33. Routing Engine Performance

Testing routing engine performance is always tricky. One GPS unit may route correctly in certain instances and incorrectly at other times. To test the 680's routing, I picked a starting address in New York City and a destination address in Southbury, CT. I compared the results of the nuvi to those of the Garmin StreetPilot c550, the TomTom ONE (to see how a device using Tele Atlas maps instead of NAVTEQ would route us), and Yahoo Maps.

Both the Garmin StreetPilot c550 and the nuvi 680 chose identical routes and estimated the trip time exactly the same. Not surprising, since both units share Garmin's routing software and both units use mapping data provided by NAVTEQ.

The TomTom ONE chose a slightly less efficient route, choosing a route that's 14 minutes longer than the nuvi's route. It's not totally fair to compare the nuvi, which retails for just over $1,000, to the TomTom ONE, which retails for less than $500, but the point here is to show that there is more to routing than just mapping data. All three sources have the same roads available to choose from. Yet the TomTom ONE chooses a different route. Why?

Mapping data is a kind of dark art that mapping and GPS companies are loathe to discuss in detail. From what I've been able to determine, NAVTEQ (who provides mapping data for almost every online map service, such as Google, Yahoo, MSN, etc.) provides more details about the roads than Tele Atlas does (Tele Atlas is used in the TomTom ONE). The additional mapping attributes, combined with the decisions a routing engine makes ultimately decides the route the GPS will take. For example, when calculating the estimated time of arrival, a routing engine might decide to discard the posted speed limit on a road and instead use the average speed of travel that NAVTEQ provides. Or a GPS routing engine might decide that more accidents occur when making right turns than left turns, and decide to take a slightly longer, although safer, route. These are just some of the things that can affect how the GPS calculates the "best" route.

The nuvi and Yahoo Maps actually choose identical routes, although the nuvi is more aggressive with its time estimate.

I tested 3 more random starting addresses and destination addresses and found similar results. In most cases the nuvi chose identical routes to Yahoo Maps. In each tested route the TomTom ONE was less efficient at calculating the best route.

Just to verify that the nuvi 680 uses the same excellent routing engine found in the nuvi 660, I compared the two units:

Not surprisingly, the nuvi 680 and 660 produce identical results.

The nuvi has an excellent routing engine. That routing engine coupled with NAVTEQ mapping data makes the nuvi very good at its primary task: navigation.