Thoughts on "closest entry point"

[danham]

Just back from a few days of great fall riding in New York's Catskills and I may have discovered a useful strategy for joining a pre-planned route other than at its start. As we know, failing to "touch" the start can doom the rest of the route, so "closest" should be a good alternative, but sometimes isn't.

Here's what we tried, with some success. We were using pre-planned routes from a previous trip that all started in a location some 20+ miles away from where we were staying this time. Start the route using any of the entry options but ignore directions to "wrong" points and visually use the track and/or purple stripe to get onto the route by whatever means. Stop a short distance =before= the next blue dot shaping point (or flagged waypoint) and cancel the route, then re-start it and choose closest entry point. This seems to force the XT to concentrate on the logical next point, heading in the correct direction, without having to know names or order of points and simply gets you headed the right way on the route with no further drama, RUTs, recalcs, or off-route errors.

Hope this helps,

-dan

[jfheath]

Thanks for those observations. Recently I have avoided using it because I suspected, but couldn't prove, that it was one of the triggers that resulted in RUT behaviour at a later deviation. CEP seems to have a random tendency to produce odd results - I needed something that would produce the same results in the same situation every time.

Closest Entry Point is pretty good at finding where the magenta line is closest to your current position. It doesn't actually aim for one of the plotted route points - it really does head for the line. Once it finds that location it navigates to that location and ignores any plotted points in your route that you have by-passed.

But sometimes, circumstances prevent it from coming up with an ideal solution, so a quick look at the preview map is important - before setting off.

I cannot find any circumstances where CEP will always go wrong. Sometimes it will, sometimes it won't - but my chief suspects for failure are:

* If The Zumo thinks that it is facing the wrong way and U turns are disabled. Bear in mind that the Zumo does not have a compass. It only knows which way you are facing once you have started moving. It remembers your last direction, and the XT does now update its position when it gets slightly different fixes from roughly the same location. This used to have the bike logo jittering around on the screen, and changing the way it us facing. Nevertheless, a biker will often turn off the ignition (and satnav) and then reverse the bike to park it. So when it is turned on again, the bike is facing in a different direction from when it was turned off. With U turns disabled, if it thinks that it is facing south, for a route that is heading north, it has no choice but to plot a route that starts off heading south.

* If Starting from a position that is in the route's 'blind spot'. By that I mean that the zumo cannot see a point where it can intersect the plotted route. This could happen (say) if the route is heading generally north, and you are due south of the start point - it cannot see the magenta line to intersect it. I have this notion that it has to be able to draw a straight line from the current position that meets the route with a right angle. Geometrically, that would be the correct shortest distance, but whether it actually does that, I don't know. But it is not a bad rule of thumb.

You get a good idea of this if you load a track and say go. No nav instructions, but if you deviate, the XT plots a dotted line to the closest point on the track. I think that this gives an insight into what CEP is doing.

* I also suspect that it may have problems if you are already on the route - but that is another casual observation - sometimes.

I have never seen it go wrong when your current position has a good 'view' of the entire route. Eg for a routing heading roughly due north, you are starting from somewhere east of the route. Not further south than the start, not further north than the end.

It's actually a very clever bit of software, but sometimes, circumstances seem to prevent it from working as intended. I am forming the opinion that after deviating from a route, RUT behaviour is actually trying to find the closest entry point back to the route, and in certain circumstances, that has real benefits. But when the CEP is always just behind you ......

I think your solution is a way of getting round the situations where CEP can go wrong, and moving to a position where it will always work.

Maybe the above will shed some light on why you have to do this.

[rbentnail]

I have seen the behavior described as "when your current position has a good 'view' of the entire route". I often have my route pulled up but have not yet pressed START. I wait until I'm moving then press it. It just seemed to me that several problems including RUT and "Cannot Calculate Route" don't show up. I've noticed too that it really doesn't matter whether I go in the direction of the start point or not. It seems that the XT just needs to be moved some so it knows which direction it's facing.

[jfheath]

Certainly the getting the XT to know which way it is facing has been part of the issue with CEP. And I notice that some of the other issues that existed after v5.90 of the software I can no longer reproduce. It iss difficult to be certain because I never found one definitive set of circumstances when I could get it to create problems.

One thing I do know for certain is that selecting closest entry point no longer re-calculates the entire route. This is easily proved by creating a route like this: The flags are all Via Points on the main road, the A59. The blue discs are all shaping points which are detours from the main road.
I stripped the shaping points out before transferring the route. This keeps the route intact, but the sahping points no longer exist.

When the XT receives it - the zig zag route remains, and selecting closest entry point does not lose the zig zag - so the rest of the route is NOT recalculated when CEP is selected as I once thought.

I don't know when this behaviour changed.

However, selecting Skip does cause the entire route to be recalculated. A new route joining up all of the flags - directly along the A59, rather than the original zig-zag route - is caclulated.

The direction that you are facing has no impact of RUT behaviour as such - although deviating from the route can cause the XT's recalculation of an imported route that triggers it. If the XT thinks you are facing a different direction, then I suppose it is possible that it immediately recalculates the route. In which case not starting a route until you have started moving may not be a bad idea.

Strictly speaking, you have not started the route - if, as I know you do, you put the start point up the road. But the XT probably regards its route as starting where you are and getting you to the start point.

Hmm. I will ponder on that. A couple more tests are called for.

Whatever - performing the procedure to fool the route itno believing it is imported - eg by loading it and resaving it - seems to prevent any RUT behaviour at all. Certainly in all of my tests when RUT behaviour was previously (repeatedly) displayed, the identical route (suitably nobbled) immediately behaved far better that I would have expected. It seems to place more much emphasis on navigating the way that I was heading, rather than calculating a route from current position to the next route point. (eg when the way that I am heading is actually 3 miles longer than the way that it had already plotted)

[danham]

@jfheath

More food for thought - thanks. On the recent Catskills trip I did have an Imported Route that refused to allow me to save it and trick it into being a Saved Route. It either told me the name was taken (?? I had changed it) or simply would not appear in the Saved menu under any name. But that odd glitch aside, when I rode it using the method I outlined above, no RUT, no problems.

-dan

[jfheath]

That may be due to an odd character in the name, or a space. It is possible it is wanting to say 'invalid file name' but doesn't have that message to choose from. Hnag on - I'll try one.

No - its not that !

How far away were you from the start of that route when you tried it ? It has to calculate a route from your current position to the start. It might have been too much.

[danham]

Probably too far from the start. If I recall correctly, maybe 20+ miles? That was the whole reason for using CEP.

-dan

[jfheath]

No, 20 miles is hardly an issue. I was thinking maybe over to the other coast !

But I have seen errors when heading for the start in a more or less direct line with the rest of the route with CEP.

Why not just choose the start point when asked to select the next destination. It takes you there and then continues as if the start was just another via point.

[danham]

No, 20 miles is hardly an issue. I was thinking maybe over to the other coast !

But I have seen errors when heading for the start in a more or less direct line with the rest of the route with CEP.

Why not just choose the start point when asked to select the next destination. It takes you there and then continues as if the start was just another via point.

I was not clear in my explanation. We were using pre-planned routes that started in a location some 20+ miles (as the crow flies) from where we stayed this time. So the goal was to join the route where it passed closest to our headquarters and follow it to its original start/finish, but keep going until we looped back to where we were staying. I thought that choosing the start point would route us =directly= to that location, not following the route. Am I wrong?

-dan

[smfollen]

I have been trying to figure out certain aspects of the zumo XTs behavior since I mounted one on my motorcycle. I’ve been quietly following the zumo discussions here for a while. I’ve learned a lot from the smart and knowledgeable crew here. Thank you all for that. I think I finally have something to contribute.

I have generally had good success starting routes at the “Closest Entry Point” but the occasional unexpected behavior caused me to wonder what exactly is a “Entry Point”, and what exactly makes one of them the “Closest”. I’ve done some testing or experimenting as well as some emailing with Garmin support. The purpose of this post is to share what I think I have figured out. A short simple two point route example illustrates.

As many here already know, when starting a route, the zumo XT asks the user to select the “Next Destination” and offers the route’s via points as choices. This originally led me to believe that the Closest Entry Point was the closest via point. I was wrong. I next thought it must be the closest via or shaping point. Still I was wrong.

CEPselect.png (88.25 KiB) Viewed 892 times

As @jfheath has explained https://www.zumouserforums.co.uk/app.php/ZXT-P26 a route, once calculated, is made up of three types of points: via points, shaping points, and additional “hidden” or “ghost” points. The additional points are added by the software during route calcluation. (They are seen as route point extensions in a gpx file and as small dots on the map if a route is converted to a track.)

It turns out that an Entry Point can be any of those three types of points. Garmin support eventualy confirmed that for me by email.

So how is the “Closest” entry point selected? I’ll note up front that (1) “Closest” does not always mean the “most desirable” and (2) what follows is my own theory, backed up by my own testing and experience, but not confirmed by Garmin.

Recall the case of following a track where, when off track, a straight dotted line appears from current position directly to the track (magenta line). Since the device must draw a line between two points, that dotted line must be from current position to the closest track point. No roads or route calculations are involved.
When a route is converted to a track, the track points are the same set of points as the complete set of via, shaping and hidden points. It would make sense that the same logic used for the dotted line is used to determine the Closest Entry Point. In other words, my theory is that the Closest Entry Point is the via, shaping or hidden point which is the shortest straight line distance from the current position, regardless of roads or routing. (Compared to calculating straight line distances, it would take a lot more time and computing power to calculate all of the routes from current position to each via, shaping and hidden point.)

For a simple demonstration of this theory, consider a planned route along a road and a rider on a parallel road trying to join that route. It would be logical for the rider to continue down the current road in the direction toward the route’s finish and join the route at the next cross road (Old Chatham Road for the case shown here).

CEProute.png (588.46 KiB) Viewed 892 times

The Closet Entry Point is obviously somewhere on Old Bass River Road (the entire route is), but depending on how far down Rt. 134 the rider is, the closest straight line point may be nearer to Setucket Road than to Old Chatham Road, in which case, the resulting route calculation will go that [less logical] way.

CEProuteFromA.png (531.06 KiB) Viewed 892 times

If the rider is just a bit further down Rt 134, then the closet point will be nearer Old Chatham Road, and the resulting route will be more logical.

CEProuteFromB.png (404.54 KiB) Viewed 892 times

I used the simulator to capture the above screenshots safely, but it is compatible with my actual road experience.

Further support of the theory can be demonstrated by converting the route to a trip so that the entry point candidates (the track points) are visible on a map. I did this in Basecamp. The track points are a bit hard to see, so I marked the ones of interest with orange squares. If the rider happens to be located at red circle A, then the closest point in the route, by straight line, is at orange square 1. A route calculated from A to 1 would follow Setucket Road. If the rider happened to be at point B, the closest route point would be at 2 and the route calculated from B to 2 would follow Old Chatham Road.

CEP.png (235.73 KiB) Viewed 892 times

Fortunately, the system recognizes that the route from B to 2 joins the original route at the corner of Old Chatham and Old Bass River Roads, so the planned route is picked up from that point without heading to 2 and doubling back.

For this simple example, the route difference is trivial. If the two parallel roads were on opposite sides of a river, with bridges 20 miles apart, the CEP and route calculation might be more important.

Note that, at the start, the zumo did not “know” which way the motorcycle was facing. If closest entry point had been selected while, or after, the motorcycle was moving, there might be additional affect, especially if U-turns are not allowed.

I'll allow some time for those more knowledgeable than me to shoot holes in this theory before I take it any further, but I think CEP logic might be applied when a rider goes off route. If true, the above might have implications for RUT.