Interesting TomTom routing insights

[ezerhoden]

Interesting look into routing engine. Try these locations and see tomtoms routing change to favor Interstate road.

Start: Glenfield Rd, Edgeworth PA, USA @40.52895?N, 80.1301?W
Finish: Flaugherty Run Rd HWY 3089, Carnot-Moon PA, USA @40.53058?N, 80.23174?W

Details: 7.4 miles 13 mintues

This is the preferred route having driven this route many times.

Start: Glenfield Rd, Edgeworth PA, USA @40.52895?N, 80.1301?W
Finish: Flaugherty Run Rd HWY 3089, Carnot-Moon PA, USA @40.52986?N, 80.23416?W

Details: 7.8 miles 17 minutes

The difference in distance of the finish points is a few hundred feet if that on the same road.
I think this gives some insight into the routing algorithm tomtom uses. The weight applied to total time of the trip and making left hand turn across busy road. The Left hand turn on the faster route across Rt. 65 onto the Sewickley Bridge is what I believe causes the issue here. There is a traffic light but the tomtom does not know this. The turn is safe, but I think this is an example of the tomtom preferring either highway use or hating left turns.

Another routing issue I notice is that if you have a planned route and miss a turn or purposely drive another route missing a planned turn, the tomtom keeps trying to put you back on your initial route overlooking much shorter/faster routes. If you instead tell the tomtom to navigate to the destination, instead of recalculating the initial route, a much better route is found.

I have to say that overall, with IQ routing, tomtom routing is superb.

 

[pianoCM]

Another routing issue I notice is that if you have a planned route and miss a turn or purposely drive another route missing a planned turn, the tomtom keeps trying to put you back on your initial route overlooking much shorter/faster routes. If you instead tell the tomtom to navigate to the destination, instead of recalculating the initial route, a much better route is found.

I think based on its IQ Routing algorithm, Tomtom think the original route is still the FASTEST route to go, if you keep go further, it eventually re-routes the new FASTEST route.

 

[mvl]

Thanks for pointing this out. I've seen this before on other navs, but never the Tomtom.

The beginning and end of the route are the same, but the middle changes based on how far the endpoint is. Usually, this happens when GPSs try to optimize route calculation time by picking some midpoints and routing via them.

I never thought this happened with IQroutes. I had thought that it does a 100% compare of all route options (other than 6+ hour trips), but you definitely found a bug that proves this isn't the case.

Makes me lose a bit of trust in IQroutes. But I agree it usually works great.

 

[gatorguy]

No consumer gps route algorithm does a 100% comparison of all route options. That would take too much time in a lot of cases. Instead they work in from each end of your route as I understand it, sampling closest nodes on possible roads and assigning a value (or cost) to using each of them. That value might be primarily mileage if asking for shortest distance, or travel time/road class if requesting fastest. Some devices also offer "most use of freeways", "least use of freeways", that type of thing. That also results in a value/cost figure. The better the designers assign value figures, the better the determined route. But because not all possible roads at all entry points are sampled, you do find instances where there is, in fact, a faster or shorter way to travel. The route chosen by your device is the fastest/shortest it found.

In reality, almost every algorithm used today is a variation of one of two, in use for years and dating back as far as 1959: Dijkstra?s shortest path or A-Star.