Latest XT routing weirdness

[danham]

Yesterday I was traveling from New Hampshire to our home on Cape Cod. I know the way, but always try to use the XT so I can learn its quirks. It didn't disappoint.

There are two possible highway routes I can take and a last minute decision based on traffic is called for. Before setting out, I used Trip Planner to create a route choosing one of them and let the XT do the navigating. It all went fine until I reached the "Stop" in my route (actually a Via) that I had inserted to force the XT to choose Route 128, not 495. It told me to exit from 128 and do all kinds of weird turns. Traffic was heavy, so I quickly stopped the route and told the XT just to take me to Home, figuring that took the least bit of fiddling with the device while driving.

Instead of continuing on 128 it kept trying to get me to exit and take local roads. I couldn't divert my attention enough to pull up the map and see what it was up to precisely, but it had the look and feel of trying to send me back to that Via, even though this was a brand new route. I stopped the route and hit Home again, and it worked properly for the rest of the trip.

It is certainly possible that I hit a wrong button while trying to avoid "creative" lane changes being performed all around me at 75 mph, but it sure seemed like that Via was "stuck" in the XT's brain and it wouldn't let go until I started from scratch again.

-dan

[Peobody]

@danham, do you have traffic enabled? The one time I experience behavior similar to what you describe was due to a reroute resulting from traffic on an interstate. That reroute was inappropriate in many ways but your description reminded me of that experience. I recall being notified by the XT of the route change but I don't recall whether that was audible, visual, or both.

[danham]

Yes, I do have traffic enabled. Interesting. Obviously I did not choose a traffic-related detour, nor did the traffic function alert me to any congestion prior to the error, but that's an interesting idea.

-dan

[Fourbears]

Just to jump on this thread I also experienced routing weirdness last week. On the map the blue line is the route I planned in MyRouteApp. I deliberately left out shaping points before number 16 thinking (wrongly as it turned out) that the XT would probably take that route. However, it decided that it was quicker (!) to take the route indicated by the red line which is over 20 miles longer. If I put the route from the M25 into Google maps it uses my route as it is the quickest and shortest so I have no idea why the XT thinks it’s route is quicker/shorter/whatever! I actually ignored it once I realised what it was doing and took a parallel route to 16 which gave me a bit of XT RUT before it calmed down and got on with it. Weird!

IMG_0126.jpeg (783.16 KiB) Viewed 1449 times

[Peobody]

The emphasis by the XT on faster roads is truly frustrating. Its algorithm identifies roads with the fastest travel speed that will get you to the route point without consideration of slower roads that will get your there quicker.

[rbentnail]

The emphasis by the XT on faster roads is truly frustrating. Its algorithm identifies roads with the fastest travel speed that will get you to the route point without consideration of slower roads that will get your there quicker.

Took a while for me to learn this, last year I guess. I pressed Go Home knowing full well how to get there. After all, I was on the road I live on, about 80 miles away from home. You'd think traveling down the road you live on would be the fastest but no, the XT wanted to get on roads with faster speed limits. It wanted me to go some 50 miles and 45 minutes longer to get home "faster". That's when I began searching for a way to choose neither faster nor shorter so it would calculate "normal" . Haven't found a way to yet. I wish Garmin would fix this bug.

[lkraus]

I'm not even convinced the XT is picking faster travel speeds. I know from experience the fastest route home from a local store and the shortest.

The fast route is 9.4 miles and uses a 60 mph highway slightly south of east and then our 55 mph semi-rural road north. Three traffic lights, three stop signs. Google's 13 minute estimate is accurate.

The short route is 8.4 miles, uses half of the highway east and then sort of cuts the corner northeast on another 55 mph road. Three lights, four stop signs, 14 minutes.

The XT route is 9.3 miles, uses half of the freeway east and then goes due north on a 45 mph road before going east for 2.8 miles on a 25 mph residential street through a small city (and two school zones at 20 mph during school hours), before becoming 35 mph for 1.8 miles before reaching our 55 mph road. Seven traffic lights, 19 minutes. Much longer in late afternoon, as those trying to turn left hold up everyone for 2-3 cycles at one intersection.

My XT is always set to Driving mode, fastest route, but it is not picking fast routes, nor short routes, nor even using the roads with the highest speeds. And it never acknowledges the traffic slowdown at that one light, though there are ways around it.

[jfheath]

I have a theory. Rather a grand word for an idea that has popped into my head from casual observations.

Somewhere in the map data is something that classifies the type of roads. I believe that there is still historic 'TrafficTrends' data which was used by the Z590 to calculate best routes on particular days of the week and particular times of day. This is true.

So the Zumo knows where faster roads are, but they may not be simply the Major roads (but it knows those too).

The inkling that I had was that the route calculation looks at the time to reach the nearest faster road that is heading towards the destination and selects that route. So eg if there are two possible ways to reach a destination 12 miles away heading north and one of them involves heading 6 miles west to reach a faster road heading north, then it believes that 6 miles west on slower roads to reach a faster road is better than 12miles north on roads that are not labelled faster.

I don't think that it looks any further ahead than that, because I have seen it do the equivalent of 6 miles west, 12 miles on fast roads then 6 miles east.

As I said - pure conjecture, but if I was to start looking at this 'for fun', I'd start with that thought in mind and then try to prove it wrong.

[danham]

@jfheath,

Some years ago I got to talk to two Garmin employees, both of whom confirmed that among the factors embedded in the maps and used to calculate ETA and "fastest" is what here in the US is known as "level of service." It is a speed based on road width, design, sight distance, number of intersections, curves -- a whole bunch of things that add up to what the road COULD support in terms of safe driving speed. It may or may not be the same as the posted speed limit, though that is of course also one of the factors.

In my previous Garmin units, there was a definite bias toward numbered routes with higher speed limits (and perhaps levels of service), but the XT seems to have a much stronger bias.

-dan

[Oop North John]

What I've noticed on my recent trip to Spain where the XT and it's version of fastest was seen to be anything but ie over 10 minutes longer over 30 minutes of travel was that the XT very quickly worked out the actual fastest route in only a few hundred yards of going off it's route attempt at the "fastest" route. I\

t's almost as if the XT has got a set of blinkers on, looks for the "fastest" road(s) nearby and works it out from that, then not even trying to see if there's a better option. Or, the algorithms are so useless and power hungry that it would crash if it looked for alternatives.