Just Ordered Two 2014 Titanium Energi

[openair]

What you're claiming is simply inaccurate.  If true, companies like Garmin wouldn't exist.  Since the same car with the same tires, and now updated firmware tracks without issue, are you going to stand by your statement that it's the tires??

 

This is a very strange statement.  There is a TSB above describing how the tires size can affect navigation and a suggestion to reduce recalibration time after installing new tires.  You, yourself, state that the Ford gps uses tire rotation as part of its location calc and only question why it would bother with rotation data.  Now someone tells you why and you're back to questioning if the tires are involved at all?

[larryh]

The following chart shows smoothing of GPS data.  Here 321 independent measurements were taken to come up with the final GPS location.  The car is stationary.   I only have the smoothed measurements, i.e. average of the previous measurements.  So at measurement 20 on the plot, you see the relative GPS error based on averaging 20 independent measurements.  The individual raw measurements jump all over the place.  Initially, with less than 10 measurements, the location is off by more than 10 feet relative to the final calculated value.  Because the individual measurements are very noisy, it takes many measurements to converge on an accurate estimate. 

 

Convergence is much faster if you know the car's speed, i.e. tire rotation speed.  Without tire rotation speed, GPS smoothing is going to have to calculate both speed and position, which takes significantly longer to do than if you knew the exact speed of the car.  If you are accelerating or braking, then smoothing takes all that much longer to come up with an accurate estimate.  If you know the car's speed at all times, you can come up with a much more precise location much faster.  Convergence will be much faster than what is shown in the chart below under all conditions:  constant speed, acceleration, and braking, 

 

The plot only shows relative error, which is different from absolute accuracy.  The dilution of precision of the measurements was 10, which implies absolute accuracy is only fair.  The final estimated position was off by about 25 feet.  

 

Edited April 12, 2015 at 12:15 AM by larryh

[larryh]

The following plot shows the error in speed computed by GPS during my commute to work.  The speed estimate provided by GPS is off by more than 10 mph at times.  If speed is off, then GPS location is also off.  For each second that passes while the GPS speed is off by 10 mph, location error will increase 15 feet.  So after 5 seconds (the length of time it took GPS to correct the error), the GPS location error is off by 75 feet (in addition to the normal GPS location error).  If GPS has access to speed information, i.e. tire rotation speed, then location is going to be far more accurate.  You definitely want to provide the car's speed to the GPS location estimation algorithm.

 

Edited April 12, 2015 at 01:39 PM by larryh

[Doug0716]

It looks like in most instances when it was off by a significant amount, it spiked back in the oposite direction.  It makes sense as doing it's calculations based on previous inacruacies with current acurate results would result in the inverse of the previous results.  This would allow most standalone GPS devices to maintain a relatively accurate position, while still explaining why, if you have the expected speed, you would want to feed it to the GPS.

 

Though, shouldn't the GPS, given speed information, be able to re-calibrate itself without requiring a reset?  Let's say if it had a consistent speed of 40MPH (allowing for variance of +/- 1MPH or so) or higher for 30 seconds (a minute, 10 seconds, whatever... they could do some testing and figure out what speed/variance and time frame is needed) but it recorded 43MPH average over that time frame.  At that point it should be able to recalibrate that for every 1MPH reported by the car is really 1.075MPH of travel.  I mean it seems that this is what it's doing after you do a reset, I guess I don't understand why the reset is required before it recognizes there is a problem, why not once per power cycle?

[openair]

The TSB above says a reset will greatly reduce the time required for recalibration.  Not that is it required.  It's a form of heuristics.  The new data, with the new tires, will eventually lead to it correcting itself but, unless the reset is done, the old data is not discarded right away.

[Doug0716]

I guess I don't see the value in long term history if you can re-calculate it regularly. If it's driving at a consistent speed (freeway, in OP's case) and every few minutes it has to jump the location back several feet, it should initiate a re-calibration as if it had lost power, pretty simple algorithm to consider.

In any case it sounds like his snow tires (recommended by TireRack, not Ford) may not be within the size specifications for Ford, which may be his bigger issue? (Though I guess I don't see why the GPS would care about any of that, it seems more like Ford is saying if they put another size tire on the car don't touch the car as it could be a liability issue)

[larryh]

Because of the large measurement errors in GPS location, it takes a few seconds for the GPS device to recognize you are accelerating or stopping unless you provide it with the car's speed.  When you are braking, it is going to continue to think you are traveling at constant speed until it has enough evidence to recognize that this is no longer the case.  Similarly, when you are accelerating from a stop, it won't recognize that you have started accelerating until it has sufficient measurements to realize that you have started moving.  Thus you see the large spikes in speed error when coming to a stop and when accelerating from a stop.

 

I have no idea how long it should take the car to figure out recalibration for a change in tire size.  However, I would be very wary of changing tire size on the car, I have no idea how many systems in the car may be impacted.  In the very least, the odometer readings will be wrong. 

Edited April 12, 2015 at 05:25 PM by larryh

[stevedebi]

I ran the GPS diagnostics this past weeking while traveling at a (more or less) constant 71 MPH. The GPS was constantly going back and forth between 71 and 69. I think it is possible that the GPS diagnostics is reporting raw data, and the car software does the smoothing, but I was a bit surprised that it was exactly those two speeds back and forth. I had 10 satellites tracking all the way.

[rmorse]

To OpenAir:

Please re-read the post.  Resetting did nothing for the problem.  We're not talking 100ft difference, or having it self-correct.   It would literally be miles off, and show the car was traveling south, when, in fact, it was traveling north!   I realize the Ford document says it's because of the tires.   I'm suggesting the TBD likely doesn't cover this case since the car now tracks perfectly with the same tires.  I'm just hopeful it doesn't die like before.  As far as GPS data being "noisy", I think you need to look at your receiver.

 

This is mostly a dead issue.  Thanks for all the comments.   Now I just need to get my other 2014 Ford Fusion Energi, the one that's one VIN digit off my other Energi, to work!

[larryh]

Should you be interested in learning more about GPS accuracy, you can visit Garmins web site:

 

http://www8.garmin.com/aboutGPS/

[lonzo71]

The receiver, how many satellites, location, and a whole bunch of other tech stuff affects the GPS. If it wasn't for the other features, I wouldn't have gotten GPS. If you don't know basic land navigation, people shouldn't be driving...sadly, too many that don't, do! :)

Edited April 14, 2015 at 03:26 PM by jeff_h

[stevedebi]

The receiver, how many satellites, location, and a whole bunch of other tech stuff affects the GPS. If it wasn't for the other features, I wouldn't have gotten GPS. If you don't know basic land navigation, people shouldn't be driving...sadly, too many that don't, do! :)

Yeah, with the vehicle nav units these days, I'm glad I learned to navigate the old fashioned way when I started driving. I still find that my sense of direction is vital when "following" the in vehicle GPS!

 

My wife's VW is especially bad at direcitons, far worse than the Ford unit, but better at knowing the vehicle location.