Lane Assist odd errors

I have a 755T and an 855, both with CN NA 2012.30. If you want to post the route coordinates, I would be glad to test.

With my 855 and the past few maps at least, I have one example as well in Flagstaff AZ, your Neighbor to the West. Travelling east on I-40 on a route that requires Exit 201, then a left at the light onto the Country Club Dr overpass, Lane Assist shows a 4-lane I-40 indicating I should be in one of the two left-most lanes. Yet I-40 is 2-lane there and I need to be in the right lane for the exit. But the ramp quickly becomes 4-lanes where I will need to be in one of the two left-most lanes.

So Lane Assist is telling me to stay left too early.

Here's one example where Junction View (using the large jcv file since the small default jcv most likely doesn't offer a JV here) is helpful. While LA is telling me to be in either of the two left lanes of four while approaching the exit, JV correctly has me in the right lane of two lanes and exiting. Once on the ramp, LA is correct (and there's no JV for that intersection).

It'll be interesting to read reports from others (alandb probably being the first) to determine if this is a nuvi 855-only issue or if it's also seen with other LA-enabled GPS units.

Great. Here are three coordinates, a common start and a good and bad end:
SB LA demo start N35.17114 W106.49331
SB LA demo bad end N35.17460 W106.50474
SB LA demo good end N35.19014 W106.49341

On my 755T, I get the same results as shown in your screen shots. So I would have to say it is an error in the map data and not firmware.

Or maybe a common error across multiple Garmin firmware variants.

Thinking through the one other location I am pretty sure of here in Albuquerque, one other possible, and the example cited by another poster, it seems a common element is the presence of a second commanded turn with LA provided an unusually short distance after the first. It will probably be a couple of hours, but I'll try to post a similar set of start, bad end, good end data for those other two--both of which are at far more major intersections than this one (one is at the main Interstate to Interstate intersection in the center of town).

For me, Lane Assist always display lanes of the next turn. So I just assumed that's how it's supposed to work. Guess not.

Here we go.

In the first case, one is heading west on Doctor Martin Luther King Jr., turning right on Oak Street, and then immediately taking a left-side ramp up onto I-25 N. In the bad case trying to do that, the LA instructs one to be in either of the two left-most lanes and to turn left, which would be really bad, as Oak Street is one-way the other way. The lane symbology, however, does match the next turn coming up--going onto the ramp, but the count down and labelling are wrong. In the good example, one approaches the same intersection, but intending to continue ahead on Oak, and the LA symbols for the very same turn are proper.

MLK LA Start N35.08408 W106.62740
MLK LA Bad End N35.09415 W106.63338
MLK LA Good end N35.08935 W106.63332

The second case is the most major intersection in the city of Albuquerque, locally called "the Big I". In this rather special case, the bad example starts a couple of miles north of the I-25/I-40 intersection, with the intent of taking a curious right-side exit ramp onto Lomas which is only a couple of tenths of a mile after the major exit ramp to I-40 W. Starting over two miles out, the LA shows the lane configuration for the Lomas exit, but the distance is counting down to the I-40 exit, and the text does not mention Lomas at all. Once the I-40 ramp is past, the same LA repeats, counting down again to the correct exit. In the "good" case I provide, the LA illuminates the same distance north as before, but this time correctly shows the structure of the intersection to which one is being counted down, and the LA matches the text.

I25 LA Start N35.13608 W106.60444
I25 LA Bad End N35.09386 W106.63482
I25 LA Good End N35.07049 W106.64004

In all four of these cases, one has two closely spaced turns, and the LA for the _second_ turn is incorrectly displayed with a distance count down to the first turn and text only mentioning the first turn. I decline to guess and the bounding parameters of cases that will fail this way, but suspect they may be common, at least the way LA was coded here in Albuquerque. I, personally have seen the issue for at least four different intersections here.

Good LA is an excellent feature, but Garmin/Navteq need to chase down this bug and figure out how to make it go away. I'm afraid this is not a simple mapreporter error reporting opportunity, and have no idea how to bring this to the attention of someone who might taken action on it.

I'm so glad that someone else finally has confirmed they had this problem too. This problem has occurred since the 2012.10 map update, and still has yet to be fixed. I've already reported this error to one of the heads of the Garmin help desk. They are now aware of the problem and have sent information to the software engineers to fix it. Here's the email message I received from the head himself about that specific issue:
"I WAS able to duplicate the issue you described both using the route you provided
and using some local routes where I know a similar set-up occurs.

I did notice that there can't be any real length of time/distance between the two maneuvers or it displays both Lane Assist options. But when they'revery close together, it only displays the second of the two maneuvers.

I did forward this on to our subject matter experts (who consult with our IT/engineering departments), but I have yet to receive any updates on the issue. I have sent them an email asking for a status update."

I never received this update. But hopefully a firmware update will be released soon that will fix this lane assist issue AND the other lane assist issue I reported about concerning the 7x5 series.

The only way to get results faster is if you guys send more reports of this to Garmin customer support about this issue. Garmin prioritizes issues based on how many people reported the same issue (I was actually told this by the Garmin tech support about my previous lane assist issue). Email techsupp@garmin.com describing the issue exactly like this guy did in the email. Also attach pictures of screen shots with your own examples. To take screenshots, go into tools, settings, display, then enable screenshots. You will then notice a little icon of a camera in the corner. Use this to capture the screenshots during a simulated route. After, plug the nuvi into the computer and find the screenshots folder on the GPS.

I've prepared a Word doc, which includes waypoint coordinates for good/bad routes through three Albuquerque example intersections, good/bad screenshots from my 855 for each case, and "ground truth" documentation of the actual lane structure from Bing Maps birdseye overhead photography screen captures.

But what would really help, I think, would be additional reports from people with other Garmin Lane Assist capable models than the 855, preferably loaded with the current 2012.3 maps, that they see this problem. It would be especially helpful if people with recently released models (and up-to-date firmware) would report here on one or more of these test cases. Garmin is more likely to care if they understand they have a generic problem affecting the whole product line than if they believe it is an isolated observation on a model sliding down the remainder sales path to oblivion.

I'll make a new compact post with waypoints and resulting screenshots, and invite those who would help both to report here, and to file their own personal reports with Garmin.

Even better would be additional example cases from non-Albuquerque locations. Look for closely spaced intersections which have Lane Assist, in cases where the proper route would give differing lane advice in both intersections.

CASE 1: San Bernardino NE westbound turning right on Tramway Boulevard NE.
SB LA start N35.17114 W106.49331
SB LA bad end N35.17460 W106.50474
SB LA good end N35.19014 W106.49341
Bad screenshot at 400 feet http://i6.photobucket.com/albums/y231/PeterAStoll/problems/G...
Good screenshot at 400 feet
http://i6.photobucket.com/albums/y231/PeterAStoll/problems/G...

CASE 2: I-25 southbound traveling past the exit ramp to I-40
I25 LA Start N35.13608 W106.60444
I25 LA Bad End N35.09386 W106.63482
I25 LA Good End N35.07049 W106.64004
Bad screenshot at 300 feet http://i6.photobucket.com/albums/y231/PeterAStoll/problems/G...
Good screenshot at 300 feet
http://i6.photobucket.com/albums/y231/PeterAStoll/problems/G...

CASE 3: Dr. Martin Luther King Jr. NE westbound turning right on Oak Ne

MLK LA Start N35.08408 W106.62740
MLK LA Bad End N35.09415 W106.63338
MLK LA Good end N35.08935 W106.63332
Bad screenshot at 400 feet http://i6.photobucket.com/albums/y231/PeterAStoll/problems/G...
Good screenshot at 400 feet
http://i6.photobucket.com/albums/y231/PeterAStoll/problems/G...

I now know this problem occurs with the 885 and the 7x5 series (I have a 765t), but it would be interesting to see if other units, like the 1xxx series and even the new 2012 series nuvis have this same problem.

We just need to have a start, and end point. Not a 'good' or 'bad' one.

By definition, one will have a preset conclusion, right?

It appears that the database representation of the lane structure at the offending intersection is proper. The problem is the "pre-display" of lane assist information for the NEXT turn, with false distance count-down information and text implying that it applies to the current intersection.

The reason for the "good" and "bad" endpoints is to show this effect--if you navigate from the provided "start" to the provided "good end" you'll just see proper Lane Assistance at the point in question. Travelling form the same start point to the "bad end", you'll see improper Lane assistance at the exact same point. At least that is what happens on my 855 with current firmware and current maps.

If you are not interested in that aspect, and just want to see error, then obvious navigating form the provided "start" to the provided "bad end" will do the job.

If we have a good or bad endpoint, the conlusion is obvious, right? If there is just a start, and an end, there is no conclusion except for the one the device arrives at.

This is how science works. There is no pre-determined conclusion. This is what we need to find out for a relevant analysis.

Observing the common element among the reported problem locations of this type as involving two closely spaced turns, both with Lane Assist provided, and recalling that Albuquerque Interstate ramps are often from frontage roads shortly after intersections with moderately major streets, I tried prospecting for problems with a selection of Start/finish endpoint pairs along I-40 and I-25 through Albuquerque. Some of my trials worked properly, some failed much as described above, and three had an extra twist to their failure. These three displayed proper LA symbols for an intersection on far approach, but on close approach switched to the next intersection LA, while continuing to count down distance and display text for the first intersection.

Here are details for one of those "delayed failures", as I observed it for starting points on San Antonio Dr NE westbound, with navigation specified to a point on I-25 north of town. For distances over two miles from the turn on to the Interstate frontage road, proper Lane Assistance appeared, as in this example captured at the half mile point just before passing through the San Pedro intersection:
http://i6.photobucket.com/albums/y231/PeterAStoll/problems/G...
But immediately on passing that intersection, the LA display changes to the wrong Lane structure and wrong selected lane initially appearing like this:
http://i6.photobucket.com/albums/y231/PeterAStoll/problems/G...
And mere feet before the actual turn to the frontage road persisting in the error and looking like this:
http://i6.photobucket.com/albums/y231/PeterAStoll/problems/G...
Immediately after the turn to the frontage road, the LA display became correct, looking like this:
http://i6.photobucket.com/albums/y231/PeterAStoll/problems/G...

Should you wish to review this case in simulation on your own GPSr, you can use these start and end points for a route:
Start N35.15985 W106.56456
End N35.23029 W106.57096

In summary of this exercise: on looking for candidate cases of two closely spaced Lane-Assisted turns associated with Albuquerque-area Interstate on-ramps, I identified ten failure cases out of twenty-five examined. One of these I had previously found by personal experience, so with this expansion I'm aware of twelve Albuquerque area locations exhibiting this problem.

I repeat my invitation (plea, really) for people owning current-production GPS receivers to try one or more of the start/end waypoint pairs posted in this thread and to report whether they see this sort of failure or proper operation. I've written a careful Word document with details, coordinates, screen shots, and even ground truth lane structure pictures, but would like to be able to beef up the section reporting that other users see the problem on other models than my 855, alandb's 755T, and rame1012's 765T.

This route does not have any LA errors on my 1490 with current 5.70 firmware and 2012.30 maps.

Here are the 4 screenshots corresponding to yours:
http://i40.tinypic.com/n48lu1.jpg
http://i44.tinypic.com/n1ypf8.jpg
http://i39.tinypic.com/15s7hpy.jpg
http://i43.tinypic.com/sgg8.jpg

Very interesting--this would seem to suggest a common firmware bug or algorithm hole in at least some 7nn/8nn models not present in yours. By the way, the third of your four picture links obtains a 1x1 pixel GIF image on my browsers, but the other three plus your word make the case well enough.

Web references suggest the 1490 was released in the third quarter of 2009, though I assume there have been firmware updates since then. The 855 was listed for first quarter 2009 release, so not really so much older. So maybe this bug is model dependent.

No errors on either of them... until the 775T hit the San Pedro Dr NE street. Then, LA shows you make a left, but Jill states correctly to make a right onto the ramp.

The 3790T gave the correct LA, and spoken directions.

I see the 3790T with an attributed release date in mid-summer 2010.

It begins to seem that Garmin knows how to do this right, but did not choose to back-propagate "the knowledge" to 7nn/8nn series receivers in firmware released to date, maybe.

When the 775T changed LA at San Pedro, it was 1 km (.6 mile) from the ramp. I wonder if this distance has a common ground elsewhere?

Of the other two Albuquerque cases of "first right, then wrong before reaching the intersection" both apppeared to change state exactly at the moment of passing a road. But the distances differed radically, one being about a tenth of a mile, and the other, I think, again being about half a mile.

It remains possible that some of my other purely bad cases might turn out to have been "first good, then bad" had I selected a start point farther away. I don't know what governs the timing of the appearance of LA symbols during approach, but have seen it over two miles away on a 40 mph speed limit road.

I found the same thing, true, here in Phoenix, too. While the directions given were correct, the lane assist arrows indicated the opposite direction as the map driving directions. Seems there is always a price to pay for the new updates.

What model GPSr did you observe this on?

...which is why those with 7X5 series nuvi or an 855 series nuvi need to email garmin tech support, or a firmware update to fix the issue might take a while to be released if its ever released all. Archae86, did you ever email report your findings to techsupp@garmin.com ?

I intend to wait to the end of the weekend, to see whether other postings here add information (especially model-dependence) that I can cite in the last paragraph of my five-page Word document with screen shots, tables, test coordinates, and ground-truth pictures.

My thanks to those who have tested and posted here already, and invitation to others to have a try.

I was using the 885T

It was the 885T.

Mt 3790 has not given me any lane assist since I got it. Could it be that I have not gone anywhere Yet?

Lane Assist is most seen on major highway exits and ramps when using a route to chosen destination. That's where I'd look for Lane Assist (and Junction View) either in real time or as a simulated route.

Good luck and report back.

My focus on Albuquerque arose from my familiarity with the area. We already have trouble reports in Arizona. To expand the known problem locations by geography a bit more, I decided to try two locations near Cleveland. Both are part of my personal life experience, I either actually erred or at least came close when navigating through them on a wedding trip a couple of years ago with my ancient Garmin GPS V. On learning of Lane Assist (and Junction View), I specifically thought these were locations where it would help.

Sadly, both involve closely-spaced Lane Assisted turns, and both foul up with the misbehavior we are discussing here--Lane Assist for the NEXT turn is displayed with count down and correct voice and text prompts for the CURRENT turn.

The first one involves travel on westbound 480, seeking to reach the DoubleTree Hotel. One needs to exit, choose the correct one of two ramps, and move on. The ramp choice is shortly after the exit, and in this case the LA symbology invites one not to take the exit at all, which would send you on down the highway a bit before recovery. Not very dangerous, but wrong if you obey it instead of the voice or map symbol.

DoubleTree Start N41.39431 W81.64932
DoubleTree End N41.41058 W81.62301

The second one involves downtown Cleveland on northbound 77, with the goal of threading the correct path through a real maze of ramps at the T intersection of Interstate 77 and 90 to get onto Eastbound Chester Avenue with the eventual goal of Epworth-Euclid church in the University Circle neighborhood.

EE Start N41.45193 W81.65693
EE End N41.50596 W81.61310

This is a really tight maze of closely-packed ramps--just the place where reliable aid from the trusty Garmin would be most welcome. From my rather distant chosen start point, my 855 successfully guides one to avoid exiting when I-70 intersects with I-480. With the big Junction View file installed there is even reinforcement there with a remarkably high resolution view.

The next LA is also good--advising one to take the center of three lanes as the I-90 intersection madness begins. This is arguably better than the voice prompt advice to "keep left". But the very next LA is wrong, advising the left of two lanes, and in conflict with the voice which advises "keep right", which is correct. The turns are very closely spaced, the possibility of confusion from the conflict considerable, and this is just a bad situation. I was tense and breathing hard with my GPS V. I'm not sure what I would have done with this 855 situation.

With all due respect, that was a while ago with old maps. With all that 'spaghetti', it's amazing GPS's can navigate it at all. Imagine it without one!

Besides, in heavy traffic, or a large city, I rarely look at the unit as my eyes are on the road, so the voice steers me right. Literally!

I have not read the entire thread, time is short tonight. The following problem happens on both Nuvi 765T and Zumo 660.

This problem exists on a few different ramps exiting off I-79 in Western PA since 2012.10 was released. I documented this to Garmin when it first appeared with a series of multiple screen shots approaching and traveling through the ramp on to the next turn. No fix yet!

Traveling north on I-79 2 miles from the exit 83 the GPS shows 2 lane arrows for the top of the ramp split, not 3 arrows with the right curved arrow highlighted. If my route at the top of the ramp is going left the left arrow is highlighted. If my route were to be taking me right the right arrow is highlighted. Also, when I reach the split, prior to reaching the road at the top of the ramp the voice and driving instructions are already giving me the next turn. All in all it just appears that Garmin has lost control over it Lane Assist database when they released the 2012 maps.

As others have said, the map route arrows and Junction Views are accurate.

Since I am unable to get a Zumo 660 that can operate with a micro SD card installed (I've had 3 in the last 6 months) why would I think they are capable of fixing the lane assist.

The maps were not old--I'm one of the fools who spent many hundreds of dollars buying map updates for the GPS V, even after they stopped delivering it in tiles small enough to allow efficient trip packing on the GPS V internal memory of about 19 Megabytes.

I believe the Garmin map was up-to-date with the streets in that maze as they lay.

With all that 'spaghetti', it's amazing GPS's can navigate it at all. Complex freeway systems like that must drive map makers insane, making sure everything is over-under correctly.

What I said wasn't a slam on you.

I've observed the behavior you describe for the I-79 Exit 83 problem running my 855 in simulation with the following waypoints:

Start N40.73138 W80.10357
End turning left N40.75944 W80.12061
End turning right N40.76283 W80.11266

I've got the big Junction View file provided by links obtained on this web site loaded on my 855. With that one, the ramp split gets a JV in addition to the exit JV! They are very pretty, but for this sort of situation I think it would be much more valuable if they got the Lane Assist working correctly consistently.

Thanks for the additional location, bear007.

FYI ...,
I have also seen 2 other exits off I-79 south of the Crafton exit but north of Washington, PA but was too busy with work issues at the time to document.

It sounds like there are isloated problems like this everywhere since the 2012 maps started. It is not specific to any GPS unit. It is a LA / map problem. Garmin has to know about it. We will probably see a fix when 2013 maps start.

These errors seem rather widespread. I would hope that Garmin would issue an update soon and before the next quarter.

Ah, that explains why Lane Assist doesn't show up for the current road when routing locally.

Well, here in Albuquerque there was very little of it indeed until map release 2012.1, when suddenly it was remarkably widespread. The little two-lane road departing my development now has Lane Assist for the place in T's into a 4-lane boulevard.

If you take your GPS on a trip, you may find yourself surprised to see Lane Assist at a different class of intersections than at home. This happened to me somewhat over a year ago, when I noticed that in the Silicon Valley area intersections which would not have had it at home usually had it.

Garmin is adding more lane assist with each map update based on how populated your metropolitan area is. For example, in this past update, I noticed just about every intersection in the New Orleans metropolitan area had lane information where it did not have it before. So a metropolitan area that is more populated will get full coverage lane assist sooner than a less populated metropolitan area.

about Lane Assist or Junction View. Kind of confuse.

Lane Assist

I'm not optimistic for a fix that fast unless they were triggered by user complaints or their own observations many months ago.

The observation of model-dependent failure may mean this is primarily a firmware problem, and so far we've seen it pretty broadly in the 7nn and 8nn lines. But even if there is a single-instruction fix, re-issuing firmware for that many models would involve a lot of QA work, and whatever else they do in a release. Many of these may be models for which they have hoped never again to release a firmware update, and for which they have not allocated further development resource.

Our observation that these seem to happen only with closely spaced intersections both of which have Lane Assist provided suggests a conceptually simple fix by selectively depopulating a "fixed" map release which has fewer Lane Assist intersections. If our model dependence hint is fully accurate, this could actually be a retrograde step for owners of models which may not have this problem at all because of differing firmware, would require developing a scheme for detecting problem pairs and choosing which to delete, QA investment, and abandonment of a fair amount of investment in populating Lane Assist locations.

I'd love to see this fixed, but unless a lot more people send in independent, specific, verifiable error observations to Garmin, I doubt it will rise high enough in their prioritization to muster the considerable resources required to fix it.

I just checked. I sent my local LA problem on I-79 in western PA on June 9th. It was sent with routing info and a screen snapshot for every LA change in the upper right corner.

How could they not say,"Oh! Look at that!" But I guess they can ignore it if they got it on their desk on a Friday late afternoon.

Three months after reporting this issue, I received this email today regarding not only this lane assist issue, but the other lane assist issue for the 705 series that I discovered:
"Thank you for providing us with information regarding your nuvi 765T. We were not able to duplicate your first Lane Assist issue on any other series. Although we do not have any plans to release any software updates in the future for the 705 series, the ticket created for this issue is still open and in review. As such, it is still possible that there may be follow-up from engineering. And as for the second lane assist issue you discovered, we are still in the process of discovering which units are having this problem, but so far, only the nuvi 705 series and 805 series units are displaying the issue. Since these units have long been discontinued, a firmware update correcting the issue will not be immediately forthcoming. Please reply with any other questions you may have."

The "second lane assist issue" I spoke of was this issue that this forum topic is about. Looks like if we ever want to see this resolved, more people are going to have to send in reports.

My LA odd error in this thread was not fixed yet. Did anyone else get their error fixed??

Yes, I ran a simulation of a route including an exit ramp that gave wrong LA directions with the previous map. At least in this one simulation of one exit ramp, the new 2012.40 gives me correct LA arrows now.

I am looking to buy a new gps with lane assist, and I just wonder if there is a difference between the different units... are all the maps the same for the lane assist?

I think the maps are the same, but I doubt that is the answer you are looking for. Most likely you want to know whether the units share this problem.

If you review earlier postings in this thread, users reported at specific locations that a 3790 and 1490 work properly at locations where the troubled units have problems.

I think the correct description of this problem is a firmware bug which only manifests itself for certain conditions in the map database.

Our tests were certainly not comprehensive, but it seems the way to bet is that Garmin may never release a firmware update for "older" models mentioned here as having the problem, but that newer models, including those sharing a firmware base with the 3790 or 1490 may lack this problem.

I appreciate your explanation for my question. Many models have different issues. Here we can learn what they are to avoid buying a faulty unit for your needs! Thanks