This site uses cookies to offer you a better browsing experience. Find out more about how to manage cookies, or dismiss this message and continue to use cookies.

GPS or Speedometer

At highway speeds my Nuvi 760 conistently reads 5 MPH slower than my speedometer...which is more accurate? But, really officer, my GPS had me going just 65 :roll:

Comments

  • gatorguy 326 Points
    More than likely the nuvi. Are you still running the original size tires that came with your car?
  • vleek 0 Points
    edited February 2008
    Tires also may be underinflated. Actually changes in tire size should produce proportional ratio GPSspeed/CarSpeedometer. If the readings are identical at lower speed, then it's probably wrong car speedometer.
  • Tim 1480 Points
    There is also a thread about this here.
  • Spyder63 331 Points
    To check your speedo why not just time yourself while on an interstate against the Mile Markers? At 60 mph it should take 60 seconds. You do the rest of the math. You just need to find a stretch where you can actually go just 60 for a short time without getting rear-ended or a lot of single digit salutes. :wink:
  • I definitely go by the GPS more than the speedo.

    In fact, I used to do custom tuning on a truck I had and used that to dial it in.
  • Tim 1480 Points
    I'd take Spyder63's suggestion and I wouldn't go by the GPS. I've taken four GPS devices, put them on the dash, driven a consistent 60 mph on the Interstate and they all showed a different (yet steady) speed.
  • Here's an interesting article GPS speed vs Radar Gun:

    http://www.news.com.au/heraldsun/story/0,21985,21999706-661,00.html

    With 24 orbiting satellites using triangulation, and GPS speed being calculated using an atomic clock, not some old worn out treads on those tires … Guys, forget about speedos on your car … Only laser guns can claim to compete with GPS accuracy and even them have lost in court.

    Cheers!

    HappyTrails! Drive Safe!
  • Tim 1480 Points
    I totally disagree with the results of that article, unfortunately. GPS has a huge margin of error that can't be ignored in cases like that. If they are so accurate then why do multiple GPS devices in my car each show a different speed?
  • I totally disagree with the results of that article, unfortunately. GPS has a huge margin of error that can't be ignored in cases like that. If they are so accurate then why do multiple GPS devices in my car each show a different speed?
    You're probably right - I just came across it by accident, thought I would pass it on. One reason why multiple GPS devices in your car each show a different speed may tie in with differences in technology, like the number of correlators for example and I'm sure you can think of other features that one cause one to perform differently than the other.


    Cheers!

    HappyTrails! Drive Safe!
  • I can understand how a speedometer could vary but I don't understand how a GPS could unless they calculate speed using slightly different formulas. (Which would be odd.)

    I've clocked my speed using:
    My speedometer
    Previous Lowrance iWay 350c
    eTrex Camo
    gps12
    Previous nuvi 350
    Previous nuvi 360
    Current nuvi 660
    Current Freedom Keychain GPS through my Motorola Q

    The only one that gave a different reading was my speedometer.

    Not saying your findings are inaccurate, just posting a different result.
  • Tim 1480 Points
    I think some of them are perhaps better of taking into consideration their own processing/calculation times and how that might impact the calculated speed. Some GPS devices are also better than others at throwing out junk data than others. But yes, it is weird to be driving down the road and have my speedometer say 60, one GPS saying 61, and another GPS saying 62, and another GPS saying 64.
  • I just realized most of mine were the same brand, Garmin. Maybe that has something to do with it?

    What brands did you compare?
  • Tim 1480 Points
    I had the same thought when I looked at your list and indeed I would expect a similar results from the same brand. I've been comparing my car (Outback Limited sedan) with most of the different brands on the market which is likely why I saw such a discrepancy.

    Something else probably related is to look at the speed your GPS shows while it is "at rest". Now many of the auto devices are likely programmed to show 0 when it is close to 0. But for example I just turned on my 760 and have left it in place for the last five minutes and it shows an average speed of 1.7 mph with a max speed of 3.1 mph. That is likely a good indication of how the margin of error of GPS impacts the speed readout.
  • Yeah, good point there... I know the Freedom can read .1 - .7 mph when I stop. I thought the reason was due to the fact that it was still swinging back and forth on my keychain lol.
  • @Tim: "But yes, it is weird to be driving down the road and have my speedometer say 60, one GPS saying 61, and another GPS saying 62, and another GPS saying 64."

    I'm surprised that your speedmeter showed a lesser number, I would have expected the opposite. It's a known fact that most car manufacturers have cranked up their speedos for a number of reasons, lawsuits for one. I think it's fair to say that that dial is an indicator, NOT a precise measurement; such that if you're showing 110 mph, odds are that you're probably doing only 109 or 108 or 107 - the variance can be as much as 5 mph depending on the manufacturers. There are some exceptions, the 2007 Toyota Camry for example is dead on - why this one and not others! It beats me!

    Cheers!

    HappyTrails! Drive Safe!
  • Tim 1480 Points
    Yea, sorry I had it backwards in that example, my car was reading the highest speed. I'll see if I can go do another test today with a few different devices.
  • Spyder63 331 Points
    In the state of Pennsylvania only state troopers can use radar and by law they have to give a 6 mph leeway. If you are doing 71 in a 65 zone you are "safe". Seems the legislators are convinced vehicle speedometer are not accurate. :roll:
  • gatorguy 326 Points
    Apparently, the margin of error also varies between units within the same model. I turned my 760 on after reading Tim's post of his experience earlier. Showed no movement in over 3 hours of sitting still.
    0.00 mph max speed
  • I definitely go by the GPS more than the speedo.

    In fact, I used to do custom tuning on a truck I had and used that to dial it in.
    thats a great idea, im a tuner also 8)
  • What software/hardware do you use? I used HPTuners, mostly for GM's but when I got out of it they were expanding quite a bit to other manufacturers.

    /hijack 8)
    I definitely go by the GPS more than the speedo.

    In fact, I used to do custom tuning on a truck I had and used that to dial it in.


    thats a great idea, im a tuner also 8)
  • Yea, sorry I had it backwards in that example, my car was reading the highest speed. I'll see if I can go do another test today with a few different devices.
    Have you had a chance to run those other tests TIM. This is an interesting issue, so was just wondering what you came up with.

    Cheers!
  • Spyder63 331 Points
    Of all of my vehicles I have found that only my Cadillacs always agree with my GPS. The rest vary by 2-4 mph, pretty much on the low side. My '63 Corvair speedo reads WAY high at speeds above 35mph so I always carry a GPS and so far have not been pulled over. For awhile I was impressed with how well it handled at 90-100 mph. Then my bubble burst when I discovered I was only going 65-70! :shock:
  • I ride a motorcycle and when I use my GPS the speed differential between speedometer and GPS is quite significant. I decided to calculate the speed according to the RPM of the engine adjusted by all the different gearing ratios to determine the rear wheel speed based on the actual circumference of the wheel. According to these calculations my speedometer is accurate making the GPS wrong. If the mechanical calculations say the speedometer is accurate why does the GPS differ so significantly?
  • Tim 1480 Points
    The simple answer is that the GPS is not displaying an instantaneous speed but rather a figure calculated based on the distance traveled (GPS has an error of 10 meters) over the time. Also see this thread: Trust Speed Reading
  • SkipD 91 Points
    I've always assumed that a GPS speed reading is only accurate on level roads (no incline). Because the elevation readings that I've seen on various GPS receivers that I've had vary a lot, it's likely that only two dimensions are used when the speed is calculated.
  • My GPS test was done on a straight flat stretch of freeway keeping a constant speed per the rev counter. At 120 km/h the GPS shows 112 km/h which is a 6.66% variance. This surely can't be accounted for by readings every 10 metres. The test was done over a distance of at least 1km.
  • popej 57 Points
    Speedometer in my car shows always about 5% less then GPS. Whenever I pass on a road a speed camera with a display, I compare its reading to speedometer and GPS. GPS is more exact.

    Actually GPS calculates instantaneous speed. Position is calculated based on phase of received signal (pseudorange) and speed is calculated the same way based on Doppler frequency of received signals.
  • I've always assumed that a GPS speed reading is only accurate on level roads (no incline). Because the elevation readings that I've seen on various GPS receivers that I've had vary a lot, it's likely that only two dimensions are used when the speed is calculated.
    I wonder if it does measure speed straight up. I think it would, especially if the doppler shift theory applied...
  • Tim 1480 Points
    This surely can't be accounted for by readings every 10 metres.
    I didn't say it only takes readings every 10 meters. I'm saying the accuracy of each position calculated can be off by about 10 meters. (Consumer GPS without WAAS is accurate to within 10 meters about 95% of the time.)

    So if the position is off by 10 meters behind your actual position one second then 10 meters ahead of your position the next second you have an error of 20 meters over a one second gap. That will easily account for speed calculation errors regardless of terrain.

    As to vertical travel, a little Pythagorean Theorem math will show that it is irrelevant to the calculation, even in mountainous driving, given the horizontal errors already mentioned.
  • Tim 1480 Points
    The GPS speed isn't accurate because the GPS isn't a speedometer. I've had about a dozen GPS devices on the dash of my car and they will all vary in what they show for a speed, even with a constant speed on straight road.
  • alanb 539 Points
    I don't know the answer to whether the GPS bases it's speed on actual distance between 2 points or horizontal distance (ignoring elevation change), but if I did my calculations correctly, the difference in horizontal vs point-to-point speed at 100 MPH on a 10% grade would be about .5 MPH, on a 20% grade it would be about 2 MPH.
  • Speedometer in my car shows always about 5% less then GPS. Whenever I pass on a road a speed camera with a display, I compare its reading to speedometer and GPS. GPS is more exact.

    Actually GPS calculates instantaneous speed. Position is calculated based on phase of received signal (pseudorange) and speed is calculated the same way based on Doppler frequency of received signals.
    I don't think typical consumer GPS's use Doppler - the signal processing becomes expensive. In fact, I think speeds are determined by a simple filtered derivative of successive position data. The filtering will reduce sample to sample "noise" errors. Bias errors will be effectively subtracted out.

    In more critical applications (military) there may be an advantage to using the Doppler returns along with the base derived position data and combining these measurements in a more sophisticated filtering scheme (i.e, Kalman filtering).

    Here is a link to a discussion of this topic (this "speed calculation" question seems to pop every year or so) that is hopefully more informative.

    http://www.gps-forums.net/speed-calculated-nuvi-660-a-t42724.html
  • popej 57 Points
    I don't think typical consumer GPS's use Doppler - the signal processing becomes expensive.
    I don't think it is expensive. Probably computing signal correlation is the big task and frequencies comes as one of the outputs of this computing. Then receivers needs only to solve some vector equitation to get speed vector.
    In fact, I think speeds are determined by a simple filtered derivative of successive position data. The filtering will reduce sample to sample "noise" errors. Bias errors will be effectively subtracted out.
    I think that it rather goes opposite way. Having speed and position receiver can extrapolate next position and use this value to do some filtering of position noise.
    In more critical applications (military) there may be an advantage to using the Doppler returns along with the base derived position data and combining these measurements in a more sophisticated filtering scheme (i.e, Kalman filtering).
    I believe Kalman filtering is a standard for GPS.
    Here is a link to a discussion of this topic (this "speed calculation" question seems to pop every year or so) that is hopefully more informative.

    http://www.gps-forums.net/speed-calculated-nuvi-660-a-t42724.html
    Yes, I know this topic repeats :)

    Somehow I think that idea of computing speed form successive position appears as a best guess form people surprised by the question. Not that I know GPS algorithms, but after reading some more detailed description of how GPS works, I tend to assume that Doppler frequency is the obvious way to compute speed.
  • I don't think typical consumer GPS's use Doppler - the signal processing becomes expensive.

    I don't think it is expensive. Probably computing signal correlation is the big task and frequencies comes as one of the outputs of this computing. Then receivers needs only to solve some vector equitation to get speed vector.
    In fact, I think speeds are determined by a simple filtered derivative of successive position data. The filtering will reduce sample to sample "noise" errors. Bias errors will be effectively subtracted out.

    I think that it rather goes opposite way. Having speed and position receiver can extrapolate next position and use this value to do some filtering of position noise.
    In more critical applications (military) there may be an advantage to using the Doppler returns along with the base derived position data and combining these measurements in a more sophisticated filtering scheme (i.e, Kalman filtering).

    I believe Kalman filtering is a standard for GPS.

    Here is a link to a discussion of this topic (this "speed calculation" question seems to pop every year or so) that is hopefully more informative.

    http://www.gps-forums.net/speed-calculated-nuvi-660-a-t42724.html

    Yes, I know this topic repeats :)

    Somehow I think that idea of computing speed form successive position appears as a best guess form people surprised by the question. Not that I know GPS algorithms, but after reading some more detailed description of how GPS works, I tend to assume that Doppler frequency is the obvious way to compute speed.
    The Kalman filter setup for this kind of problem is computationally expensive. You could have around seven measurements (derived position components, derived range rate components from the Doppler, time bias, etc.). It would also probably have to be an extended Kalman Filter because of the non-linear nature of the measurements.

    I still think that commercial GPS's simply take 4 measurements (per sample time) in order to get a set of non-linear equations where the unknowns are the vehicle position components and a time bias - satellite positions are known from downloaded data, typically in an Earth Centered Earth Fixed (ECEF) coordinate frame. These equations can be iteratively solved for the desired unknown variables using simple, established numerical methods. Velocity data is then obtained as a filtered derivative.

    An alternate (non-Doppler) way to get the velocity components would be to set up the problem by differentiating the position equations to obtain a set of non-linear eqs. in the unknown vehicle velocity components (using the position components previously obtained). This results in a set of equations that use the angle information to each satellite (direction cosines) and the known satellite velocities and positions to iteratively solve for the velocity components. A lot more complicated that a filtered derivative.

    Anyway, just my best guess of what is actually being performed.
  • As I stated in the other thread, motorcycle speedometers are made to count revolutions of the front wheel, and they are notoriously optimistic - by as much as 10% or more. If you maintain an indicated speed of 60 mph, for instance, your actual speed will vary depending on the amount of rubber you wear off of the front tire, and the lean angle of the bike (lean the bike over and the effective circumference of the tire - where the tire contacts the road - decreases) among other things. I don't know about internationally, but here in the USA speedometers are allowed to read optimistically (indicated speed is greater than true speed), but it's illegal for them to be pessimistic (indicated speed is lower than true speed) so the manufacturers make sure that if a speedometer is inaccurate that it will never give a pessimistic reading, so that a driver cannot use inaccuracy as a defense for a speeding ticket. Since a GPS is calculating speed based on location and time it will be as accurate as its position measurement - it will be much closer to your true speed, based on timing a run between accurate mile markers.
  • Tim 1480 Points
    I'm locking the other thread as redundant.
    Since a GPS is calculating speed based on location and time it will be as accurate as its position measurement
    That is a factor-- there are also accuracy errors in the calculation itself. I've had a dozen GPS devices on the dash at once and disagreements with my current speed (while maintaining a speed) could differ from one another by +/- 4 mph at times.

    I consider GPS not very reliable when it comes to speed readings. Heck, I went for a hike over the weekend and it said my max speed was over 22,000 mph with a few other times when I was hiking over 70mph. :)
  • jonnop 101 Points
    I consider GPS not very reliable when it comes to speed readings. Heck, I went for a hike over the weekend and it said my max speed was over 22,000 mph with a few other times when I was hiking over 70mph. :0
    If you are that fast why did we not see you in the London Olympics. Sorry, silly of me -at those speeds we would not have seen you! :D
  • alanb 539 Points
    I have also seen the wild speeds mentioned by Tim when standing still or moving at low speeds, especially when hiking under dense canopy or in urban canyon condtitions. But on the open road at normal driving speed, my experience with my 3 Garmins is that the GPS speed is very consistant and more accurate than my car's speedometer, which always reads 2 to 3 MPH faster than the GPS.

    Another interesting observation is that I have an UltraGauge device on my Hyundai Elantra. It is an OBD2 connected device that monitors many different data elements on the car's electronic control system. The UltraGuage speed matches my Garmin speed very closely. And since it has access to the same data as the car speedometer, it makes me think that Hyundai calibrates the speedometer to show fast on purpose.
  • Car speedos can be off by 10 Km/h in Canada.That is the acceptable legal
    standard.I've used my Nuvi 755t in several cars.It never quite matches the car speedo,its either higher or lower.Thats because every car speedo is a bit different from the other.With all the tech. involved I trust my GPS more than speedos in cars.
  • If the guy or girl in the little box says "You are over the speed limit" and you heed the warning and as a result do not get an expensive speeding ticket then mission accomplished so why worry if gps or car reading is off by +\- a couple mph? :? :lol:
  • Still a good idea to check speed limit signs on the road. I've noticed that CNNANT 2013.2 gets about 50% of the speed transitions at the right places, and it misses all of the 80 MPH sections on I-15 in Utah.
  • Regarding the 10m accuracy rating. Personally I don't think that 10-meter error is random, but would be similar along adjacent points of a track. In other words, if my position fix were 10 meters west actual, it would be 10 meters west of actual on the next however many position fixes. It would gradually drift in another direction, becoming very close to actual, or 8 meters north, then drift toward the south, whatever. A track would show a curve weaving around my actual track, and not bobbing back and forth.

    That's just what I think happens, nothing at all scientific or measured about it.

    I do know that I trust my GPS receiver's speed reading more than my speedometer's, based (as another post earlier) on passing "Your speed is" radar signs, as well as timed passes through mile markers on Interstates.
  • I certainly trust mine WAY more than my speedo, too.......however....disregarding accuracy errors, the GPS only knows where it IS when it updates, not how it got there. If you were able to let it update then drive forward a distance then back up to your update point, the GPS would say you haven't moved (disregarding accuracy errors.)

    I experienced this frequently while mountain biking off road. If I was riding on a relatively straight road, the GPS and my bike computer would record distance pretty comparably. BUT, if I was off road on switchback trails, the GPS would always show less distance traveled than the bike computer did. I rationalized this as the GPS measuring a straight line distance between it's update points, whereas the computer recorded actual distance around the curves.

    In marine charting, it's not unusual for the position fix to show a docked boat jumping from one side of the dock to the other every second or so based on accuracy errors.

    We also experienced the same effect tracking Iditarod dog sled competitors. The actual trail distance as measured by snow machine odometers was always greater than their GPS distance. For that reason (among many others) no one can tell you the actual distance of the race.
  • Boyd 1985 Points
    You lost me on all that. :) But automotive GPS units have something called "road lock". Since the device is intended for use in a car, it assumes you are driving on a road and if the actual position is not on a road the software thinks it is just an error in the reading. It will then "snap" your vehicle position to the nearest road and this is not what you want when travelling off-road.

    If you use a non-routable topo map, then road lock will not be supported and you should get more accurate results.
Sign In or Register to comment.
↑ Top