Flying today: GPS track accuracy

No hacking today, went flying instead. Ended up being a passenger the whole 3 hour flight, crappy IFR weather. But as always I captured the flight on my little AMOD GPS tracker and mapped the track through my custom GPS track mapper project. The result is quite surprising; not only is data missing, but the track is wrong! In particular when we left Modesto we did not fly south, we flew more or less west, and the GPS recorded the wrong position. What happened?

My GPS tracker logs NMEA sentences once a second and records all sorts of neat data. Unfortunately I’m not graphing it yet, so I turned instead to GPS visualizer to visualize some auxilliary data: specifically the # of satellites we had for the fix as well as the horizontal dilution of precision, a measure of fix accuracy.

As you can see in the graph, the HDOP spiked really high for about 10 minutes around 23:20, or exactly when my position data was bad and then disappeared. I’d naively assumed the GPS tracker wouldn’t even record data if it didn’t have a good fix. Apparently it does. The Wikipedia article suggests I should be ignoring fixes with HDOP > 10 or so. My tracker also records vertical DOP; in general GPS isn’t as good at vertical position as horizontal, so that data may be even worse.

One other graph: # of satellites as a function of HDOP. My GPS apparently needs to see 5 satellites to give an accurate position. No idea if that’s true for all GPS chipsets (I think 4 is minimal), but this tracker uses the decent SiFR III chipset so I imagine it’s doing about as good as it can.

I’ve got a long list of things to do with my GPS track mapping program. I hope to make it a web site for others to use some day.