| |
| ProfessionGPS |
|
|
| POIenFrance |
|
|
| Visiteurs |
Actuellement en ligne: 248
Dont membres: 2
+ d'infos...
|
|
|
| GPS Chipsets and Receivers Compared |
5. On the trail (pedestrian)
 This is an area where GPS use has always been subject to avid discussions in the forums, both because having a GPS while walking in an unknown town or in the wilderness is a "no-brainer" but also because results are not always very satisfactory due to difficult environments (buildings in town, tree cover), poor placement options and not in the least due to the 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 but let's see how the GPS chipsets at hand deal with various sets of environments. In keeping with the other tests, I placed various combinations of receivers side by side in my backpack and logged the results.
Downtown Paris - ST/SS3/uNAV
Similar to the first series of driving tests, the first pedestrian ones were done in the dense urban environment of the "Halles" area in Paris where there's a lot of interference on hand for GPS signals although not the urban canyons one would find in Manhattan where I will be doing some testing shortly too. Here are pictures of the results and comments are below:
Walking was done upper left to upper right. Again uNav produces a puzzling result that is not usable and this time SS3 and ST struggle too with a lot of drifting likely due to multipath and the "predictive" algorithms that would work at a sustained speed can't help much here. A bit of a disappointment that the powerful correlation of SS3 can keep the fix but not always match the actual track.
Bry sur Marne - footbridge and parc - ST/XT1/SS3
Sticking to difficult environments, let's see how these chipsets deal with footbridges located under a train track (see photo).
While a bit surprising at first, the result is consistent with what was seen in Paris, and the recorded track alongside the footbridge seems to have been "pushed away" by a powerful magnet. Fortunately on the bank of the river, the track of the loops is quite consistent with the actual walk. Xtrac v1 does show that it was quite a performer for walking purposes as geocachers had found, this leaves SS3 unimpressed though!
Le Perreux - downtown - ST/XT2/SS3
Moving on to the more GPS friendly downtown of Le Perreux with its fairly large streets and medium sized buildings.
  Fig.p10 - Fig.p11
The good news here is that GPS comes back into its own with an SS3 track that closely matches reality and so does ST...when it can get a fix. On the other hand XT2 does show that it was (well) tuned for driving and lost most, if not all, of the effectiveness of XT1 for pedestrian use.
Bry sur Marne - downtown - Nemerix/Sony/SS3/uNav
Back in Bry, downtown this time with the other chipsets whose performance I did not hold high hopes for in light of the results of the driving tests.
The above comments apply here too althouh this time it's the uNav chipset that doesn't hold the fix and the Sony one that "wanders". SS3 is off to a difficult start dealing with walking alongside a wall and "bouncing" off as a result but then falling back on its feet.
SiRF's "static navigation" setting - XT2/SS3
Since we're discussing GPS use on foot, there's one setting available of SiRF based receivers that you probably want to check for before setting out, it's called "Static Navigation" (SN) and here's the link to the tutorial about it and other advanced SiRF settings.
 Fig.p16
The tutorial goes into more detail about this, but SN filters out low speed movement having a rather dramatic impact on SS3 more so than on XT2 so make sure it is deactivated on your GPS before setting out for a hike if you want to follow your track to match reality as closely as possible.
If you have questions or comments, you can use this discussion thread in the forums
|
|
|
|
|
|
