GPS accuracy of osm2go on the N900 while mapping house outlines

A mapper who tried to map houses with osm2go (and who politely offered the above screenshot) asked whether the GPSr’s accuracy of the N900 was poor. The above screenshot suggests it is. For a couple of reasons, I doubt that the result would be (much) different using another device:

  • GPSrs are usually optimized to provide more accurate positioning data during movement. While moving slowly (e.g. as a pedestrian), you will see worse results regardless of the receiver used.
  • GPS reception is poor near walls. This behaviour also applies to devices like the WBT-201 or Garmin’s GPSMap. If possible, try to use a bike. Ride straight intersecting lines passing several buildings in one go instead of moving around each individual building.
  • Mapping house numbers in cities by foot usually requires some guesswork when entering the data after the trip. The set waypoints usually are offset by several meters.

I’m using the N900 for over a year now, and I’m very content with it, as I was aware of the aforementioned issues. It is a bit inaccurate while hiking, but on the bicycle it’s almost perfect.

2 thoughts on “GPS accuracy of osm2go on the N900 while mapping house outlines

  1. joachimp

    If you can get access to the raw nmea sentences or have the dilution of precision data of your GPS lock and/or access to the satellite count you can use gpsbabel to filter out possible bad data points. I use this to dramatically improve the quality of the tracks from the source nmea data: gpsbabel -i nmea -f “$file” -x discard,hdop=2,vdop=2,sat=7 -o gpx -F “$file.gpx”

    The last six months(24/7 tracking) of my time in the local city:

  2. LeeOnCandler

    I’m willing to learn here, but if GPSs are “optimized to provide more accurate positioning data during movement” and GPS satellites are moving at about 8,700 mi/hr, how does +/- 50 mi/hr make much difference?

Comments are closed.