| Versions |
  |
|
| Author |
 Topic  |
|
trs2004
19 Posts |
 Posted - 13 mars 2008 : 23:20:08
|
Just how accurate should a GPS track be?
After applying the patch for SNoff, I tried the windowsill test again. The result were a bit better than before, with most of the wandering inside a 12m radius. How does this compare with other models?
Next I went for a walk. Once I got a fix, I put the unit in an outside pocket of my coat where it stayed for the whole trip.
Anomalies
1. The GPS track starts about 144m NE from where I was standing while waiting for a fix. This might be a spurious data point during startup
2. The GPS track shows me back-tracking about 70m
3. The GPS track shows me taking a 160m detour to the far end of a parking lot but doesn't show that I crossed the street at the corner.
4. The GPS track shows me crossing the street and then returning after 12m.
5. The GPS track shows me on the wrong side of the street for 90m. This is open sky without any buildings nearby.
6. The GPS track again shows me crossing the street and returning
This walk was done with the unit in mode 3. I'll have to try it again in mode 1 to see if more frequent sampling will improve or degrade the track.
For general location and just to help remind me where I went, it might be acceptable but if I really need to know my actual path on a trip, I think I might have to look elsewhere. Too bad since it had so much going for it, driverless, easy to use.
Am I expecting too much from this class of GPS logger? Will I be just as disappointed by other models? |
 |
|
|
gpspassion
93389 Posts |
|
|
trs2004
19 Posts |
Posted - 14 mars 2008 : 01:02:54
|
Thanks for those links. Now I know a bit more about what would be considered normal for a GPS. I especially like the part where
quote:
GPS system's natural "inaccuracy" of approximately 10 meters. While this inaccuracy is easy enough to "rub out" with a properly tuned firmware inside the receiver in driving conditions (60 kph is 17 meters/second) that you can predict to some extent, it's much more difficult if you're walking at 5 kph or 1.4 m/s, well below the accuracy of GPS each second. It's unlikely that much can ever be done to fix this latter aspect ...
Since I was testing in mode 3, @5 sec I was still under the 10m threshold. I'll have to try it in mode 5 @10sec instead which should put me over that point, around 14m |
|
|
|
trs2004
19 Posts |
Posted - 14 mars 2008 : 23:53:05
|
After resetting to mode 5, I went to walk the same path with the unit again in an outer pocket of my jacket. Here is the result.
1. There still seems to be a problem at this corner.
2. The sideways jog appears where I went under a train bridge. All of the steel seems to have confused the unit.
As for the rest of the track, the AMOD AGL 3080 did a very good job of tracing my path, accurately showing where I crossed the street. After reading the links provided by gpspassion, I have a better understanding what the built-in limitations of GPS units are. Today's results have renewed my faith in this unit.
As a final test, I will have to retrace the route in mode 1 to see how much of a difference SNoff makes compared to the original configuration. |
|
|
|
gpspassion
93389 Posts |
Posted - 15 mars 2008 : 14:42:35
|
Thanks for sharing your tracks, yes the problems you're seeing seem pretty "standard" to me, they are in areas where multipath (bounced signals) could kick in and at low speeds GPS algorithms can't cancel them out easily. The ways to improve that would be using a large choke-ring antenna mounted on a pole ;-) Just one thing, in the extract you quoted above I'm referring to the general update rate, the logging update rate doesn't come into play, the unit just records what the GPS module outputs, there is no tweaking done at that level.
On another matter I just got rechargeabl 1000mAh NiMH AAA batteries to operate the 3080, I'll see what kind of battery life I am getting. |
Discounts and Assistance/Réductions et Assistance (Club GpsPasSion) / Où commencer? |
|
|
|
trs2004
19 Posts |
Posted - 16 mars 2008 : 19:36:36
|
Here is my latest test track. The same route as before but this time using mode 1 from the SNoff patch.
1. A wild spike when the unit found itself after a cold start.
2. Wandering for a bit on the wrong side of the street.
3. Still a few problems at this corner but then continues correctly, even under the train bridge.
4 and 5. Wrong side of the street
6. The corner is almost correct but then continues on the wrong side of the street.
Of the three modes I tried in a pedestrian setting, I found that mode 5 @10 sec gave the best results. In my case, more is definitely not better. Since I don't plan on taking many pictures from a moving platform, the longer time between data points will not be an issue. I will most likely be standing still for at least 10 seconds while composing my next picture.
|
|
|
|
TonyW.
USA
29 Posts |
Posted - 28 mars 2008 : 16:25:08
|
I was happy to see the utility for the logger was available but I can't get their site to come up .
After two days of trying I'm going to ask if anyone can make it available for a download somehow ?
Thanks. |
Edited by - TonyW. on 28 mars 2008 23:20:54 |
|
|
|
gpspassion
93389 Posts |
|
|
TonyW.
USA
29 Posts |
Posted - 29 mars 2008 : 12:06:31
|
I was talking about the firmware update utility.
I finally got it this morning.
Thanks |
|
|
|
collinm
40 Posts |
Posted - 01 avr. 2008 : 09:50:19
|
| somebody can test if the new firmware improve the precision? |
|
|
|
wavshrdr
USA
152 Posts |
Posted - 16 avr. 2008 : 19:57:44
|
| I just bought a new AMOD from Semsons. It came pre-loaded with the new firmware. So far my testing of it looks pretty good. Acquisition times are very good as well. The new modes are really useful. I'll be trying them all out. I only wish this thing was usable via a USB cable to use as a GPS for on of my laptops as well. |
|
|
|
TonyW.
USA
29 Posts |
Posted - 17 avr. 2008 : 18:23:32
|
| That would certainly be another nice feature.I'm wondering if that could be done through software ? |
|
|
|
wavshrdr
USA
152 Posts |
Posted - 20 avr. 2008 : 03:34:42
|
I did some extensive testing today while riding my Xootr push scooter. I went on about a 5+ km trek around an urban area. I don't know how to post a track on here so show you the results but I am quite satisfied with them. My average was about 13km/h and max speed about 29km/h. The AMOD did an excellent job with my track.
Considering that the accuracy can be degraded at any time, my track was never off by more than about 1.5 meters. 95% of the time though it had me EXACTLY where I went with my scooter. I was staying to the bike paths for a lot of my route and they were denoted by white lines 1 meter apart. It showed me consistently always between them. I had it suspended from the collar on my jacket and it had no issues picking up a signal even while bouncing around as I kicked the scooter.
This thing is definitely a keeper for me. Add a little LCD, BT and make it usable via a USB cable as well and it would be tough to beat. I love the fact it takes normal AAA batteries and has given me about 13 hours or so real usage on Sanyo Eneloop batteries. These batteries are awesome as they don't self-discharge near as quickly as normal NiMH batteries do. So even if I don't use them for a month, they still have a good charge on them while my other ones will be most likely flat. |
|
|
|
wavshrdr
USA
152 Posts |
Posted - 24 avr. 2008 : 05:35:52
|
| Anyone know of an easy way to get the AMOD in sync with a camera? Is there a way to access the internal clock of the GPS on the AMOD so I can make sure the times are in sync for both devices? I tried exporting a log while marking a waypoint and it always seems I am off from a few seconds to a lot more than that. Any advice? |
|
|
|
collinm
40 Posts |
Posted - 24 avr. 2008 : 09:15:11
|
surely amod sync it's with with satellite...
sync your camera with atomic clock... check website who use ntp |
|
|
|
Topic |
|
|
|
| This page was generated in 1,45 seconds. |
 |
|
- 
- 
[TOPIC] AMOD AGL 3080 Photo Tracker - v2.5 update
