I dropped my prototype Pi out into the garden with the GPS installed. It’s running the code from the previous post.
Here’s what I got:
This is much more like what I was expecting. The core is < ±5m from the origin – the top right area is the path as I returned the kit back into the office. However, there’s still something wrong: it’s about 10m north to the back door, and yet it shows 30m on the ‘map’ based upon the change in latitude as I walk north, and the radius of the earth. The scale here seems wrong yet it comes directly from this equirectangular approximation:
dx = (lon - base_lon) * math.cos((lat + base_lat) / 2) * R dy = (lat - base_lat) * R
Throughout this test where the core is stable, the receiver was locked on to 7 – 10 satellites. On the walk back inside, this reduces to 6 or fewer so that may explain it. Next step is to include the accuracy measurement from the receiver to ultimately apply a bias for merging these values with the double integrated accelerometer distances.
P.S. Ignore all the wires, they are for URF and LEDDAR testing; the GPS is connected via a black USB cable.