile phone nor the everyday car speedo are consistently accurate?

For a different view on speed and distance traveled you might need to consider a Halda rally trip meter used by amatuer car rally participants and or for professional WRC options such as;

Note all professional options require some form of calibration for speed?

Note no need to check the shoe size of angels or measure pin heads with any of these options.


Wow, what passions!
Here’s my take on theory vs reality.
The Quora reference above claims GPS position error “up to 5 metres”.
In extreme situations with poor satellite signal (such as inner-city surrounded by high-rise buildings) you get total signal drop-out and no meaningful result.
On open highways, particularly with modern GPS’s with both GLOSNASS and GPS receivers enabled, you can expect somewhat better accuracy.
A simple check is whether the GPS can accurately display which side of a dual-carriage highway you are on.
Aust Design Rules specify speedo’s should be between -10% thru +0% accurate. My experience with at least 20 or 40 hire cars (Toyota Corolla, VW Golf, Audi A1, A3, A4) is that all new cars on new tyres read -5% on their speedo’s (matched against two Tomtom GO 520’s, a Garmin DriveAssist 51, and two ODBC displays). That is static speed displayed in cruise control on a straight, flat road at highway speeds.
You often get a discrepancy of 1 kph and rarely 2 kph during acceleration or braking as each device seems to have a different sampling rate and averaging interval.
Different models and manufacturers of tyres and tyre wear do affect speedo and ODBC readings, not GPS.
Roadside radar speed displays usually match my GPS, though sometimes they seem affected by other vehicles or perhaps braking.


This article (May 2018) also provides information on the accuracy of GPS in Australia. …currently about 5m but the government is investing $260M (last years budget) on Satellite Based Augmentation System (SBAS), which aims to correct GPS accuracy to around a metre.

If an accuracy of a metre or less is achieved, then GPS based speed measurements will become substantially more reliable and accurate. This would bring accuracy from about 8-10% down to about 2%. A 2% accuracy is likely to be equal to or better than a cars speedometer.

Until SBAS (and using devices which can use this service) is rolled out, one is playing Russian roulette if one only uses GPS to measure vehicle speeds. Presently there is a risk that one may pass through a speed check/radar with the GPS under reading actual speed by maximum of 8-10km/hr (viz. In worst case scenario a GPS indicating 100km/h whereby actual speed is 108-110km/h).

With SBAS, worst case scenario this would come down to about 2km/h or 102 km/hr actual speed when the GPS indicates 100km/h.


GPS speed is an interesting debate, but to me it has become esoteric and technical while ignoring ‘why’ are we worried about it. When all is written, GPS computed speed is imperfect yet accurate enough for a purpose.

A speed display is to avoid an infringement. How many drivers on any road conclude ‘I should not be going over 60 kph’ here because it would be unsafe as compared to ‘I should not be going this fast here and need to slow down.’ Setting and enforcing what is usually an arbitrary limit is seen by some to be ‘part of the toolkit’ to reduce the road toll; it is certainly a revenue raiser; but if one steps back to look at what constitutes safe driving I propose GPS for speed measurement has as much merit as a mechanical speedo, or its electronically simulated counterpart, excepting it is more problematic to accept it for the purpose of issuing infringements, although that could be addressed by rational drafting of related legislation.

What is the purpose of a limit and which speed is more accurate? In the context of road safety why would 1 or even 5 kph matter? Is it because speed kills and that is an absolute above all other concerns? I suspect there will be at least one rebuttal that a crash at say X+5 kph will be more serious than one at only X kph, and nobody will change that or my alternative mindset. But avoiding a crash is about more than speed, it is about attention, reaction times, vehicular braking, keeping safe distances, and safe driving, not just ‘speed kills’. And therein lies one problem, that it is easier and less costly as well as profitable for government to get revenue from arbitrary rules than it is to patrol and take bad drivers off the roads and keep them off.

FWIW I always drive with my GPS running, and day in day out it shows a constant relationship to the digital speedometer display (within seconds of lags) depending on terrain and changes of speed. I trust it is probably a tad more accurate than the speedo at constant speed since when driving, if I use the GPS speed I am literally one of many cars travelling down the road together, but if I use the speedo I will soon be leading a parade, giving me cause to think my speedo might read a bit higher for any given speed than most. It could be everyone else on the road is technically speeding, but would it always be everyone but me?