XT2 vs XT - RUT Testing

[jfheath]

@smfollen has already done some testing on this and has come up with some results which suggest that the RUT issue has not disappeared with the XT2. Link here

I want to be able to confirm that my previous investigations have not yet been addressed, so for that I need to do some identical testing.
My plan is to have the XT2 and the XT running side by side on the same test route.

It'll take a couple of weeks to do the tests - when I have the time and weather to get out on the bike. This is really a place holder.

First the route. My original route has been hijacked as the XT2 does not calculate the same route as the XT. My replacement route (local) was also hijacked because it involves a road that has been closed since January, and only reopened 3 days ago.

The proposed Route

RUT Test Map.png (126.37 KiB) Viewed 2024 times

The purple line indicates the route that I will be taking according to BC this is 19 miles 31 minutes.
The blue line is the route that both the XT and XT2 calculate. According to BC this is 21.6 miles and 32 minutes.

There is a starts and and end point and a couple of intermediate points. The first intermediate point is less than half a mile after the start. I intend to skip this point as that forced the XT to recalculate the entire route. Any deviation after that and the route behaved completely differently and would initiate repeated U turn requests until if got less than a mile from the original route.

The routes separate and rejoin without encountering any additional route points - in other words they are free to choose whichever route is best.

The area in between to two routes has no roads. So it has to choose one route or the other - it cannot find a different solution. This mirrors my original tests when I first diagnosed the issue on the XT.

The 4 Via points have all been created as Waypoints.


The Tests - XT routes on the left, XT2 Routes on the right.

Screen Tests.png (81.62 KiB) Viewed 2023 times

Same route installed on each device, but the route has been prepared in different ways.

• 01 - A route exported direct from Basecamp as is. Typically this route would show RUT behaviour under the circumstances described.
• 02 - The same route duplicated, but given a different name. This has had its mImpoted byte changed to false - so the XT thinks it has been saved, rather than imported.
• 03 - Test 01 route has been loaded and resaved on the XT, and simply copied on the XT2. In both cases it has been given a new name. The suggestion is that the Zumo thinks that these routes are not imported, and will treat them differently.
• 04 - All of the via points were created as Waypoints. This makes it possible to create a route using the Zumo screen using the identical route points. So this will be test 04.

The tipping point is about 3 miles after the roundabout where the two routes split.

And that's it so far. I'll know more when I get out on the bike.

[Sean OZ]

Looks like a well thought out testing route.
Will be very interesting to see how the XT2 behaves compared to the original XT. Look forward to seeing what hapens.
Thanks again jfheath>

[jfheath]

A comparison between XT and XT2 when trying to provoke RUT behaviour.

I can provoke RUT behaviour in the XT by preparing a route which has just two alternative ways to get from one point to the next. The Zumo calculates one route, I take the other. But first I have to start the route, pass through the start and then Skip a route point to force the Zumo to recalculate the entire route. The sort of situation that I have come across many times when road works, traffic, accidents etc require me to change my plans. So for this test, I put an additional route point half a mile after the start - just so that I can skip it.

Click the images to see a full size version.

Image 1. Just started route. Note that bottom centre there is a green flag and two orange flags. The green flag is my location.
The first orange flag is the start of the route.
The second orange flag is a via point that I intend to Skip.
The only other plotted points on the route are the last orange flag via point and the finish flag.

XT2 vs XT Routing 01.png (719.14 KiB) Viewed 1841 times

Conclusion - Both routes are the same at this point.
---------------------------------------------------------------------



Image 2. 2 miles. I passed through the start point, and I stopped briefly to press the Skip button. This is about half a mile later - I stopped in a lyby after the skipped point to see what had happened to the routes on both devices.

Note that in both cases, the orange via point flags have gone. There is no magenta route behind me. The green flag marks my current position.
Up ahead the route is still heading East (ish) along the A59 before truning north. The orange Via point and chequered flag are both in place.

XT2 vs XT Routing 02.png (730.3 KiB) Viewed 1840 times

Conclusion - Both routes appear to be the the same at this point.
---------------------------------------------------------------------


Image 3. 9:49am. Approaching A59 Roundabout


This is about 3 miles later, approaching the A59 roundabout. Curious that with identical routes and identical maps, on sat nav thinks that the next Via Point is 2 miles further away than the other. Both had recalculated the route after I had pressed Skip. Perhaps one has calculated a slightly different route from the other. If so, I didn't see it - but that remains a question mark - I will take a look at the distances on the simulator.

One of the screen shots is taken slightly later than the other - but that is of no consequence. The significance of the roundabout is that I am not going to follow the instructiosn. I am taking the 2nd exit - the yellow road heading towards Bolton Abbey.

XT2 vs XT Routing 03.png (515.97 KiB) Viewed 1834 times

Conclusion: There is a slight difference in time to be investigated, but I expect both Zumos to compain when I do not take the third exit.
---------------------------------------------------------------------


Image 4. 9:53am Turn Back Loops and U turns.

These are to be expected. The point where it is faster to continue ahead - identified by repeatedly moving the bike's current position and asking it to navigate to the end point - is a few miles further ahead. The XT2 seems to prefer turn-round loops initially, but it later demanded U turns.

The XT - well - - I messed up. I had reset the XT and forgot to check the U turn status - so that too found loop back routes or side roads. Later, I switched the XT to allow U turns. More later. Never mind - this is about the XT2, and what that does when it reaches the tipping point.
Note that the distances to the last Via point are both increasing - because the Zumos are expecting me to go back to the roundabout.

XT2 vs XT Routing 04.png (458.16 KiB) Viewed 1833 times

Conclusion - U turns / loop backs are expected behaviour on this section of the road. Both Zumos are beahving normally.
---------------------------------------------------------------------
Image 5. Approaching the Tipping Point - 5 minutes later - where it is faster to go ahead, than it is to go back.


Half a mile before the junction, the XT2 has already identified that it can take the road I am traveling and get there faster than if it turns me back.
Look at the arrival times shown in the images for the last screens.
At the same point, the XT is still trying to turn me back by using the road on the left.

XT2 vs XT Routing 05.png (495.5 KiB) Viewed 1831 times

Image 6. The Remainder of the XT2 route on the preview map.


Both images are the same location. I zoomed in and took the right hand image because I noticed something different: The position of the green flag. ie the start of the current section of the route - after it had recalculated. Note also that the orange Via point and the end point are still in position. This is exactly what I believe should be happening - at each point, before deamnding a U turn to pick up its last calcaluated section of route to go back - it should calculate the distance or time to the next route point (via or shaping). If it is less than its previous calculation then yes - let's go this way please.

XT2 vs XT Routing 06.png (806.08 KiB) Viewed 1829 times

Image 7. XT Overview Map at the same location.

This is the XT screen, 2.5 miles after the point where the XT2 decided to route ahead.

XT2 vs XT Routing 07.png (312.98 KiB) Viewed 1824 times

---------------------------------------------------------------------

Conclusion: Positive outcome for XT2 - But not yet Proved.


Why? Well - I messed up. I wanted to have U turns disabled on the XT - because the map on the XT reveals all of the turn-around loops that it has created, and when the XT is displaying RUT behaviour - it keeps these turn-around loops as part of its route back.

But early on in the test, I wanted to see if the behaviour was the same as the XT2. So I turned U-turns back on again. This may have casued a recalculation that resulted in that odd milage difference. But later, I needed the turn-around loops to PROVE that the XT was displaying RUT behaviour. Because the roads were so narrow, I couldn't stop the bike (although it kept asking me to perform a U turn !!).

So when I was stopped to get screenshots used in Image 6 and 7, I also took the ones below - just after I had disallowed U turns.

Now thes images are very interesting:

XT2 vs XT Routing 08.png (490.88 KiB) Viewed 1824 times

The left hand image shows that immediately after changing U turns back to disallowed, the route changed. I don't believe that the route recalculates normally, But now, the route was navigating ahead, not back. Just prior to stopping to take this shot, I have a commentary that says it was asking me to perform a U turn at the junction.

So was turning off u-turns the trigger for it doing this ? Is that something that could be used if anyone gets trapped in a RUT ?

I have never seen this before - although disallowing U turns does force the satnav to look further ahead than normal. Maybe as soon as I reached the junction and turned right, the XT would have given in anyway. In which case this whole test is invalid - because in previous RUT tests, the XT hasn't given up until it was less than a mile from rejoining the original route. Starting to navigate ahead just 2.5 miles from the tipping point is not so bad. It certainly isn't what I described as RUT behaviour.

But now look at the right hand image. What has happened to the Via Point flag just before the end flag ? It is no longer there.

Now this, I have seen before. In fact I have a video of it happening. It seems that if you press skip, the entire route is recalculated. If the route is imported, the nature of the route then changes. Subsequently, if you miss an instruction to take you to the next route point, that route point is removed from the route. Having said that - it didn't notice it happening when I was testing RUT behaviour in other locations.

Whatever - none of this happened on the XT2, and on this test the XT2 behaved immaculately.

I'll listen to my in-ride commentary - see if I have missed anything.

Next time I go out, I will run this route again on the XT and get more screen evidence.

[Sean OZ]

Thanks, jfheath.
Fantastic to see that your "RUT" testing of both the XT and the XT2, shows that the XT2 doesn't experience RUT problems,
(If I have read your excellent post correctly)

[jfheath]

Thanks, jfheath.
Fantastic to see that your "RUT" testing of both the XT and the XT2, shows that the XT2 doesn't experience RUT problems,
if I have read your post correctly.

It suggests that at the moment. Not yet proved. To be certain, I have to prove that the XT does display RUT behaviour when the XT2 didn't. I have on test where it seemed to display RUT behaviour, but then calculated ahead ok - but much later than the XT route did after the route had been 'fixed'.

One thing is for sure - the XT2 behaves much better with an imported route than the XT does. [edit - in a later test, see a few posts down, it seems to prove that it will get stuck in a rut]

I went out again yesterday and hoped to prove it but decided I needed a route modification as it was likely to see my finish point too soon with u turns disallowed, so i enabled them again. doh!

I did confirm one thing though. If you are in a RUT situation on the XT, then stopping and altering U turns to allow/disallow seems to force the XT to calculate a new route from your current position, rather than looking for the closest point.

[smfollen]

If you are in a RUT situation on the XT, then stopping and altering U turns to allow/disallow seems to force the XT to calculate a new route from your current position

Just a thought: If stopping to mess with the zumo, why not just stop navigation then restart the route, selecting Closest Entry Point? I've had good success with CEP.

[Peobody]

I've had good success with CEP.

Me too, but then I have had undesirable experiences as well. The most common undesirable is when CEP is behind me. IIRC, RUT is is triggered by the combination of skipping a point followed by going off-route and then allowing a recalc. RUT will start then. CEP won't help because that entry point will likely be where you left the route. If indeed toggling u-turns works in getting the XT to navigate forward, well, that would be a useful option.

[jfheath]

I did a test today. I got the XT2 to follow my verified test ride and behave as per the previous tests. I dont know what I did, but it plotted the basecamp route and then when i skipped an early point, it plotted the route that I was not going to follow. Perfect.

I went for a ride with both XT and XT2 following the same route. The XT2 behaved in the same way as the XT - ie it got stuck in a Rut, repeatedly asking to go go back. The XT2 was slightly better it started navigating ahead when it was 1.3 miles from the next Via point. Just before that it was navigating 29.8 miles in the opposite direction. The XT produced the right result at 1.1 miles.

This was proven RUT behaviour. It also happened at a previous deviation in the route - for which I have no proven test results.

Yet previously on other routes, the XT2 seemed to be behaving better. The only other thing that was different was that on todays test, I forgot that I had been playing with the Tread App so that it could synchronise. Previous tests I hadn't yet touched the Tread App.

I need to check that out, but at present, I am disappointed.

-------

The CEP suggestion is a good one and it works. Except - I rarely mention it because of the conditions that come with it. When you need it, you are off route. Many routes are circular - out for the day. If you select CEP and you have wandered towards the last half of the route, then CEP may find the return leg. Similar conditions can arise on a route that zig zags. You just have to be careful when you use it.

[smfollen]

@Peobody @jfheath

The most common undesirable is when CEP is behind...

The CEP suggestion is a good one and it works. Except - I rarely mention it because of the conditions that come with it. When you need it, you are off route. Many routes are circular - out for the day. If you select CEP and you have wandered towards the last half of the route, then CEP may find the return leg. Similar conditions can arise on a route that zig zags. You just have to be careful when you use it.

Thank you both. You make good points and I agree entirely.
I find CEP to be helpful, but it can't be used blindly. After selecting it, I take a look at the map before I start off to be sure the route makes sense.
Also, if I've planned a route with a close loop, or one that overlaps, I tend to break the route in two - out and back. I think I learned that from an old jfheath post - thank you!
And yes, a zig-zag route could certainly challenge CEP.

[Oop North John]

For the first time today I had a full blown RUT. Was heading home and chose the option that Basecamp had calculated and was displayed as a "track" and was local so knew it was sensible. Went off this unsaved / un-copied route that came from Basecamp and for the next few miles there was the request to do a u-turn or effectively one using a road / roundabout that achieved the aim. Was still doing it when I was less that half a mile from home and it wanted me to do 20 miles / 40 minutes of travel . The XT2 did stop the RUT'ing when I tuned onto my home street with about 400 yards to go

The use of the track did help, but one problem is if you're doing an almost out and return, then working out which is the out, and which is the return bit can exercise the brain cells.
Will start to "copy" the routes in the XT2 from now, had hoped that after a couple of u-turn attempts it would behave, obviously this was comprehensively disproved today