Exploring Local
Mike Dobson of TeleMapics on Local Search and All Things Geospatial

Dopes, Nincompoops And Other GPS Users

December 30th, 2009 by MDob

I know that I promised to share my thoughts on the types of strategies that Google might be using to update their map database of North America. The topic is complex and I have been working my way through several models, but am not yet satisfied with my efforts.

In addition, I have been sluffing around doing the kinds of family things that seem to pop-up at Christmas. Today, however, I was shocked out of my reverie, by an email from my friend Duane Marble that contained a link to an article from PC World Online about the folks that relied on the GPS to get them home to Reno from some place in Oregon, only to wind up lost and imperiled on a snowy, local road. Ok, so now you know – this is a Rant!

The article that brought this on was trivial, but the contributions of the pundits who decided to comment on the blog caused me to seriously consider finding a method to euthanize them. Somehow, without the benefit of formal study, a group of technology experts has evolved who want to tell us how technology works, how maps work and how GPS works, without any of them actually possessing an in-depth understanding of any of the subject matter areas involved in the process.

The first item to catch my eye was by a guy who claimed he used his Droid cell phone for navigation because it would always have new maps and the maps on his Garmin were old. He thought the Droid was a better idea because the cloud would always contain the most up-to-date data. I guess that would be true, but the latest, up-to-date map data does not mean correct, navigable map data that is fit for use in routing engines designed to provide turn-by-turn directions.

Well, the same fellow especially liked his Droid since he could use Google Maps’ Satellite View to see what the route looks like before taking the suggested path.
Apparently he believes that Google has its own satellites that they redirect to take a new photo each time you request a view – especially so you can see if it has snowed. Well, they do, …don’t they?

I guess he probably believes that Google also has its own cellular infrastructure that works anywhere in the world, so that their Droid phones will always be able to access Google to retrieve those new maps and those up-to-the-minute satellite images. In addition, he might even think that Google has launched its own GPS satellites in a larger constellation, with better coverage and stronger signal strength than that provided by the Department of Defense, allowing the $2 high-quality GPS receiver in his phone to unerringly access GPS for PTN calculations (Position, Timing and Navigation).

OK. So none of this is true – no Google satellites, imaging systems or cell network. Instead Dingerbob needs to be told that cell phone reception coverage quality is variable and in some locations non-existent. GPS signals are weak and signal reception can be negatively impacted by weather conditions, foliage, and obstructions to the signals, including topography, among other variables.

Satellite imagery can be more out-of-date than paper maps and in heavily wooded areas you may not even be able to see the roads. While everyone would like to have imagery that reflects leaf-off conditions and clear, cloudless days, there are tons of images that do not meet these conditions. Moreover, you never get a leaf-off conditions in forests composed of evergreens – like those so common in Oregon. Finally, the way that Google and everyone else seam images together, you might not be able to “see” anything on the images and, to add to the problem, the imagery used by Google and Microsoft rarely is properly registered to the geometry of the map bases which they overlay. Can we say “Orthorectified”, Mr. Rogers?

Of course, some of the pundits who responded to the PC World blog clearly thought that the problem in the report was that these two old fogies did not understand the technology and should not have been paying attention to the GPS.

Oh, please! The “You should have known better argument” sounds like the one used by Garmin (whose device was in the car being driven by those who became ”lost”). Quote the Garmin Spokesperson (you can find these comments in a graphic attributed to Garmin on ABC News “…Drivers must always remember that GPS devices provide route suggestions; they do not cause drivers to make driving decisions.” Say what?

Crafted by a lawyer I guess. And, I must agree that neither my TomTom nor my Garmin have ever actually reached out and taken control of my vehicles. To be honest, I had not realized that if my PND told me to turn right at the next corner that, when I arrived at the next corner and made sure that I could turn right (hmm, street is available to the right signal is green, no pedestrians, no obstructions, no vehicles in lane), I should ignore the $300 device and just keep going. Wait, why did I buy a PND? Wasn’t it something about navigation? I suppose Garmin will now be rebranding its products with the tag “GPS – Gets People Somewhere”.

Apparently the spokesperson for Garmin has come up with a new and unique definition of turn-by-turn directions. Read my lips – OK read my writing instead. People buy Personal Navigation Devices to help them navigate roads in areas with which they are unfamiliar. Often, the people who buy PNDs are the same people who are not very good at reading maps, which is exactly why they want a device that can perform map-use functions. If PNDs do not perform a wayfinding function, what possible use do they have in this universe?. Unlike paper maps, they are too rigid to be used as toilet paper and too large to be used as a suppository. Gosh, I think their only use is, now let me get this just right, as a PERSONAL NAVIGATION DEVICE. How quaint. You mean these things are really supposed to work? If that is the case, Garmin’s and TomTom’s stock prices will never recover.

Of course, the Garmin spokesperson may also be the person who supplied the comment in the PC World blog that “GPS Systems are fine. Some people, however, can be really, really, clueless.” Of course. User error! How convenient.

In the formal statement actually made by the Garmin spokesperson, the quote continued “It is the responsibility of drivers to exercise common sense at all times when driving, including deference to the posted road signs and road conditions.” Perhaps the Garmin spokesperson should have added “No, our devices do not work when you need them to do so. We suggest that you throw the friggin thing out the window when it doesn’t work.” Or, perhaps, the spokesperson should have said “Wadda ‘ya expect for 300 bucks, something that works as well as a two-buck map?”

Before all of you go off in the wrong direction (Yuk, yuk – it has taken entirely too long to use that gem in this piece), what we are talking about is “Fitness for Use”. Consumers have a right to expect that devices they buy perform the task that the manufacturer represents it as performing. Companies that create flawed products that don’t deliver the advertised functionality can be sued for non-performance. Better yet, products that fail in their intended operation usually fail in the marketplace.

But enough of my rant. The interesting question is “Why does a PND fail to calculate the correct route?” I think the answer is relatively straightforward and believe that the problems are on the data side more than on the algorithmic or hardware side of the product. My list of Fail Factors, in order of decreasing importance, is as follows:

Map data incomplete
Map data out-of-date
Map data incorrect
Map data ambiguous
Routing algorithm bugs/inefficiencies/tuning
GPS receiver inefficiencies
GPS signal degradation (weather, foliage, obstruction, etc.)
PND hardware design inefficiencies
PND software design inefficienies
User error (always a possibility – rarely the root problem)

If you are a frequent reader of this blog, you know my belief that most navigation problems result from map data that is inadequate for navigation, address finding or the solution of other problems that depend on data related to the position and description of geographical features. That is why creating a viable and effective map updating process is so important.

It is clear that the models used by NAVTEQ and TeleAtlas result in the provisioning of less than satisfactory navigation data. However, before NAVTEQ and TeleAtlas started compiling navigable map databases twenty-five years ago (or so), the situation was even worse. Both NAVTEQ and TeleAtlas have spent a great deal of time thinking about the map update problem and have spent considerable sums of money trying to improve the quality of their map databases. I think they have made great strides towards solving many of the problems that plague compiling accurate and up-to-date navigable map databases.

Google, however, was not satisfied with their efforts and thinks that it may have discovered a better way to improve the quality of these databases. I think Google is headed in the right direction, but anticipate that it does not have the experience to remedy the problems it so freely criticized before it became a producer of navigable map databases. But more about that next time.

My best wishes for a Happy New Year and may next year be your best year ever.


Bookmark and Share

Posted in Garmin, map updating, Navteq, Personal Navigation, PNDs, routing and navigation, TeleAtlas, TomTom

2 Responses

  1. Toby

    Thanks for this. Too many articles simply assume that anyone who gets lost/stuck following their GPS is an idiot.

  2. Don

    I didn’t read the stories carefully, but I think I read that the users discovering that they could save 35 miles of driving by choosing the shortest route rather. In an area with lots of logging roads this is a poor choice even in good weather. Could they have set up the PND to the “shortest” option rather than the “fastest” one?

    “Thanks for you comment, Don. I agree that user error is always a possibility and in the case referenced, the users could have set the option to shortest, rather than to fastest, but I do not know if this was the case.”

    “What I find curious about all the press that this issue is attracting is the constant suggestion that taking a paper map with you is a possible remedy for the inadequacies of PNDs and the map databases they use. Paper maps are not the solution to navigation probems. The answer is in the production of more accurate and more comprehensive navigable map databases and data mining the paths that people use to travel between places. If the map database cannot solve navigation problems, perhaps “collective” intelligence can help create useful navigation products.”