Tuesday, January 26, 2010

Receiver performance metrics tuned for APRS

I've improved the "first receiver" detection algorithm a bit. It's now a bit tougher and shouldn't get as many false positives. For example, the packet:

SRCCALL>DST,DIGI1*,WIDE2*,qAR,IGATE:data

is no longer considered to be heard first by DIGI1. Often in these cases the packet was transmitted as SRCCALL>DST,WIDE2-2, and then digipeated by a non new-N digi which changes that to WIDE2-1, and then by DIGI1 which changes that to DIGI1*,WIDE2*. And DIGI1 gets a too big range estimate for that.

On the other hand, the following packet:

SRCCALL>DST,DIGI*,WIDE2-1,qAR,IGATE:data

will give the receiving point to DIGI1. That packet is probably an original WIDE2-2 packet with a single used hop, or a WIDE1-1,WIDE2-1 packet which has been digipeated by a callsign-substitution fill-in DIGI1.

Because this changes the receiver performance metrics heavily (in the good direction), I've deleted the old "heard" databases and receiver histogram data of APRS igates and digipeaters. Sorry for that. They'll grow back soon.

1 comment:

Daniel said...

I have a quick question on this.

My iGate is the only one in my home town (and some of the next door towns) and as such almost always gets the first packet because it is the only station to receive the packets from nearby mobile stations.

I have noticed that as a result all my packets get put into the 'bogus, unreliable' graph.

Should I be worried about this?

http://aprs.fi/info/graphs/G7JVN-5