This site uses cookies to offer you a better browsing experience. Find out more about how to manage cookies, or dismiss this message and continue to use cookies.

Why does the Nuvi keep trying to take me on back roads?

1246716

Comments

  • cuc tu 91 Points
    What in Garmin's algorithm could possibly cause the device to pick a longer/slower route when it knows (or should know) that a shorter/faster route exists.

    ahhhh... but that is just it. :) I'll have an article on this sometime. But the GPS does not test all combinations of route possibilities. (Any GPS.) It can't. There are too many possible combinations.
    That is hard for me to buy. The processing power of my GPS is on par with my fastest PC of 10 years ago.

    I can't believe that a simple two dimentional analysis could be so complex. This is where I tought my ignorant, "if I was a programmer" speal about it is so simple to take every route, multiply speed x distance, and get a number, then pick the lowest, and there you go...well, I can see the complications in that "every route" part, but I can also see ways to significantly limit those complications.

    Anyway, I still see that a clear definition of the problem is required.
  • Tim 1480 Points
    well, I can see the complications in that "every route" part, but I can also see ways to significantly limit those complications.
    Sure-- and that is why what I described (forcing a route that even the device thinks is faster than what it initially picked) is rare.

    Consumers demand fast routing times and fast re-routing times. One of the easiest ways for developers to achieve this is to reduce the amount of fringe cases that are checked that rarely find a faster route.

    There is a good white paper from NAVTEQ that gives a primer of route calculation to developers. Unfortunately they've removed it from their website. I have a copy, but I'll have to check and see if there are any copyright notices before I post it. After reading through you can see how certain types of good routes might not be found. Rare... but it can happen.

    A good example of this in action is some of the newer TomTom devices. (There are a few other smaller companies I've confirmed are doing this too. Basically, the GPS quickly calculates an initial route and you start driving. Then, later on in the drive you might see the GPS say "faster route found, do you want to take it?". And this happens without any sort of traffic receiver connected so it isn't as if there is any new data it didn't have before.

    image

    Where you commonly see this happen, and where you occasionally see the oddity I mentioned above, is when you have a longer route that starts and ends on a slower road, has a faster road in the middle, yet while on the faster road there is a shortcut from faster road to slower road and back to a faster road again. Once a route "steps up" a road class, tend to not consider alternatives on slower class roads until you are closer to your destination. And part of the other way routing works is that typically a device will sort of start to build two routes... one from the starting point and one from the ending point... each towards each other seeking higher class roads.

    I'm not saying any of this is the reason why the Garmin picks a bad route in the examples we are looking at-- but it was related. :)
  • mvl 191 Points
    Tim, while a bit off topic here - I don't think your Tomtom example is caused by what you think it is:

    Because IQroutes has time-of-day specific routing, it runs a background route recalc every time the anticipated arrival time moves 5 minutes. The idea here is that if you stop for gas, the fastest route may now be different (eg new rush hour activity may have developed). If you have every voice preference turned on, you can hear the recalc even if it makes no change. So without a traffic receiver, when stuck in an accident or IQR-10mph-bug traffic, you'll hear "arrival time is now n:nn, you are still on the fastest route".

    In fact, the above thing made Tomtom some money (at least from me). I was in a 1-hour construction delay, and as I'm slowly crawling my 720 nagged me with "you are still on the fastest route" every 5 minutes. After hearing it for the 10th time, I got so angry that I upgraded my data plan while in the jam, and signed up for PLUS traffic immediately when I got home.

    But the point of this was, I think your Tomtom example is one where the best time-of-day path changed mid-route because of an unpredicted backup, and the Tomtom was asking for permission to switch routes purely because of the new time. I still think Tomtom runs the same route algorithm each time it routes.

    Actually, I think Tomtom does a "try every route" algorithm for trips under 3-5 hours. Most optimization algorithms don't work if the road speeds are different based on the time you drive them. Above 5 hours, it appears that it shuts IQroutes speed off, and runs optimization logic from the 5-hour point to your destination. I've never seen cases where an on-route via point changed a Tomtom route, except if the arrival times were within 5 seconds (the minimum Tomtom time rounding), or if the total trip was over 5 hours.
  • Tim 1480 Points
    I disagree. I thought of that and have actually run calculations on the original route picked as well as the new route it suggested, both for the original planning time and the then current time. (As well as running the test simply with IQ Routes turned off.) None of those cases accounted for the 8 minute faster route in the screenshot.

    I have seen routes of about one hour in length where you can place a via point and the TomTom will show a faster route than what it originally picked as the fastest. I'll try to see if I took any notes on the route the various times I've seen it.
  • All this talk of testing routes and fancy algorithm issues is interesting but please re read my post on what happened to me on a 3500 mile trip: a consistent series of not just poor routing but unreasonable routing. I'll repeat the simplest example. Going from Boca Raton to Atlanta you get on the Florida Turnpike which merges with I 75 to GA.

    My 265WT with the latest map wanted me to get off the Turnpike and take Rt 4 to Orlando and then 408 to get to 75. This would mean going North East and then West as opposed to just staying on the Turnpike. Since I had a map I decided to pass the point where it wanted me to turn and kept on the Turnpike, my time to destination dropped by 30 minutes. Again 75 and the Turnpike simply merge!

    By the way, Google maps correctly plots this run.

    This is not rocket science and it doesn't take a software engineer to figure out there is a problem. I can't believe that a company would not publicly address this issue if only to say they are working on a solution and that they will take care of their customers who put their trust in them.
  • Tim 1480 Points
    Is this what it did?
  • Is this what it did?
    Unless, that section of the FL Pike was closed, that routing is absurd!
  • Might as well put my two cents in.

    The number one complaint I see on Amazon is the routing, and it is generally across the board on just about every Garmin Nuvi model. Many of these complaints I take as a grain of salt, but many are quite mystifying.

    I've owned two Garmin units (260 and now the 765), and have driven often with my girlfriend's c330. I've found in most cases they route the same. However, I have noticed that my 765 has consistently routed me differently in a few areas (going to my girlfriend's house it uses the freeways ... going home it tries to get me to use streets for half of the trip, then puts me on the freeway). I had assumed the reason was because of the traffic function and just ignored it. Does it bother me? Not really.

    I also was going to a hospital to see a relative and my 765 verbally wanted me to make a right turn near the hospital, while the picture on the screen showed me going straight. I made the turn and it circled around the hospital and had me turn left into the parking lot. Had I followed the magenta(?) line, I would have entered on the right just a block or two away and not circled completely around. I found this rather odd, and consider this behavior a problem (left hand not knowing what the right hand is doing), but again, it didn't really bother me. After all, it did get me into the parking lot.

    Do I trust these devices? Sure, to a point. It might not always be the fastest, shortest, or most economical, but it will get me there. And if I know where I'm going, I don't often use the GPS. And when I'm going to an unknown area, it has always gotten me there.

    Should Garmin re-engineer their routing engine? You bet. But there are a lot of things Garmin should be doing and aren't. One should be able to see the different routes on-screen (fastest, shortest, most economical), then let the user decide and select which one is more appropriate. They should also add to their avoidance menus (always use highways / never use highways). That right there "could" solve a lot of routing issues. Will they? I'm not holding my breath.
  • Tim 1480 Points
    Well said dsrussell.
  • Marc 301 Points
    edited August 2009
    Is this what it did?


    Unless, that section of the FL Pike was closed, that routing is absurd!
    So I just put this in to my wife's 265T as well as my Magellan 1340. Neither chose this strange route, but did the realtively simple straight line approach of Floridas Turnpike to I75. The 265 has the 2010.1 maps in it. Perhaps there was an accident on the Turnpike and you got routed around it. Why don't you do a simulated route now and see if you still get the same route or the more logical one.
  • Tim 1480 Points
    Perhaps there was an accident on the Turnpike and you got routed around it.
    Precisely my thoughts as well. The route is bad, but we need to rule out the traffic receiver... regardless if the traffic receiver was correct or not.
  • SergZak 340 Points
    edited August 2009
    I tested the Florida route that Tim generated from Google.

    Tested: nuvi 765T FW3.60, v2010.20 maps, no avoidances enabled
    Result: The generated route stuck with using Florida's Turnpike throughout
    Note: Enabling ALL avoidances on the unit resulted in the exact same route

    Tested: nuvi 760 FW4.80, v2010.20 maps, no avoidances enabled
    Result: The generated route stuck with using Florida's Turnpike throughout
    Note: Enabling ALL avoidances on the unit resulted in the exact same route

    Tested: Mapsource v6.13.7, no avoidances enabled
    Maps tested: CN NA NT v8, CN NA NT 2009, CN NA NT 2010.20

    Result: Despite the maps used, the generated route stuck with using Florida's Turnpike throughout. Nothing I did made it deviate from this route.
  • Tim 1480 Points
    I shortened up the route in that map for illustrative purposes. While I doubt it would make a difference on this route, did anyone try the original route from Boca Raton to Atlanta and then zoom in on the area in question?
  • Marc 301 Points
    Not sure what you mean Tim. If you mean did I program in the full route from Boca to Atlanta, the answer is yes.
  • Tim 1480 Points
    Not sure what you mean Tim. If you mean did I program in the full route from Boca to Atlanta, the answer is yes.
    Yep, that's it, good. Just wanted to make sure you weren't using the same (very short) route I built in the Google map to highlight the area.
  • The OP didn't mention anything about using a detour around a traffic problem on the Pike, so I doubt that traffic reporting would have influenced the routing. If he input the trip before hitting the road traffic reporting should have had zero influence on calculating the route?
  • Marc 301 Points
    Have to put in my two cents here, and say I agree with dsrussell. I see this as a tempest in a teapot. If people are expecting consistently near perfect routing from their GPS device I suggest they return them or sell them on ebay. It is a tool and a highly imperfect one at that. For any trip longer than about 30 miles if you don't look at a map before you walk out the door, there is a fair chance you will not get an optimal route from your GPS. For routes shorter than that, you still may not get an optimal route, but it probably doesn't make that much difference.

    I gave one of the easiest to use GPSRs to my wife because she is at the technical level where a folder in Windows is still somewhat of a mystery. But even she realizes that for a trip down to her friends house in Conneticut, it makes sense to look at Google maps and sometimes ignore the route that her 265 is telling her to take. For finding addresses around New England where street names change about every half mile and most streets aren't labeled, it is great. As a trip planner for a tour across the USA if you solely depend on your GPS you are being very foolish.

    Sure send your complaints into the manufacturer, but don't expect instantaneous fixes on problems that are more complex than you realize. Your mind is a wonderful tool, use it to help plan your routes.
  • I'd never give up my Garmin because without it I wouldn't know how big my carbon footprint is!!! :lol:
  • Marc 301 Points
    The OP didn't mention anything about using a detour around a traffic problem on the Pike, so I doubt that traffic reporting would have influenced the routing. If he input the trip before hitting the road traffic reporting should have had zero influence on calculating the route?
    I don't think so. If you have avoidances checked and traffic checked in the submenu, if the traffic problem exists at the time of initial routing, it will be included, just as avoid u-turns or any of the other choices for avoidance. The easy way to find out if something like this is going on is to do the same route in simulator mode and see what happens.
  • Yes!! And that is just one example.
  • Tim,

    I thought my reply would follow yours but it shows up at the end. I responded to your question asking me if the route between Boca and Atlanta you linked to was the one my unit gave me. The answer is yes.

    And there were no accidents or notifications either by the Garmin or on the radio.

    Marc,

    I agree with what you said that we shouldn't expect perfection. But if you read the summary of my trip and that of others you will see that at least some units are consistently sending us on bizarre routes that defy logic and are not minor route discrepancies. Unfortunately I didn't discover this issue until it was too late to return it or else I would have. For the record I have had 3 other GPSs including 2 Garmins. While they occasionally missed a beat, I've never experienced such consistently poor routing.
  • Tim 1480 Points
    And there were no accidents or notifications either by the Garmin
    If you create that same route right now and zoom in on the area, does it still show the bad route? Which map version do you have?
  • Tim,
    To answer your question: If you create that same route right now and zoom in on the area, does it still show the bad route? Which map version do you have?


    The simulator has me take the turnpike to SR-50 to Ocee (at 116 miles from my start) then exit 156 to I 475 N to Atlanta.

    I have the 2010.10 map.
  • cuc tu 91 Points
    I just tried my 765T and it did not use that obscure route.

    3.6 FW and 2010.10 map
  • mvl 191 Points
    Just curious and to reiterate my earlier "learning engine" post.

    Has anyone had problems after using the new map for a month or two (with the Nuvi on most of the time)? This would give time for the tuning engine to learn proper speed differences.

    It seems like everyone who's reporting issues is a new device or new map owner.
  • BullieverBulliever 0 Points
    edited August 2009

    The simulator has me take the turnpike to SR-50 to Ocee (at 116 miles from my start) then exit 156 to I 475 N to Atlanta.

    I have the 2010.10 map.
    That's the route my 755T w/2009.11 maps give me,. On simulation, It tells me (Voice) to drive 320 miles on SR-50 which is not only insane, its impossible! Garmin has lost it's marbles.
  • Marc 301 Points
    edited August 2009
    Tim,
    To answer your question:
    If you create that same route right now and zoom in on the area, does it still show the bad route? Which map version do you have?


    The simulator has me take the turnpike to SR-50 to Ocee (at 116 miles from my start) then exit 156 to I 475 N to Atlanta.

    I have the 2010.10 map.
    So this is interesting. Apparently we have two 265Ts with the same map, yet when run under simulation mode give different routes, since mine does not do this. Could you see what you have checked in the avoidance menu? Also, let's compare firmware.
    Mine is Software 5.30
    Gps SW 3.10B
    My avoidances are U-Turns and Traffic

    Finally, my route perefernce is fastest time.
    Did I miss anything?
  • BullieverBulliever 0 Points
    edited August 2009

    So this is interesting. Apparently we have two 265Ts with the same map, yet when run under simulation mode give different routes, since mine does not do this. Could you see if you what you have checked in the avoidance menu?
    I believe it maps it correctly (FL Tpke to I-75N to I-475 to I-75N), however the voice directions and written directions do not match the mapped route and are totally insane...My 755 directs me (voice) to drive 320 mile on a road that ends in approximately 60 miles....320 miles would put you about 250 out in the Gulf of Mexico!!!

    755T
    SW: v 3.60
    Map: NT 2009.11
    Avoid: Unpaved Roads
    Routing: Faster Time
  • Marc 301 Points
    This is a very different problem then. The routing engine would be ok. It is the interface between the route and voice directons that are screwed up. I never actually run the route, so mine could do the same thing. Have you let it run the simulation and see if it actually sends you out into the ocean? That would put some egg on Garmin's face!.
  • This is a very different problem then. The routing engine would be ok. It is the interface between the route and voice directons that are screwed up. I never actually run the route, so mine could do the same thing. Have you let it run the simulation and see if it actually sends you out into the ocean? That would put some egg on Garmin's face!.
    No..the mapped route is correct...the TTS directions are "out to sea"(pun intended!)...Garmin has so much egg on it's face already that they wouldn't notice anymore being added..

    If you want to hear it for yourself, simulate the route using Kissimmee, FL as your start point and Atlanta as the end point...it will shorten the time it takes to get to the SR-50 error.
  • Marc 301 Points
    Not sure I completely understand. I have seen on a Magellan where you are supposed to turn right and the voice says turn left. However, if you follow the voice it immediately knows you are off route and reroutes you to correct the error. Is this what is happening, or once you go off route does it just let you continue?
  • Not sure I completely understand. I have seen on a Magellan where you are supposed to turn right and the voice says turn left. However, if you follow the voice it immediately knows you are off route and reroutes you to correct the error. Is this what is happening, or once you go off route does it just let you continue?
    I'm running it again I have 6 more miles before I reach the CrazY ZonE!
    I'll watch and listen more carefully this time...now .4.4 miles to go!


    With simulation you cannot go off the route. Here's what happened.

    @ 1 mile it tells you to keep left on Florida's Turnpike

    then it says: continue 1.6 miles then keep left on on SR-50 ( it should say keep left towards SR-50 because you do not want to exit the Tpke).

    I have seen this type of error on Garmins before - which can be confusing for people unfamiliar with the area and dangerous at turnpike speeds


    It also illustrates it with Junction View. (Not one of the best JVs I've seen)

    Then it says: Keep left on SR-50 ( it should say keep left on FL Tpke)

    then it says: Drive 342 miles on SR-50

    The map keeps you on the FL Tpke never exiting onto SR-50 which at this point runs parallel to the Tpke.
  • Marc,

    Sorry I don't know how to use quotes. To answer your questions:

    Preference is for fastest time
    Software 5.00
    GPS SW 2.55b
    Avoidances disabled (all unchecked)

    Also since I noted that I don't have the latest updates, I just did that and ran the route again. Got the same exit on route 50 direction, though reading above, that could be a grammatical error rather than a routing error.

    I am going to do a hard reset later today after I get the battery recharged and see if that helps.
  • Marc 301 Points
    I am not sure that Bulliever and mitrajoon have the same problem. One sounds like a routing problem, the other sounds like the GPS is giving voice and visual commands that don't jive with the route. It would be interesting mitrajoon if you update your firmware and see if your problems disappear. The frimware and the slight difference in avoidances seem to be the only difference in our settings.
  • BullieverBulliever 0 Points
    edited August 2009
    I am not sure that Bulliever and mitrajoon have the same problem. One sounds like a routing problem, the other sounds like the GPS is giving voice and visual commands that don't jive with the route. It would be interesting mitrajoon if you update your firmware and see if your problems disappear. The frimware and the slight difference in avoidances seem to be the only difference in our settings.
    Not from what he has posted...I believe we both have the same issues.
    The simulator has me take the turnpike to SR-50 to Ocee (at 116 miles from my start) then exit 156 to I 475 N to Atlanta.
    That matches the route description that I am getting. SR-50 doesn't have numbered exits it's a state highway/city street, not a limited access road. "Exit 156" is on I-75 near Macon, Georgia, not on SR-50. 50 runs East - West across the middle of FL from Titusville on the east coast to Weeki Wachee on the west coast.

    Why don't you just put the route in your Nuvi, simulate it and see/hear it for yourself. To shorten it use Kissimmee, FL to Atlanta, GA. Without doing so you're just beating off here.
  • Marc,

    Sorry I don't know how to use quotes. To answer your questions:

    Preference is for fastest time
    Software 5.00
    GPS SW 2.55b
    Avoidances disabled (all unchecked)

    Also since I noted that I don't have the latest updates, I just did that and ran the route again. Got the same exit on route 50 direction, though reading above, that could be a grammatical error rather than a routing error.

    I am going to do a hard reset later today after I get the battery recharged and see if that helps.
    I'm running the latest FW on my 755 (with 2009.11 maps) and we are getting the same result, so I doubt a reset is going to do anything to remedy the error. Garbage in Garbage out! The boys girls in Olathe, KS really need to get out in the world every once in awhile.
  • On the second route, the google route is probably better than Garmin, but isn't optimal. The exit it gives is different than the Garmin.

    But does Google Maps have the exit number marked incorrectly? I just want to be sure I'm following the correct (bad) route.
    Sorry, I've been off line this weekend. The google exit number I believe is accurate. I'll verify next time I'm down that way.

    As far as the question about which problem are we trying to fix, in my mind there is only one problem. I think the Garmin does a decent job of estimating time of arrival. When doing mostly highway driving, it actually does an excellent job. I'm almost 100% sure that the problem with the Garmin is that it does not take into account "real" local road driving speeds. I'm pretty sure if they just tweaked their algorithum to take into account stop sign/light times on local roads it should take care of the problem. Has anyone found any online map or GPS unit that would have you take hwy 41 on the MKE to CHI route? I doubt it. It is obvious that saying on the interstate is fastest.... unless you are calculating local road speeds incorrectly.

    I'm not sure if we really should be tying this issue to estimated time of arrival. I think the underestimation of local road speeds manifests itself in the ETA when traveling on a city street... but I'm more concerned with getting the fastest route. I'm less concerned that the ETA is off by a few minutes.

    I have noticed when driving on city streets, the ETA will show, lets say 5:45. After sitting at my very first stop light, it will update to 5:46. If I were programming the Garmin, I would at least try to estimate the number of lights on a city street, the number you may hit and the length of each light before calculating a best route or an ETA. Obviously the calculation could not be 100% accurate since there are so many factors that could affect the number of stops you might have to make. But, since the Garmin starts increasing the ETA after the very first stop (full red light) I'm thinking it isn't taking stop lights into consideration at all.

    Just a note on the Blockbuster route. Of the 5 or so times I tried the Garmin route, 2 have been rush hour. The others have been weekends. Again, this is a route that all other systems that I have tried route differently. I wouldn't mind if the 765T routed me off of the highway if it spotted heavy traffic. Heck, that's what I bought it for, but it tells me to take that route even on days with no traffic.
  • Tim 1480 Points
    Agreed. Our focus on ETA isn't because we care so much about an accurate ETA, but rather the cumulative ETA is the only way we can take a peek inside the road database to determine what they think the speed for that road will be.

    Have you played around with building that route into smaller segments and analyzing each one? For example using the same starting point but ending just at the highway exit, how does the estimated transit time compare with your expectations? Likewise from the highway exit to the destination.

    Figure out where the estimated transit time is furthest off from your expectations and then we're starting to narrow down where the issue is.

    We don't have enough info to really pinpoint the issue-- but if someone where to make me take a best guess right now where the issue is-- I'd say that despite other devices and online mapping systems using Navteq maps that the issue starts there, and ends with how Garmin is applying the road classification.

    And don't forget-- while a map doesn't "time" stop lights they do try to take them into consideration by road classification. And, GPS devices are assuming no traffic unless otherwise fed with data. So sure, a device without traffic data (historic or live) is going to have bad time estimates in bigger cities during rush-hour.
  • I don't want to upset anyone, but I'd like to point out that this thread is intertwining 3 separate issues.

    1)Bizarre routing (more than simply off a little)

    2)Poor time of arrival estimates

    3)Disparity between visual and voice directions

    The original thrust was on bizarre routing (note the title of the thread). The concern of the OP was that his unit directed him to side roads that take much longer to reach his destination while ignoring a perfectly good freeway. I and some others have had similar experiences.

    If I understood Tim correctly (apologies if I didn't) this is the issue that Garmin is looking into.

    The other two issues maybe related to a common software problem, but for the sake of discussion, wouldn't it be simpler if each issue was treated separately in different threads?

    If you agree and are experiencing issue two or three, may I suggest you start a discussion of the problem? This thread could then stay focused on the routing issue. If you disagree, no reason to flame me, just do what my wife does and ignore me :)
  • Agreed. Our focus on ETA isn't because we care so much about an accurate ETA, but rather the cumulative ETA is the only way we can take a peek inside the road database to determine what they think the speed for that road will be.

    Have you played around with building that route into smaller segments and analyzing each one? For example using the same starting point but ending just at the highway exit, how does the estimated transit time compare with your expectations? Likewise from the highway exit to the destination.

    Figure out where the estimated transit time is furthest off from your expectations and then we're starting to narrow down where the issue is.

    We don't have enough info to really pinpoint the issue-- but if someone where to make me take a best guess right now where the issue is-- I'd say that despite other devices and online mapping systems using Navteq maps that the issue starts there, and ends with how Garmin is applying the road classification.

    And don't forget-- while a map doesn't "time" stop lights they do try to take them into consideration by road classification. And, GPS devices are assuming no traffic unless otherwise fed with data. So sure, a device without traffic data (historic or live) is going to have bad time estimates in bigger cities during rush-hour.
    There isn't any one section that is off more than the other, other than it starts when exiting off the hwy and on to city streets. You say that the GPS does try to take stop lights into account, but I beleive that is where the problem is. My Garmin does not seem to. Like I said previously, at the very first light I'm stopped at, the Garmin will start increasing the ETA. If it did take stop lights into account, then it would not need to at the very first stop. Driving highway, the ETA is as good as I could hope for. City streets, not so much. I think you're on to something possibly with how Garmin is classifing the roads. Could it be that it isn't classifing city streets correctly? If it is, then it just isn't allowing enough (any) time for stops.

    To test my theory, the next time I'm taveling on city routes, I'm going to try to calculate the amount of time I'm stopped at stop lights during my route and compare that to the difference in the initial ETA and the actual time of arrival. If my theory is right, the two will come out to be about the same.

    I'll let you know what I find. I might not have this until Saturday, so if anyone else out there can give this a try, let me know.
  • Tim 1480 Points
    If you agree and are experiencing issue two or three, may I suggest you start a discussion of the problem? This thread could then stay focused on the routing issue.
    In my opinion issue 2 is a symptom for identifying the cause of issue 1. So I don't mind if issue 3 is split out, but I personally think number 2 is helpful for diagnosing issue 1.
  • Not the most extensive test ever, but I did time my stop lights on the way home today. I only have three, but by luck I had to stop at all three. Initially, the Garmin said I would arrived home at 5:24. I had to stop at my first light just as it turned red. At nearly exactly 1 minute, the garmin increased my eta to 5:25. At the next light I stopped for about 50 seconds. The garmin did not adjust the time. My third and final light, I had to stop for about 20 seconds. At almost exactly 10 seconds into the light, it increased the ETA to 5:26. I'd like to try this a few more times, but it seems to prove that Garmin isn't allowing for any stop time on local/city streets.
  • cuc tu 91 Points
    To make it easy, could you not just plan a long route on a lighted street and calculate the expected drive time based on the speed limit, then compare to Garmin's ETA? No driving/stopwatch required...
  • Marc 301 Points
    Try this. Run a simulation and see what speed the GPS puts you at for those roads. If it is below the speed limit there is some type of allowance for traffic and lights.
  • Tim 1480 Points
    I'd like to try this a few more times, but it seems to prove that Garmin isn't allowing for any stop time on local/city streets.
    I agree in principle, but not precisely. There is an implication that the road database doesn't take these things into account. That isn't really correct. They do try to take those things into account. Now if they get it right or not for each road classification/type is another matter-- and likely what you are seeing. But I'll just nitpick on the wording so we don't get any rumors started.

    So based on what you describe, the GPS thinks you will transit the stretch of road faster than you actually likely will. This is a bit different than saying the GPS expects you will hit every light green. While the net result is similar (actual drive slower than projected) the same could be true for any type of road.
    Try this. Run a simulation and see what speed the GPS puts you at for those roads. If it is below the speed limit there is some type of allowance for traffic and lights.
    Yeah, that's probably a better way to explain it. :)
  • cuc tu 91 Points
    I just plotted a route in mapsource along a main city road.

    26.4 miles with lights probably at least every 1/4 mile or more.

    The ETA is 48.39 minutes, which would take an actual 45.25 minutes if one was able to travers the entire distance going the 35 MPH speed limit...

    I think that three some odd minutes total waiting at over 100 lights is underestimating the time, never mind the slinky effect...
  • If anyone is looking for another example of bad routing - try this route from the intersection of rt109 and Pound St in Medfield, MA, 02052 to Boston Logan Airport Terminal A.

    Nuvi 755t - Map 2010.10 - FW 3.60 (Fastest Time, Avoidances Disabled)

    7.8m Right on High St
    0.7m Left on Washington St
    4.6m Right on Arborway
    0.1m Left on Washington St
    2.4m Right on Dudley St
    0.1m Left on Harrison Ave
    0.3m Right on Melnea Cass Blvd
    0.6m Keep Left onto Mass Ave Conn to I-93N
    0.2m Left onto Frontage Rd to I-93N
    0.4m Right on I-90E
    0.9m Keep Left on I-90E
    1.8m Exit 26 right to Logan Airport
    0.3m Keep Left onto Road to All Terminals
    0.9m Arrive at Logan International- Terminal A

    Distance 21m Time 36m 38s (Way underestimated)

    Curiosity got the better of me and I actually drove this route last Wednesday at 2.30pm. It took 57m 42s timed on a stopwatch. Traffic was light and I don't think I could have done it much faster at 2.30am with zero traffic. There are so many traffic lights and intersections on this route that my Nuvi doesn't seem to account for.

    Google Maps gives the following:

    1. Head east on MA-109/Main St toward Hatters Hill Rd
    Continue to follow MA-109 6.6 mi
    2. Merge onto I-95 S/MA-128 S via the ramp to Canton/Providence Ri 4.0 mi
    3. Continue on I-93 N/US-1 N (signs for Boston/US-1/Braintree/I-93) 15.3 mi
    4. Take exit 20 to merge onto I-90 E 3.4 mi
    5. Take exit 26 0.1 mi
    6. Keep left at the fork 0.3 mi
    7. Keep left at the fork 0.3 mi
    8. Continue straight 0.1 mi
    9. Continue straight onto Terminal A
    Destination will be on the right 0.2

    30 miles Time: 40min (which is about right)

    I have driven to Logan Airport four times in the last week - if I ignore the Nuvi it takes me 35-40min - following the Nuvi almost 50% longer. This is not an acceptable margin of error. I have also experienced the inexplicable diversion off and back onto the highway on a couple of occasions.

    I know no GPS is perfect, but I shouldn't have to check with Google maps all the time to see if the route is reasonable. My IPAQ with TT6 seems to route much better. I do like the 755t features, but lets get the basics right.
  • SL 0 Points
    Not the most extensive test ever, but I did time my stop lights on the way home today. I only have three, but by luck I had to stop at all three. Initially, the Garmin said I would arrived home at 5:24. I had to stop at my first light just as it turned red. At nearly exactly 1 minute, the garmin increased my eta to 5:25. At the next light I stopped for about 50 seconds. The garmin did not adjust the time. My third and final light, I had to stop for about 20 seconds. At almost exactly 10 seconds into the light, it increased the ETA to 5:26. I'd like to try this a few more times, but it seems to prove that Garmin isn't allowing for any stop time on local/city streets.
    In my opinion, Garmin assumes you will not stop at all, whether driving in hwy or local and I think that is acceptable consumption because no one knows which light and for how long you will stop. It happened with me and I never stopped around 10 lights and sometimes I spend 2-3 minutes in couple of light. However, when you stop at any light or there is delay, accident...etc, Garmin will adjust your time accoringly, and I think the way it does it, it calculates the remaining distance and divide it by the posted speed. For example, if your distance is 65 mile, starts and ends on hwy, and the posted speed is 65, then you will get there in 1 hour, if you drove 58.5m/hr then you will get there in 1hr6m....etc
  • Wulf 0 Points
    Thanks, mitrajoon for trying to get this back on track. I am the OP and I have more to add. I last stated that a few hard resets "seemed" to fix my original problem of the 205 taking me on back roads when the Interstate was better. Remember the 205 had nothing as far as traffic warniongs or any radio at all so that can not influence it.

    Anyway, my daughter took the 205 after it seemed to be working OK again and she changed a few of the settings, she can't remember exactly what, but I know she somehow set it to "off road". I reset it to my original settings. Now it is doing the same thing, trying to take me off every exit for mile after mile when the interstate is the fastest way. It doesn't matter what avoidences I have set and it does not matter if I have shortest distance or fastest time or least fuel set. I will try hard resets again and see if it fixes it again, but this is not the way these things "should" work.

    So all of your figuring that it may be any of these things are eliminated.

    Time of arrival is just a estimate based on time and distance so I can't see it being an issue if it is off a little. Poor routing is a much bigger issue, when you don't know where you are.

    Without all the bells and whistles on the highte models, the 205 should give good basic RELIABLE, CONSISTANT, directions every time, reguardless of traffic and detours, since it does not know about them.

    This was my original complaint and as mitrajoon pointed out, this thread has taken on a few different issues. I'm somewhat encouraged that some say Garmin is looking at this issue, but I wonder how they will fix it, if they ever do, fix it.
  • Tim 1480 Points
    Time of arrival is just a estimate based on time and distance so I can't see it being an issue if it is off a little.
    The fastest route is found by looking at route possibilities and figuring out which one should have the smallest value for time.

    Time = distance * estimated speed

    Since the distances it reports for each road are correct, that leaves estimated speed as the logical place for error.

    So if the speed estimate for a road segment is way off for some reason, then the time of arrival (as a symptom of estimated speed) could be the cause of poor routing choices.

    This is why Marc's suggestion will best help figure out where the issue exactly is. If during the simulation the car is consistently showing 45 mph on a road that is "stop and go" with the "go" being around 30 mph then we've pinpointed the issue.
Sign In or Register to comment.
↑ Top