GLONASS
  • Bouje
    Posts: 9
    after using my oregon 450t, I've decided the screen is too small for my preferences, so I've turned my sites to the Montana. The oregon seemed also to have a little bit of trouble in the Congaree national park, but when I say little I truly do mean little... For the most part, it seemed to track me reasonable well but there seemed to be a lag at times... What I'm wondering is can getting a system that had both GPS and glonass capabilities help with this? I've heard the new model of the oregon coming out soon will be glonass capable?? Will the Montana ?
  • alanb
    Posts: 2,245
    See the posts by Boyd part way down on this thread regarding the garmin GLO receiver http://forums.gpsreview.net/discussion/26778/x/p1/. It would appear from his preliminary testing that he did see improved tracking using the glonass enabled receiver.

    Regarding a Glonass capable Montana model, I have read rumors that such a device is in the future and it seems likely since Garmin has added this to the new eTrex and Oregon models. But as far as I know Garmin has made no announcements about a Glonass enabled Montana. It is anyone's guess about when or if a new Montana model will be announced or what its features will be.
  • Boyd
    Posts: 11,631
    The current Montana series uses the same chipset as your Oregon 450, so I would not expect any difference between them. Of course the new eTrex models have GLONASS, but they feature small screens and pushbuttons. There have also been complaints that they're slow to zoom/pan with complex maps and especially raster imagery (BirdsEye, Custom Maps). There was speculation that Garmin traded off performance to get GLONASS.

    Regarding the GLO, yes I'm very impressed so far. But don't read too much into that if you're thinking of a handheld like the Oregon 600. The GLO provides position updates 10 times per second (10hz). No Garmin handheld has ever featured that, they all only update your position once per second (1hz). Garmin doesn't even publish specs for that kind of thing anymore, but I'd be surprised if the new Oregon (or rumored Montana) provide anything faster than a 1hz update.

    Whether the faster update helps provide better accuracy, I can't say. It seems logical that if you're using waypoint averaging though, having 10x more points to average would be better. The main reason they are offering the 10hz rate on the GLO is because they're positioning it for aviation use, where you can travel a lot farther in 1 second than you can on foot.😀
  • ddabcd277
    Posts: 238
    It seems logical that if you're using waypoint averaging though, having 10x more points to average would be better.

    Doubt, because no new data from the sattelites. Just same data multiplyed 10 times. But really very useful for aviation or in high speeds.
  • Boyd
    Posts: 11,631
    Are you sure about that? Are you saying that the device only gets a fix once per second but interpolates a new position 10x per second? That wouldn't make much sense, because the application software itself could do that.
  • ddabcd277
    Posts: 238
    No, I ment different. I mеnt that it won't increase the accuracy those 10 points if you are using waypoint averaging. Because those 10 points have same accuracy as 1. More points doesn't mean better accuracy. It doesn't metter if you have 30 or 3 points.

    I have heard this from a surveyor. He told me that when when he is taking the coordinates of a point with 2Hz geodetic gps he is taking a point even every 10 seconds because the position of the sattelites haven't change in a shorter period.

    That is why maybe in the new Garmin handheld it is written to make one sample and then make a new sample after 90 minutes in the waypoint averaging. This is the best way for increasing the accuracy of the point.
  • Boyd
    Posts: 11,631
    Sorry, but that just doesn't make sense to me. When you use a Garmin handheld with waypoint averaging, it takes a new sample once per second.

    Try this sometime with a handheld: set tracking to record by time with a one second interval. Now put the GPS stationary and let it record the track for, say, half an hour. When you look at the log it will be a "spiderweb" showing how your much each sample has varied over time. Older devices like the 60csx should show quite a large scatter pattern when doing this.

    However, if you average all those track points, you should have very accurate coordinates of where the GPS was sitting. This is what waypoint averaging does. So I can't understand why it wouldn't be a good thing to have 10x more data to average. In my scenario the tracklog would contain 60 x 30 = 1800 points. You could obtain 1800 points in 3 minutes with the GLO, as opposed to 30 minutes with a handheld.

    BTW, I have not found any Windows software that records more than one point per second, so I haven't been able to look at the full data stream from the GLO yet.
  • Boyd
    Posts: 11,631
    I did a test today comparing the GLO, Montana and 60csx. I recorded tracks for about 30 minutes on each device and used double-precision coordinates in Globalmapper to convert the tracks to points, yielding about 1800 points from each unit.

    As expected, I got big shotgun blast from the Montana and 60csx, but they should average to within about 5 meters of ground zero. Now the GLO looks disappointing at first, with some points wandering way off to the East.

    image


    But when you look at the actual data, it tells a very different story. In fact, many of the points are right on top of each other and form a very tight and accurate pattern. The image is quite deceptive - only 41 out of 1862 points were more than 10 meters away. In fact, 98% of the GLO's points are within 10 meters of ground zero and 87% of them are within 3 meters.

    Now the Montana and 60csx also got just about everything within 10 meters, but clearly they don't have nearly as many points within 3 meters. This GLONASS stuff is looking pretty good to me.😃

    image
  • babj615
    Posts: 346
    Very interesting, Boyd, I enjoy your test reports. Added another question in the other thread.
  • Boyd
    Posts: 11,631
    FYI, I just realized that WAAS was not enabled on either the Montana or 60csx (thanks babj615). There's no option for that on the GLO... in fact there's not really any option for anything. It's just a little black box. 😀

    Will try to do another test in the future with WAAS on the Montana and 60csx and see if that affects the results.
  • ddabcd277
    Posts: 238
    Boyd you are incredible! Very nice!😃 My statement is that if possible to make such a test which most probably wont be possible. Briefly if possible to set the GLO to 1Hz and if you record lets say 1 hour with GLO at 10 o'clock in the morning (1hour - 3600 points) and on the next day at the same time at 10 o'clock with GLO set to 10Hz (1hour - 36000 points) and if you average this data from the two days you would have same accuracy from the averaged points. This is what my statement is. It doesn't matter the amount of the points but the clock inside the gps. I now see that I didn't explain it good in the previous posts. Excuse me.

    About the Windows software - Visual GPS most probably would do it. Not sure.
  • Boyd
    Posts: 11,631
    Unfortunately you can't "set" the GLO to do anything. There is no user configuration. I am not sure how an application program and the GLO agree on an update rate, but there doesn't seem to be any way that a user can change this. All the programs I've tried only record 1 point per second. However, when I use the GLO with Mobile PC, it becomes much more responsive than it was with my USB GPS so I assume that is due to the higher data rate.
  • ddabcd277
    Posts: 238
    Ok. If you wish we can do it in this way. Record two days. One hour at the same time, same conditions. One of the days I will filter the data and take out every 10th point. (3600 points) Should be same. Then compared it with the average points from the previous day 36000 points and we will see the result. I will average them. No problem.
  • Boyd
    Posts: 11,631
    Sorry, I will have to leave that to someone else. All this testing is starting to feel like "measurbating" to me. 😋 But I am repeating yesterday's test right now and will post the results soon. This time I made sure that the Montana and 60csx have WAAS enabled.

    I am also not sure about those "rogue" points that appeared with the GLO. They look suspiciously like the track from my house to the test location. Since I was using Globalmapper to record the GLO track, I may have messed something up as the interface is rather awkward, especially on a tablet.
  • Boyd
    Posts: 11,631
    Here are my new tests. All I can say is... what a difference a day makes! This time I was certain to enable WAAS on both the Montana and 60csx and made sure that the GLO track was properly recorded. I've included a closeup of the GLO. As you can see, many of the points were right on top of each other.

    You can draw your own conclusions. Personally, I'm just blown away by the GLO results. 😀

    image


    image


    image


    image


    image
  • Boyd
    Posts: 11,631
    Boyd said:

    BTW, I have not found any Windows software that records more than one point per second, so I haven't been able to look at the full data stream from the GLO yet.



    Turns out that I already had software for this, just didn't know how to use it. :lol: OziExplorer seems to work, but the key is that you must go to File > Configuration > Moving Map > Store Track Point Interval and enter a value of 0 (zero).

    Will do some further testing of this soon.
  • babj615
    Posts: 346
    Thank you Boyd! This is certainly a very interesting topic!
  • Tim
    Posts: 19,398
    Boyd, I might have missed this in your notes, but I assume this is looking at precision and not accuracy? (Not that it matters any less, very impressive, I'm just curious.)
  • Boyd
    Posts: 11,631
    I'm going to have to show my ignorance here about the use of those terms. :oops:

    The devices were stationary and recorded one point every second. The cross in the center of the image is a known point and each dot represents one of the ~1800 points recorded by the GPS over a 30 minute period.

    I believe I've seen accuracy stated along the lines of "there's a 95% probability that your fix is within 5 meters of the actual position". I can infer that kind of thing from this data. In the most recent example above, the GLO had a 75% chance of being within 3 meters and a 100% chance of being within 4 meters. In other words, all of the recorded points were inside a circle with a radius of 4 meters.

    Is that what you're talking about?
  • babj615
    Posts: 346
    Accuracy vs Precision:

    image
  • Boyd
    Posts: 11,631
    I see. Well, you can draw your own conclusions as to which device matches those different examples. 😀
  • babj615
    Posts: 346
    Based on your data, Boyd, I would say the GLO was the only device exhibiting both accuracy and precision.

    As for the Montana and 60csx, well, I think they just need a bigger target before they may begin exhibiting any precision, but with all that scatter, I wouldn't venture to consider them accurate.

    IMHO, of course. YMMV😃
  • Tim
    Posts: 19,398
    Boyd said:

    Is that what you're talking about?


    Yep, pretty much since you were working from a known (versus random) starting position. babj615's graphic illustrates it well that you can have high precision but poor accuracy.

    With regards to your figures, the scientific community (or at least those I've worked with) generally assumes 95%. (So in terms of this conversation it would be at what distance to the closest 95% of points fall within.)

    As to Garmin, I suspect they have tweaked things over the years but one of their engineers once told me they tune their "accuracy" field on the display to be at 50%. So if the GPS says it is currently accurate to within 5 meters then half of the readings will be outside of 5 meters and half will be within.
  • Boyd
    Posts: 11,631
    I just returned from a short walk carrying the Windows 7 tablet while the GLO was in my shirt pocket. All I can say is wow. I think these images speak for themselves. I have never seen a track log this accurate.

    First I used Globalmapper and it accurately tracked my position on the screen but only records one point per second. On the return trip, I used OziExplorer at the full 10 points/second. I recorded over 2400 points during about a 4 minute walk).

    At the end, I walked around in a rough circle - I did not exactly re-trace my steps the second time around. As a point of reference, you are looking at the NJ 2007 orthoimagery at 1 foot per pixel. So those big blocks in the bottom image are one-foot squares.

    Very impressive for a $90 device IMO. Add a smartphone or small tablet and the correct app, and you have an impressive tracking device.😀


    image


    image
  • Tim
    Posts: 19,398
    Wow!
  • Boyd
    Posts: 11,631
    Today's test wasn't so good. I'm wondering if OziExplorer was having a hard time keeping up with the 10hz data from the GLO? This was a longer walk than my previous test. Am going to try another test but use Globalmapper to record the GLO at 1 hz this time....


    image
  • Boyd
    Posts: 11,631
    Here's my new (and final) tracking test. My initial excitement about the GLO accuracy has been dampened somewhat; it actually seems about equal to the 60csx and Montana for this kind of use. Nothing wrong with that though, and the static tests posted earlier still impress me.

    I've tried to document this as well as possible in the hope that it may help others in the future. One thing to keep in mind when looking at the images below that are shown at such extreme zoom... the NJGIN imagery metadata says it is accurate to +/- 4 feet (~1.25 meters) at 95% confidence level

    I purchased the GLO as a receiver for Garmin Mobile PC running on a Windows 7 tablet in the car, and it works perfectly for that. But I will continue to use my Montana out on the trail 😃

    image

    image

    image

    image
  • ddabcd277
    Posts: 238
    Maybe because GLO is more sensitive than the others that is why you have better accuracy on a point but not when recording a track. Let me explain I have heard that when sirf III was released it was the most sensitive receiver but there was an opposite effect this high sensitiity didn't increase the accuracy because with the good signals the gps also received "bad signals" - modified signals, reflected signals etc. That is why it "jumps" a lot. But it also metters what kind of filters the GPS is usings.

    Also for better accuracy the DoP factor is determining:

    Bad DoP factor:

    image

    Good DoP factor:

    image

    If the DoP factor is close to 1 this means the accuracy is good. (Hard to be calculated. A lot of matrixes for calculation.)

    Source:
    http://en.wikipedia.org/wiki/Dilution_of_precision_(GPS)

    Also the mask of the satellites.

    Just writing for those who is interesed.
  • babj615
    Posts: 346
    Boyd,

    Do you know what the track log recording settings (records method, interval) were (for each unit) in that final test?
  • Boyd
    Posts: 11,631
    Yes, for all tests the tracklogs were set to record by time with a one second interval. That is the rate at which Garmin handhelds provide position updates, so when you record with this setting you're getting everything that the chips put out.

    With the GLO, I used Globalmapper and it is hardwired to record one point per second from a GPS device.

    In the earlier test(s) OziExplorer was able to record the 10 position updates that the GLO provides each second (10hz update), but I'm convinced that it just can't cope with it very well (see earlier post). With Ozi, the only options I could find were to specify distance-based track recording. However if you set that to a value of zero feet, it records everything that the GPS receiver puts out.

    Now look in my earlier post about the OziExplorer problems. You will see that the track points from the 60csx and Montana are very evenly spaced, which you would expect from the 1 second interval. But there's a spot near the top of the 60csx track where there appear to be some gaps between points. I wonder what happened there? Did it lose the fix for a couple seconds? Or maybe I pushed some buttons on the 60csx and the processor was too busy to record a point at the same time?
↑ Top

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!