CHOICE membership

Automotive GPS Speed Accuracy vs Speedo


When we bought our current Honda CRV 4 years ago, I noticed that the newly built-in GPS system lacked both a ground speed display and speed/red light camera alerts.

I bought a TomTom Go 620 from JB HiFi which we use whenever we are travelling on long trips. Constant display of ground speed and no need to use a magnifying glass to read the display and it came with free lifetime map updates.

The built-in Honda display is used for what it is best for, namely displaying our collection of MP3 recordings of hit songs from the past 60 years.

Why waste money on upgrading a GPS system which was not up to speed in the first place?

Cost of map update on built-in GPS

The built in variety are crippled in that respect because of liability issues. By providing those functions they risk us owners comparing the GPS speed, that is computed differently from the speedometer speed and complaining, regardless which is more accurate, and vis a vis the ADR, as well as being accused of supporting unsafe driving by flagging camera locations.

They have no liability for the portables.

The push back on some ‘fun’ commercials is evidence enough for me that is the case.


Yeah GPS accuracy for speed is hampered by the inaccuracy built into the GPS systems to stop accurate placement of weapons etc. The military of some Governments and some of their departments get centimetre accurate readings but the timing used for non secure use is out by seconds or parts thereof. This is why most GPS are only accurate to some number of metres and for the purposes of surveying etc a placement must be left for a few days to further refine the coordinates. This is done by allowing many GPS signals to be captured and the data is then plotted to give a more accurate reading (the plot is like a circle with the most likely position the point in the centre). The way smartphones work is that they also use wireless signals to further refine their accuracy in realtime but even this is out.

So someone relying on the accuracy of a GPS to argue or determine their legal speed is an erroneous concept and could be costly.


Wandering, but in the earlier days when the US owned and operated the constellation, they would periodically ‘test’ what is called ‘SA’ (eg public availability) by changing where you were for some seconds at a time. (first person experiences with that - it reinforced it was a secondary nav device in those days). As GPS navigation became more prevalent they stopped doing this in the name of world competition as well as aviation safety. Today GPS navigation as well as instrument approaches are widely certified, country dependent.

A good treatise on GNSS with a few short words of history is at the US FAA site.

Note that instrument certifiable aviation GPS are a quite different standard and individually tested in each installation, as compared to automotive versions.


There are also elevation effects which aren’t corrected by car GPS and also the GPS assumes movement on a flat plane when the earth is in fact round.

There are also others as well:

In my last job we used a Trimble differential with Omnistar signal to get accuracy down to about 0.10m. Without the Omnistar signal, the accuracy was submetre…but only when multiple points were taken for averaging.

A car GPS has the following accuracy:

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.

Even using say the 4 metre accuracy, if a car is travelling along a road, each sample point taken by a GPS can be 4m out. If there is a shorter sampling duration, the accuracy of a car GPS providing accurate speeds is low and could be out by >10% or more. Reliability of the speed measurement would only occur when speed is high and sample duation is long (sample time say corresponds to say 100m or more and then the accuracy could be <8% (4m + 4m over 100m).

This is why car GPS readings aren’t admissible in court when challenging a speeding fine as they are not accurate.

Unfortunately, there are many who think that GPS are more accurate than a car speedo and use the GPS speed when driving…and why some receive fines for speeding thinking that have being obeying the speed limit.

Maybe the manufacturers of the GPSes should state clearly the accuracy of the speed detected by their devices.


The GPS IS more accurate than the car speedo. Our car speedo is 5 to 8% underspeed. It is normal to have a little under reading to allow for tyre wear variation. It is so great on our car that the mind automatically adjust and travels at 108 in a 100 zone. I had to remind the wife to drop to 106 after new tyres fitted.
I complained about the excess to the manufacturer. They just said they had to do it & couldn’t re-calibrate.

Plug in an ODB device into the maintenance port & we could see the computer speed - just a little over the GPS reading!
(apologies - off topic)


If you are going constant speed on a flat level straight roadway that is probably the case, but otherwise may not be at a point in time, especially in hilly terrain or winding roads, and especially winding hilly roads.

An obvious question is how do you authoritatively know your GPS is more accurate although it probably is?. I remember the olden days where we would time ourselves between designated mile test markers with stop watches to get a close enough approximation. Are modern distance markers placed accurately enough for same? Has your speedo been checked on a speed calibration dynamometer?

That is because the ADR specifies the speedo reading be +10%/-0% actual speed for the tyre size originally fitted, and the speedo will read higher on worn tyres than new ones, as you wrote, as the circumference changes accordingly.

Have you considered which sensors and data constants the computer uses to determine speed? Bingo, for mechanical and computer eras. Rotational speed of the wheel multiplied by its specified circumference. It is neither fancy nor sophisticated. The filtering from that to the speedo is for the ADR, but the computer value may or may not be accurate depending on the tyres and wheels fitted at the time, and some of us do change tyres and wheels from time to time.


Each model of car GPS may use a variety of algorithms to determine the speed read out it displays This is necessary to provide a speed read out that does not change values too rapidly (filter) and become unreadable. The computer will need to use TWA ( time weighted average) to provide a stable base calculation, followed by a decision to either keep or discard any large variation as out of range. How often each new data point is processed, eg ten times per second or once per second is also relevant. Without detailed information on the strategies and algorithms used there is only observation as a guide, and this may be flawed when comparing between different makes or updated versions of the same GPS system.

In typical survey and civil construction using a GPS base station the absolute accuracy of the RTK (real time kinetic) system can me measured in millimetres (5 - 10) including relative elevation.

Our in car systems could achieve similar if they had access to a local base station. It would be interesting to know how the self drive vehicles use GPS, how critical accuracy is to operation and how they use inertial feedback as a correction to GPS variation. No doubt their error is not +/- 5m or more?


For a quantitative discussion there don’t seem to be many numbers available to check.

We have ideas about why both GPS and speedo measures may be wrong. We have estimates based on standards and ways speedos are calibrated and others based on the putative instantaneous accuracy of GPS and reflections about how well smoothing algorithms aggregate readings and what effect that might have on accuracy.

What we don’t have are any actual measurements of both that would be good enough to settle the question, in this country, on the road, which system is more accurate and by how much.

Like 17th century European botanists who have never been outside their parlour discussing the flora of the antipodes. Pass the tea.


I presume you might be into multi-variate calculus feeding into statistics then. Or are simply trolling the topic.

It depends on the specific vehicle and specific GPS system and specific road, although the overall differences from sample to sample will probably be negligible in context, and the GPS will be more accurate according to the preponderance of reputable sources.

For our ‘safety camera’ (many of us think they mostly provide ‘safety’ for the state treasuries) laden society one important difference is that the speedo will always be within +10%/-0% of the speed assuming OEM tyres and wheels are fitted. The GPS accuracy might be out +/- and the leeway for a fine is top secret :wink:


Neither. I am not suggesting building a complex model of all the possible parameters just a more empirical approach. Nor am I trying to be annoying or to draw a reaction. I am sorry if you find me being a die-hard empiricist an annoyance.

All I am after is some physical measurements instead of theory, even if only to show the range of variation in both systems and give some clues about where the strengths or weaknesses might be. The range of complications already brought up suggest to me that theory might not be a very accurate predictor of outcome.


Theory? No, the discussion is how they work. How GPS receivers compute speed is fairly well described in this thread.

Empirical measurements? So because of the ADR you will find vehicles with speedos off from 0 to 10%, and for GPS what do you propose to use to check their accuracy beyond trust, unless you have a mate with a police certified radar or lidar gun? If you do, please get him invovled!


There are a few timed check distances & overhead radar checks around. They are generally flat & level. Done those checks many times. Maybe not 100% accurate, but GPS consistently within a digit.


Not arguing your point, and a bit dated and might be subscription walled, but


I was talking about predicted versus measured outcomes. If you don’t like the word ‘theory’ we don’t have to use it, makes no difference.

How about the implementation in the mass produced car mapping system whose primary aim is to get you from A to B not measure your speed? Could there be any short cuts or mistakes that don’t make best use of the principles? You point out situations yourself where GPS may not be as accurate as you would like.

You are probably right. But you may be wrong, but we will never know until it’s measured. I don’t have a specific method for conducting the test but that is immaterial. I am just pointing out that what is supposed to happen, what is expected to happen, doesn’t always turn out that way in this wicked world. Sod’s law applies.

Looking at the resources, expertise and tried and true manufacturing systems used to produce the Hubble telescope you would predict that it would work properly when deployed. This was not a case like the mass produced car mapper where the result was incidental, it was a case where a great deal was riding on getting it right. A wise man would bet on it. Many did.


But that is the crux of what you asked for. I suspect you are more interested in the game than the topic :wink:


That is where we agree to disagree.


Based on the relatively cheap GPS implementations you can get and how easy they would be to fit to vehicles just to measure speeds, you would have to consider why this isn’t more in place in vehicles to supplement speedo readings. Places where GPS is not usable would render the use impractical such as underground, in covered areas and similar but as it would be a supplementary reading you could revert back to the more mechanical measure as needed. If they were so accurate then you would perhaps think that ADR would regulate that new vehicles would have this system mandated. But for the most part they don’t and I guess (I have no basis to assert fact so is just pure conjecture) this is partially based on the error margins that @TheBBG linked to in an article above (while these reference flight safety the error margins are the same whether in flight or on the ground):

“SPS accuracy would be of the order of 25m (2 drms, 95%) in the horizontal plane and 43m (95%) in the vertical plane…//…PPS provides a predictable accuracy of at least 22m (2drms, 95%) in the horizontal plane and 27.7m (95%) in the vertical plane…//…PPS is primarily intended for military and select government agency users. Civilian use is permitted but only upon special U.S. Department of Defense approval”

What most get from their phones and similar is “local” stationary and wireless reference points as well, these add to the accuracy. Again some of the accuracy concerns of pure GPS were addressed in the article (bold added by me):

“for other countries with large infrastructures, the GPS signal does not meet the accuracy, integrity, availability, and continuity requirements critical to safety of flight. Enhancements to the Global Positioning System (GPS) such as the Wide Area Augmentation System (WAAS) and Ground Based Augmentation System (GBAS) provide the necessary corrections for meeting safety-of-life flight requirements”.


A police friend once told me when I mentioned the tyre diameter error that you’ll almost never get charged first time if you drove to the speed limit according to your speedo reading, provided police could confirm that the error existed.


A NRMA article noted that Speedometers fitted to modern cars have an accuracy of +or- 10%. There is no law that mandates greater accuracy. Vehicle electronics are 21st century, the humble Speedo is stuck in the last mellinium.