Accuracy question

[matanglawin]

What is the minimum level of accuracy (measured in meters) needed to produce acceptable tracks and waypoints? I have observed that my Nuvi205 registers 4-6 meter accuracy in the areas that I usually frequent or pass by.

Thanks for your time.

[Borgie]

What is the minimum level of accuracy (measured in meters) needed to produce acceptable tracks and waypoints? I have observed that my Nuvi205 registers 4-6 meter accuracy in the areas that I usually frequent or pass by.

I think 4-6 meters accuracy is already very good and the best that we can get in the Philippines with our GPS devices that's not dedicated to GPS surveying. For tracks, it would be best if recording is set to "Most often". For waypoints, some GPS handhelds (like my 60CSx) have the capability to average the location of a waypoint, resulting to a more accurate mark. In your case, with that level of accuracy, it should be good enough.

But some factors also come into play: like did you record the waypoint while inside your vehicle, or directly standing in front of the POI? When waypoint was downloaded in Mapsource and overlayed in the roadguide map, is it in it's correct position (or side of the street) as observed in the field? Maybe that particular road in our roadguide map is inaccurate also, resulting to a wrong location of the waypoint in the map, how would you resolve this? In a very big POI, like for example SM-MOA, where would be the ideal place (in MOA) to record the waypoint? Similarly, for a POI located in a street corner, from what side of the street intersection would you place your mark (this will affect routing, esp. for 1-way streets)?

Hope this answers your question partly and did not create more questions...

[matanglawin]

Thank you very much mrBorgie! That actually reinforced some of the concepts that I gathered from reading and from actual experience.

I only record waypoints of POIs with the unit handheld and usually standing in front of the establishment. Tracks are recorded on board my vehicle.

This question of mine actually came about after I discovered that the tracks I recorded in the subdivision where I reside "fit" the google earth image better than the road map that I have in my Nuvi. Inisip ko na lang na baka kailangan 1 meter or less talaga ang "minimum level of accuracy."

[amf]

Accuracy is more critical in recording tracks than in establishing POIs. That is the reason why I use an external GPS antenna when recording tracks.

[matanglawin]

Accuracy is more critical in recording tracks than in establishing POIs. That is the reason why I use an external GPS antenna when recording tracks.

Absolutely. So how is the accuracy like with the reradiating GPS antenna sirAMF? What figures does it give you?

[Borgie]

Accuracy is more critical in recording tracks than in establishing POIs. That is the reason why I use an external GPS antenna when recording tracks.

+1

[amf]

Thanks!

With an external antenna, my Nuvi usually shows a 3-4m accuracy (a few times at 2m and 6-8m without), iPhone 9-17m (17-47m without) and my BB 8900 4m.

As far as POIs go, as Borgie intimated (ideal place to record the POI-- corner? left/right?), Google Earth polygons/traces will help a lot. If say a GE polygon of SM Mega Mall is submitted and incorporated in the map, a colored bird's eye view of the polygon is shown while navigating. The name of the building (POI) will just be simply located in the middle of the polygon. You may have noticed a few of these in the map.

That's why contributions of GE traces/polygons of rivers, parks and buildings are very welcome and not very hard to do (and does not consume gas ) though very addicting.

[Borgie]

As far as POIs go, as Borgie intimated (ideal place to record the POI-- corner? left/right?), Google Earth polygons/traces will help a lot. If say a GE polygon of SM Mega Mall is submitted and incorporated in the map, a colored bird's eye view of the polygon is shown while navigating. The name of the building (POI) will just be simply located in the middle of the polygon. You may have noticed a few of these in the map.

That's why contributions of GE traces/polygons of rivers, parks and buildings are very welcome...

Yup, GE traces should be welcomed. Would submitting a few dozen traces/polygons qualify for contributor status? This might be the way to go for those who think that their area is pretty much covered in the roadguide map already and they're having a hard time getting suitable POI's/tracks for contribution. And why a few dozen, if you ask me? Because...

...not very hard to do (and does not consume gas ) though very addicting.

Would be nice to see traces/polygons in our roadguide map similar to those malsing maps...

[amf]

Yup, GE traces should be welcomed. Would submitting a few dozen traces/polygons qualify for contributor status? This might be the way to go for those who think that their area is pretty much covered in the roadguide map already and they're having a hard time getting suitable POI's/tracks for contribution. And why a few dozen, if you ask me? Because...


Would be nice to see traces/polygons in our roadguide map similar to those malsing maps...

More than a few dozen would suffice as it's a pretty easy thing to do.

Yeah malsingmaps have them all over their maps. Pretty neat. I would certainly encourage contributors to do some GE tracings/polygons. Would be very useful in Makati and some other urban areas.

Thanks!

[B2Bomber]

was meaning to ask the same question earlier too, as I was getting 6m accuracy earlier while I was logging my hotel's WP.
IMO, short of using cm-accurate DGPSes, maybe if it was possible, we could "log" WPs by turning on the track log for a minute and do a mean position of the POI and publish it as the WP coords -- like what we normally do for final position fixes at work... ofcourse if there is something we could use to (at the very least) convert the tracklog values into computable values for spreadsheets, etc.
as for tracks, I'm not privy as to how routes for this endeavor are programmed but, IMHO - tracks on opposite direction ought to be more accurate ways of "charting" roads. Surveys are usually done on opposite passes to offset errors (pitch/roll/position jumps/bad data/etc.)

[Borgie]

was meaning to ask the same question earlier too, as I was getting 6m accuracy earlier while I was logging my hotel's WP.
IMO, short of using cm-accurate DGPSes, maybe if it was possible, we could "log" WPs by turning on the track log for a minute and do a mean position of the POI and publish it as the WP coords -- like what we normally do for final position fixes at work... ofcourse if there is something we could use to (at the very least) convert the tracklog values into computable values for spreadsheets, etc.
as for tracks, I'm not privy as to how routes for this endeavor are programmed but, IMHO - tracks on opposite direction ought to be more accurate ways of "charting" roads. Surveys are usually done on opposite passes to offset errors (pitch/roll/position jumps/bad data/etc.)

Hi B2Bomber, nice input. But afraid if we do this, some contributors and junior members not as tech savvy and techie might be turned off. Would be nice to be as accurate as possible without complicating the process... still user friendly kumbaga...

[B2Bomber]

oops, forgot, this is leisure positioning. :-P anyway, I will try to find a user friendly way of doing this -- at the least for the waypoint part (for now). er, when submitting WPs, are we supposed to comment on the accuracy too or send in a screendump of the constellation/accuracy report? TIA

PS: OT: yesterday navigated using the WP of my hotel, I got from the morning just before arriving Dr.Nightmare said "Arriving at destination on the right"... but the hotel was on my left, what was on the right (across the street) was Oriental Therapy -- an MP. whew...

[Borgie]

er, when submitting WPs, are we supposed to comment on the accuracy too or send in a screendump of the constellation/accuracy report? TIA

PS: OT: yesterday navigated using the WP of my hotel, I got from the morning just before arriving Dr.Nightmare said "Arriving at destination on the right"... but the hotel was on my left, what was on the right (across the street) was Oriental Therapy -- an MP. whew...

You can if you like. But best to check the wpts accuracy against the demo map, i.e. if the GPS position is to the left of the road but when overlayed in the demo map it's to the right of the road, guess it has to be moved to the left of the road for the roadguide map until somebody submits a better and more accurate road track (assuming of course that the road track's not accurate). Perfect example of this is your OT comment...

buti nalang hindi ka naligaw at napunta sa sinabi nyang destination...

[B2Bomber]

Thought of 2 ways to get a mean position fix (for waypoint coordinates):

Track log - autofixes
1. make sure you've got sufficient accuracy (check your constellation/pnr status screen)
2. enable 1sec track logging
3. stay in the same spot for at least a minute (that ought to give you 60 fixes) and record a track log
4. convert your track log into .gpx file and QC it in mapsource -- check for any position jumps and delete them. once you think the fixes are tight, save the .gpx file.
5. load .gpx file to MN DNRGarmin and convert it to a comma delimited .txt file
6. load the .txt file in a spreadsheet application and compute for mean/average of your lats and lons.
7. use the resulting mean lat and lon for your WP coordinate and check against an existing road track

Multiple Waypoints - manual fixes
1. make sure you've got sufficient accuracy (check your constellation/pnr status screen)
2. save several waypoints coordinates in the same spot
3. load your saved waypoints in mapsource and QC your waypoints - check for position jumps. It might be helpful to turn on proximity and use a value similar to the accuracy you had while taking the fixes so your WPs' range ring appear.
4. once you reckon your WPs are tight, save as .gpx
5. follow 5-7 above.

HTH

Still thinking of how to do the same for road tracks. Just to clarify, tracks are meant to be for the road itself right and not for individual lanes (i.e. southbound/north bound) right? If so, I guess, merging several overlaying tracks and m-a-y-b-e doing a least-square smoothing would work...

More on WP details: what's the preferred location for saving waypoints for certain areas like a city/town/barangay(do we include these in the contributions)/other wide areas? For city/town do we need to save the wp at say, Rizal's monument in each town? (under the pretext that KP0 in the Philippines is Rizal's monument in Luneta) or (if they're still there) at the KP0 markers on the roads? For buildings, like malls, I believe WPs near the entrances would suffice right? (unless you plan to walk around the perimeter and do a track log to come up with the mean position)

NB: I prefer DNRGarmin's comma-delimited .txt over Mapsource's tab-delimited .txt mainly because DNRGarmin saves it in degree format (DDD.ddd) whereas Mapsource saves it in degrees/minutes format (DDD/MM.mmm) with no tab between the lat/lon. I'm open to suggestions to do the same in Mapsource. Ofcourse if you can do fast macros for excel (no time for me now) then you can use Mapsource's tab delimited .txt and just extract from there, no need to use DNRGArmin.

[pokz]

More on WP details: what's the preferred location for saving waypoints for certain areas like a city/town/barangay(do we include these in the contributions)/other wide areas? For city/town do we need to save the wp at say, Rizal's monument in each town? (under the pretext that KP0 in the Philippines is Rizal's monument in Luneta) or (if they're still there) at the KP0 markers on the roads? For buildings, like malls, I believe WPs near the entrances would suffice right? (unless you plan to walk around the perimeter and do a track log to come up with the mean position)

For cities and towns, we prefer to locate the waypoints in parks as our common set-up here is that they are usually clustered with the city/town halls and churches. So basically you have the center of the town. For the barangay level, we either have a choice of putting it between the public school or barangay hall. But on one hand, why the need to put a dedicated waypoint for the barangay if you create one for the barangay hall/school? Better to just name them properly (usually public schools are named after the barangay itself) and just put the correct address and we should have no problem finding the place.

Yup, you're right that entrances to buildings and the similar establishments are enough and the more preferred waypoint location.