Frustrating last hour of ride with XT

[Peobody]

Wow, I just returned from a 3.5 hour loop ride and am really frustrated with the XT; also its map.
The map issue was 4.1 miles on a gravel road even though the unpaved road avoidance is enabled. The route was created in Basecamp with same avoidance. The map did not have the road identified as unpaved. This has happened numerous times before but until today I have been able to see gravel in time not to make the turn. Today, the road was paved for about 100 ft then turned to a one lane gravel road. It was too narrow to turn around so me and my 900 lb Goldwing spent 4.1 miles on dirt and gravel. Luckily it was in good condition. Dodging the occasional protruding rock was the only challenge.

As for the XT problem, I wanted to test a recalc while the track was showing on the map. What I did was waited until about 10 miles from home then went off-route and let the XT recalc. I forgot that I had one more shaping point so it understandably wanted to re-route me back to it. I told it to skip that point which appeared to have worked since the skip waypoint button disappeared but the XT kept trying to re-route back to the original route. I got within yards of my house (the end point) and the XT was still trying to u-turn me back the way I had come to the original route. It appeared to be insisting that I resume the route at the point where I left it instead of re-routing me to the end point. It became worthless. I would like to understand what happened and why it happened so please share your insights.

BTW, I never saw the track. The magenta route line stayed in place during all recalcs by the XT trying to bring me back to the original route.

[jfheath]

I have observed this behaviour, and have given technical support a deatailed report about it.

It seems to be that when you deviate from the route that the XT has plotted, it seems to be doing something similar to the algorithm that is used when deviating from a track-trip (a track that has been converted into a 'trip' for the XT to follow). For the track trip, there are no shaping points and via points, but the track stays in place and is shown as a magenta line and navigates just like a route. Except when you wander away from it - the XT appears to aim for a nearby point on what remains of the original route, and finds a way to get you to that point. The couple of times I have tested this, this has worked very well.

But when it deviates from a route and it recalculates - it seems to want to head to the closest point, rather than find a new way to reach the next shaping point or via point. From my observations, the closest point is behind you, initially - ie where you left the route (point A), so it plots a point from where you are now (Point B) to go back to Point A. Point B then becomes that point of the route. Ignore the instruction and continue to Point C - which is where it gives you another instruction. Point B is the closest. And so it goes in - laying a breadcrumb trail of the locations where it has told you to turn back. Nothing breaks this chain - until you make you find your own way to a point on the route and that distance is less than the distance back to the last time it gave you an instruction.

I thought at first that this only happened after the XT had recalculated the route for some other reason (eg pressing skip), but I'm not sure now - the XT seems to recalculate the original route on other occasions.

Tiphaine dealt with this issue, and indicated that it may take some time to reproduce, observe and convey her findings to the development team. I did write again subsequently when I found out that it was the same behaviour as a trip track.

I haven't heard anything since.

I recorded all of this on this forum somehwere. It might help if others reported it too. I think that it is a really serious programming error. There is nothing that can explain this away. In my case the XT wanted to take me on two sides of a triangle. I wanted to take the other side. The XT's route was faster and used a main road. Mine was more fun. It continued to navigate me back - happy to take me 29 miles - reversing my side of the triangle and then following its two side, rather than take me ahead to the shaping point which was only a mile away.

I don't think my description of what it is doing is just a guess. I turned off U turns, so when I ignored its instruction, it couldn't just take me a mile up the rouad and ask me to do another U turn. Instead, it had to take me ahead and find a side road where it could loop me around to head me in the opposite direction. Each time I ignored it, these loops can be seen on the map, still in place - so it had only recalculated the route from the last turn round loop.

I reckon that they have been focusing on improving the processing time rather that on getting the most sensible result. But thats only my opinion !

[edit]
Having said that, I was in the car today, driving with the XT. I set Home as mey destination from Manchester Airport. The weather was almost as bad as the driving, and I found myself and unable to get across to the lane where I needed to be for the route that it had plotted - lots of spray, lots of surface water and cars already in the lane that I needed to be in, going faster than me. I didn't even bother - stayed put and continued on - there are plenty of motorways around there ! The XT tried on three occasions to route me back to the junction that I had missed - coming off, going round the roundabout and head back to the junction. I assumed that it as going to continue doing that, but it didn't. After the third effort, it gave up and plotted a new route in the way that I was going. I checked the XT map to see the new the route it had chosen, but decided to go a different way. It instantly calculated a route that was very close to the one that I had decided to take.

That is how I would expect it to behave. v6.50 of the firmware. I think its time that I tested the behaviour that we have both described again).

[Peobody]

The behavior was exactly as you described in your triangle example. When it first started prompting for u-turns I assumed that it would eventually get a clue and recalc to the next point but it never did. I am glad that I was in familiar territory when it happened. I wondered if my skipping of a waypoint affected anything but it doesn't appear so based on your experience.

[rbentnail]

I have seen this serious flaw several times with the XT. Happened again to me yesterday in Virginia. The only way I could get the XT to stop directing me to u-turn AFTER I HAD REJOINED THE ROUTE was to stop the route and make a new one on the XT on the side of the road. Note that I had NO via points on my route, only shaping points, and that after I had rejoined the route I had passed over a shaping point. But the XT was happily and repeatedly telling me "Make A U-Turn" over and over and over. I tried skipping whatever point it was trying to get me to go to but that didn't work. I tried getting off the route completely, stopping the routing, restarting the route and picking "Closest Point" but it just wanted to send me backwards 6 or 7 miles again even though I was only about 200 yds from my route.

This thing makes no sense at all. I'm seriously thinking of going back to my 595.

[jfheath]

That is very odd, Russ. That suggests to me that the route may be locked in place - and I think you are right to expect that it would resume the original route if you had no Via points that would force it back.

You should have been able to reload the route, restart it and choose closest entry. Or had it modified the route itself ?

[Peobody]

Last year my old XT performed as expected on my 10 day, 10 route, +4000 mile ride to Colorado. Those routes did not contain any ad-hoc shaping points. They were all waypoints with only the stops set to alert. My ride yesterday only contained ad-hoc non-alerting shaping points except for begin and end points.

Could be problem be associated with ad-hoc shaping points?
Is this problem only associated with the 6.50 firmware?

[jfheath]

Last year my old XT performed as expected on my 10 day, 10 route, +4000 mile ride to Colorado. Those routes did not contain any ad-hoc shaping points. They were all waypoints with only the stops set to alert. My ride yesterday only contained ad-hoc non-alerting shaping points except for begin and end points.

Could be problem be associated with ad-hoc shaping points?
Is this problem only associated with the 6.50 firmware?

The only time that I ever encountered unexpected differences between ad-hoc / Waypoints / shaping points / Via Points was when testing closest entry.

In early versions of firmware (around 2.30), CEP tagetted Via Points. Nothing else.
Then later, I thought it was interesting that it seemed to target shaping points as well, but only if they had first been created as a Waypoint.
Then my observations were that it targetted any route point, closely followed ignoring all route points and finding the actual closes point.
These manfested themsleves while I was testing it with versions 2.30 to 2.90

But I wonder if there are other situations where there is a subtle difference in the way that the different route points are treated .

Like the fact that changing a shaping point to a via point changes its name and moves it , but not if it done the other way around.
Like the fact that the XT may change the name of some adhoc points, but not if they were created as a waypoint first.

I don't know about 6.50 firmware - I tested a few things after it came out and reinstalled 6.40 and it did the same - but it wasn't specifically the issue that you describe.

When I get chance, I'll take a more detailed look. (Note to self: Revisit)

I obtained pretty good results from my XT the other day - in the car. SImply navigating to a destination and then back home again. I had the Ford car satnav running at the same time - and although it displayed more pictures of junctions than the XT, often they were displayed far too early - which would have been a disaster on the motorways that I was travelling where there are a number of junctions in very close proximity. The picture on the Ford stanav was telling me to turn off, but it was as I was approaching the junction before.
The juncton that I needed to take was half a mile further on. The XT screen and instructions were much clearer and easy to glance at and get all of the relevant info in one quick look.

Verbal instructions on the XT were also sharp and to the point. All deviations (and I did a few) were dealt with exactly as I expected they would be. The XT showed a nice little diagram type image at complex junction showing the lanes, colouring the lane that I should be in, and indicating the presence of sliproads coming in. I didn't have screen shot enabled, so I couldn't get a picture of this, but it was very helpful. I'll try and reporduce it and get some screen shots.

It was nice to be able to think something positive about the XT.

[Peobody]

It was nice to be able to think something positive about the XT.

I admit that it does have moments of brilliance. In fact, early in my experience with it I really liked it. That was back when I knew very little about it. The more I have learned, the more frustrated I have become with it, and Basecamp too. We are stuck with Basecamp as is. Hopefully the XT is still a work in progress.

I am going to recreate the routes for my upcoming 14 day trip the way I used to (all waypoints, no ad-hoc shaping points). I remain bothered by the fact that all waypoints automatically become Saved waypoints but less so now that I am resigned to deleting all routes and waypoints (except Home) after each long trip. I would have preferred a 'clean-up' approach but Garmin didn't make that feasible.

[Fxwheels]

If the route lets you down for some crazy XT's reason, always have a track to follow. X-out the route and switch to track. Just try not to miss a turn.
What programmers have to work on is how to get back on planned route from any point. Supposedly the "Closest entry" should do it, but not always works as expected.
As for the gravel roads - Basecamp although have them as broken lines "= = = ", but I found out that it's not always accurate. On some occasions even a solid line can be a dirt/gravel.

[rbentnail]

I have yet to figure out the "track" thing. The one time I tried it I had a magenta route on top of a magenta track and could not figure out what line to follow where. This also happened to be one of the times the XT wanted me to u-turn to join the route at "the closest point" when I was riding on the road the route was on. I gave up, stopped the route, disabled the track and pressed Go Home. Then I promptly ignored the route given and went the way I wanted to. The XT of course quit, and told me "Cannot Calculate". That's when I started thinking about selling this POS. I have 0 confidence in it.

edit: ditto on the dirt roads. Even though I have unpaved road avoidance selected, when left to decide the route the XT will still send me down one occasionally when there are better choices. It doesn't seem to matter if I select Faster Time or Shorter Distance. IMHO it's unfortunate that one has to be selected. I think I'd be better off if I could select neither just so the thing knows to chill out a bit. When planning routes I often use Google maps to see if all the roads are paved but even still, nothing is guaranteed in NC/TN/VA mountains. It's understood that a gps can be notoriously unreliable. Hence all the signs like "YOUR GPS IS WRONG! GO BACK!"