Synopsis: A method for using a local reference point to improve GPS accuracy. Implemented on a Android droid x 2 platform.
The Android as a GPS GeoCaching gaming platform has one major short coming. Frequently waypoints are set and when the game is played, the apparent point may be exactly on target or the apparent point can be 10 or 20 feet off or more in some direction. Every geocacher knows this and accepts this as just part of the game. Recently, a solution to this was suggested by an Android game feature I was developing. I wanted the user to be able to play anywhere. This was accomplished by having the user being able to go somewhere in the world and play the game there and the waypoints in the game would be played in relation to there. All the game had to do was calculate the difference between there and where the game begins when played in the normal manner. This logic was implemented and the game(s) can now be played anywhere. The obvious limitation being – If the game depended upon actual knowledge at the real waypoint then the game could not be played.
The waypoints used in Portals were taken one afternoon using a Garmin 76 Csx with an accuracy about 15 to 18 feet. Most of the waypoints were in a small area within 200 feet of one another. The Portals player has to roam around opening virtual portals modeled after real doors. At each portal, a question needs to be answered to complete the step. Nothing unusual about that, most multistage geocaches have the player use some bit of information at a waypoint to help locate the next stage or provide some clue for the final. Much to my surprise the various waypoints were frequently off sometimes 30 or 40 feet. In some cases the virtual location wasn’t in front of the real door but in the middle of the street or in some cases, across the street.
The solution after some slow slow insight: Use the play anwhere logic to adjust the game’s waypoints to allow for today’s GPS apparent location. This is possible because Portals waypoints were a taken in a fairly small area, in a period time of less than 2 hours, and with an accurate GPS. The only thing that had to be done was take a GPS reading at a known waypoint. The GPS reading on any given date depends upon the whim of the GPS gods and the apparent location is sometimes off by many feet. No matter, I only needed today’s apparent location from a fixed point and I had the apparent location for the fixed location. The other waypoints were all in relation to the given point on the measurement day. All I had to do was determine the offset between today and the original measurement day and I could add the offset to all the other points as the game was being played and the waypoints would maintain their relative distance even though the apparent location had shifted.
Field Test Sept 26, 2011 15:00 Field Reference Point Metal square in the center of Needham, Ma town square. N 42.28069 W -071.23651 The readings are taken from the metal plate. The known distance is approx 48 ft from square plate to the “Circle of Peace.” The Garmin 76Csx used the same points both days.
Uncalibrated Droid is + 14 feet. Calibrated Droid is -4 ft
Observed is how far the detected zone was from the expected zone: I walked to the point the calibrated Droid took me and then estimated how far to where I expected to be.
Send comments, suggestions email: firstname.lastname@example.org