Weird Routing Behaviour (3) - Not the Repeated U Turn stuff.

[jfheath]

I've been doing some route observations for a couple of videos - this for a presentation that I am doing in February.

It wasn't my intention, but I captured soemthing on the video which I had observed a few times, but for which I had no evidence.

I wish that I hadn't described the earlier phenomenon as RUT behaviour - because that is just one symptom of what is going on.
One of my first thoughts when I cam across it was that in the same circumstances, routes were behaving in exactly the same way as track-trips ie tracks that have been converted to trips. I once played with these, and they are really quite useful - unless you have to go in completely the oppositie direction. In that case, the navigation gets stuck in the same RUT loop. Taking you back. But where to? Well to the place that it last asked you to turn back - which is immediately behind you.

But that too is not correct. If you tell it to not allow U turns, then both routes and track-trips will reveal that they are actually trying to find a way back to the unaltered route. To find the Closest Point. As soon as you get close enough to any point on the magenta line, it will head for that. Allowing U-turns alway puts that point just behind you. Not allowing U-turns makes it look much further ahead.

It seems that pressing skip and then deviating from the route is not necessary to trigger odd behaviour. Simpy pressing skip changes the nature of the route. It is just that you don't notice it until you deviate from the route.

So here are a couple of observed and recorded example of different things happening on the same route, after pressing Skip.


Here is the test route: (Click to see a larger image)

Route Point Video Map 1.jpg (58.38 KiB) Viewed 1144 times

I stopped just ahead of the white rouandabout - after 03 Via Pt and before 04 Shaping Point. I needed to Skip the 03 Via Point which I had not visited, and to which the XT was still trying to get me to turn round and visit. After skipping the route was still plotted ahead at the roundabout (2nd exit) to take me to 04 Shaping Point.

I set off again to the roundabout and took the third exit, rather than the 2nd - which was the way that the route was originally plotted. As soon as I took the third exit, the route recalculated to find a new way to get me to Shaping Point 04. This is shown on the screen shot taken from the video recording. (Click to see a larger image)

ZUF 01.png (802.38 KiB) Viewed 1144 times

This clearly shows the way ahead and after a mile doubling back to visit shaping point 04. Up ahead, just after the junction is shaping point 05. I ignored the demand to turn left to visit shaping point 05 and the XT calcuated the way ahead on the originally plotted route - not minding that I had missed point 04.

This is exactly how it should be. And is exactly how the 590 and 595 would have behaved.

This was a Saved route


2nd test with an imported route, exactly as transferred from Basecamp.

More or less the same thing happened, and I stopped to Skip 03 Via Point in the same place - just before the white roundabout. After skipping the route was still plotted ahead at the roundabout (2nd exit) to take me to 04 Shaping Point.
I then set off again.
This is the screen shot just after taking the third exit fromt he roundabout, having ignored the satnav which was directing me to take the second exit. (Click to see a larger image)

ZUF 02.png (775.03 KiB) Viewed 1144 times

Note that 04 Shaping Point has disappeared. The route is plotted ahead to visit 05 Shaping Point and continues to the right turn for 06 Via Point. When I passed 05 Shaping Point, the Zumo changed the shading of the magenta line to indicate that I had entered the next section. But shaping point 04 had gone, it had been removed from the route.

My interpretation of this is that having got lost, it was now heading for the closest entry point. The route just ahead of 05 Shaping Point is probably the closest - and as we know, any route points before the closest entry point are ignored.

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

Summary observations comparing the two routes.

Prompts to Skip the Missed Via Point
After 03, the saved route would make 2 U turn requests and then pop up a window asking if I wanted to skip. It did this 3 times before I pulled off. to skip Via Point 3. The imported route did not issue any prompts to skip the Via Point.

Missed Shaping Point

Missed shaping point at 02 was handled the same for both types of route.

Missed point 04 was handled 'properly' by the Saved Route (ie as the 590/595), by finding another way to visit it, and only ignoring it once the magenta route ahd been rejoined - before 05 Shaping Point.

Missed Point 04 was wiped out completely. As if I had stopped by the roadside and selected 'Closest Entry Point'.

Missing 06 Via Point

Unseen - 06 Via Point is off road. There is no way to get to it.
Saved Route repeatedly demanded U turns after I had gone past it. When I turned round and went passed it again, it continued to do this. It kept asking me to turn round and go back until I got home again.
Imported Route. Allowed me to drive straight past, and just as if it had been a shaping point allowed me to continue ahead to the finish point.

This behaviour is inconsistent with how it behaved at Via Point 3, which was also off route. (Although 06 was closer). But I had pressed skip after not visiting 03 Via Point.

Very curious. It seems to be making a sensible decision - if you are wanting to take main roads and your route point is just off the main road. But if that shaping point was deliberately placed to take you on a scenic detour, and you happened to miss the turning, or the road to take was blocked - it would not find another. Would it then take you along the main road to join the route further ahead. This suggest that it would. Whatever - it seems that it would alter the route.

[Peobody]

Something that I think is missing is a statement about whether these tests confirm that a Saved trip behaves better. As I understand the results,
- The Saved trip routed you to shaping point 03 whereas the Imported trip didn't
- The Saved trip demanded you visit Via 06 whereas the Imported trip allowed you to bypass it (I assume it continued to navigate ahead).
I consider both behaviors by the Saved trip to be appropriate and expected.
Did I miss anything that indicates that a Saved trip is not necessarily better than its Imported version?

[jfheath]

Something that I think is missing is a statement about whether these tests confirm that a Saved trip behaves better. As I understand the results,
- The Saved trip routed you to shaping point 03 whereas the Imported trip didn't
- The Saved trip demanded you visit Via 06 whereas the Imported trip allowed you to bypass it (I assume it continued to navigate ahead).
I consider both behaviors by the Saved trip to be appropriate and expected.
Did I miss anything that indicates that a Saved trip is not necessarily better than its Imported version?

No- Both routes tried to make me visit Via Point 03. The saved trip gave 3 prompts on the screen to ask if I wanted to skip the Via point.
The imported route didn't give any. In both cases I stopped in the same place to skip the Via Point before continuing to miss out shaping point 04.
The saved trip continues to behave exactly as I would expect. The imported trip has different behaviours which are 'new' to me. ie I have seen them, but didn't have enough information to know what I was seeing. I certainly didn't link them to saved/import behaviour.

Better ? I don't know. I thought about that and I haven't quite decided. I still prefer the saved trip behaviour - because I am familiar with it from the 590/5. But I absolutely hated it when I first got the 590. It was a massive jump to work out what was going on. (The 660 behaved a bit like a saved route in an XT with no Via points - it stored the entire route and you could pick up any part of it and it would navigate you from there).

The positive side of this new behaviour is it might appear as though the satnav is trying to predict the rider's intention. (Thinks like an XT) If you miss a shaping point, do I assume that you want to go back to it when I can see the plotted route just a mile ahead and I can get you back onto the route. I think the latter, 'cos otherwise you would have made it into a Via point.

And if you miss a Via point - does the same apply. Yes I think so. And I wont bother offering the opportunity to skip it if you are still on the magenta route.

So that was my thought about it - had I missed something that was really quite useful - and it doesn't sound too wrong. The answer is - yes I had missed it (or not pinned it down before with evidence), but I still think it isn't what was intended.

It ignores the fact that I might be wanting to take that quiet mountain back road, but there are road works at the junction.
And if this was the intention, then why does it behave as expected when a saved route is loaded ?

No - I still believe that when skip is pressed or the route is recalculated, the nature of the route becomes more like the track-trip route which relies more on the route than on route points. And when you deviate, it heads for the closest point on the original route. And that seems like a slip up.

But it explains to me where the Do you want to Skip ? pop-up windows had gone. I rarely see them these days, because I use mainly imported routes from Basecamp. And on occasions I have been able to go past Via points without it complaining - only to realise after - but by then its too late to find out what happened.

It is good to know that there are some other behaviours that crop up for any imported route) that converting it to a Saved trips fixes in an instant. I used the save current active route method for these tests.

I made myself a wooden block with a RAMball mount at one end for the Zumo. And a camera mount at the other - so the stanav can just sit on the passenger seat while I drive. I cant understand why the video seems to move about - becasue both are rock steady. I discovered that the camera has image stabilisation - so it was adjusting to an inbuilt spirit level. So I turned that off, which made it better - but it still does it a bit. There's some other setting in there. Oh for a cheap camera which just takes video !

It is quite revealing seeing what the satnav does and then being able to replay it.

[jfheath]

I was just looking at this behaviour again the other day and came to a very concise description of what is happening - and it fits in with what I always thought was happening in the early days of investigating what I later called RUT behaviour.

• If you press skip the route recalculates and it then becomes like a Track-Trip.
• If you deviate, it heads for the closest point on its recalculated route and it Autoskips any route points that happen to be bypassed as a result - removing them from the route completely.

Skip is the trigger. But only for imported routes.

[FrankB]

Skip is the trigger. But only for imported routes.

So it is the Repeated U Turn stuff after all?

[jfheath]

Skip is the trigger. But only for imported routes.

So it is the Repeated U Turn stuff after all?

It certainly seems to be related. I noticed a couple of years ago that track trips were behaving in the same way as route behaved after it had been recalculated after pressing Skip.

The disppearance of the route points from the route and rejoinging the magenta line is a fairly recent observation. It seems that having pressed skip once, any route point that is by passed results in that point being skipped automatically.

I'll post the videos from which those two stills (above) are captured. But I have to find them and load them to YouTube...... I may be some time.

[Peobody]

I made four attempts to reply to the original post because I could appropriately express that I could not understand the purpose of the post. My struggle is with identifying what is new from what has already been identified as being solved with the mimport fix (eg: changing routes from "Imported" to "Saved"). The only thing I see that I don't recall being discussed relative to the mimport fix is the disappearing shaping points after a skip.

What would clarify it for me is an answer to this question: Are we are better off restarting a route after a skip recalc regardless whether it is an Imported route or a Saved route, or, is a skip recalc in a Saved route not of concern beyond the potential for a changed remainder of the route?

[jfheath]

It is just me calrifying my thoughts on this. Everything I wanted to say is in the first post - but I had only just discovered this new behaviour, and my point was that the mImport fix or the resave fix prevents this from happening as well as the RUT behaviour.

For me - on the XT, I will not be using Skip again. It seems to completely change the nature of the route when it recalculates.

If you get into trouble, it is better to reload the route and choose Closest Entry Point, or better still have the track displayed at the same time as the route.

Skip works perfectly well from my tests on a route that has been altered to 'Saved'. Such routes have never displayed RUT behaviour after pressing skip and then deviating. Same routes same places. 'Nobbled' routes always behave perfectly ie calculate to the next route point - which may still mean you get asked to U turn if that way is better. You remember my "Yahoo !" moment when I went out on my test route which had consistently behaved in the same RUT way after pressing skip and subsequently going off route. The nobbled route and the resaved route both behaved perfectly (no RUT) after doing the same thing in the same places.

I have yet to prove that an imported route displays RUT behaviour if you don't press skip. From my early observations I would say that it does not - because I only observed RUT after I had pressed skip - and it took me a while to realise that was a cause. But I haven't tested that aspect since.


But this deleting shaping points and via points - this was a new observation when I started the post. I was just trying to make sense of what it was doing - and it seems to be consistent with how a track trip behaves - which doesn't have shaping or via points.

Anyway - here are the videos from which the stills are obtained. They relate to the map that I included in Post #1

First showing the behaviour for 04 Detour Shaping Point. I've put it as a link rather than as an embedded video.

• Video 1 - showing the Saved Route behaviour: Link 1.
  Skip had already been pressed at an earlier point in the route. I ignore the 2nd exit off the roundabout and take the third. You can see the missed shaping point, and the recalculated route to get to it, soon after exiting the roundabout.
• Video 2 - Showing Imported Route behaviour. Link 2
  
  The recalculation is immediate and heads for the original route a mile or so ahead, discarding the shaping point.

Then later on the test showing what happens when I ignore a Via Point - located in a field !!

• Video 3 - showing the Saved Route behaviour: Link 3
  No attempt to re-route - the route behind is discarded, and a new route ahead is plotted
  It keeps trying to get me to go back and will not give up, even after turning round ...
• Video 4 - Showing Imported Route behaviour: Link 4
  Abandons the route and the Via Point completely

[FrankB]

I have yet to proove that an imported route displays RUT behaviour if you don't press skip. From my early observaitons I would say that it does not - becasue I only observed RUT after I had pressed skip - and it took me a while to realise that was a cause. But I haven't tested that aspect since.

Fot the route I tested, about a year ago, to go into Rut, I did not have to skip

[jfheath]

Do you happen to know if the route had recalculated for any other reason, Frank ?