ADS-B and Mode S decoding ✈️

[xoolive]

Let's use this thread to discuss about decoding capabilities for the raw data obtained from the Impala shell or directly from your own ADS-B receiver.

pyModeS may be your library of choice: that's why it is directly integrated in traffic.

[ElSabio97]

Hi @xoolive and @junzis !

I wanted to ask you for some extra-advice.

What would be your approach to getting RA (resolution advisories) events and then plotting some of them?

At this moment I do the following:

1. I edited the raw query type from @junzis pyopensky OpenskyImpalaWrapper to download from the acas_data4 table instead of the rollcall_replies_data4.
2. I download aproximatelly 200.000 acas_data4 messages.
3. Then the resulting data frame's raw messages column is processed to get which ones are BDS 30. This is done to get the threat's icao24.
4. Now, with the few raw messages that are BDS30 and also give the threat's icao24, I manually choose an icao24 and the threat's icao24.
5. I write both icao24 addreses in the icao24 filter argument of a history query using a list. Also start/stop datetime filter arguments are set 10 minutes before and after the RA datetime.
6. I export the query dataframe to a pickle file.
7. With that done, I go to the traffic GUI, select the pickle file and then plot the trajectories, altitudes and vertical rates of the two planes.

It's a long process I know, but this is the best I have done so far.

The thing is I recently learned while reading through the ICAO DOC 9871 that there are other ways to get the RA information. As far as I'm concerned: BDS10, BDS30, BDS61, BDS65 and on the ADSB's TypeCode 28 SubType 2.

And now I guess the best way to get this information is by using the rawdata or extended query not by using the acas_data4 table because it won't be there since I believe it only contains DF16 messages. And then using pyModeS functions to decode the messages.

But as you know, raw messages downloads take forever, even if there is a bound for an airport or shape. I have been searching and reading about how to minimize the download time with no luck.

Sorry for the longest comment on github history.

[junzis]

Unfortunately, I think there is no faster solution. Since Mode S information is not decoded by OpenSky, you have to download the raw messages and decoded them locally. This process always takes a very long time. You have to be selective in terms of the time scale for your study.

[ElSabio97]

Thank you for your reply!

I was wondering how @xoolive & Co. got it on this article.

I suppose the download process took days. And the funny thing is that I could get my hands on the .csv used on that article (If you read this, thanks again Martin), but it only weighs 45MB!

[xoolive]

For the DASC paper, the decoding was done by Matthias with the java parser for the paper, directly from the raw data feed.

[stinerk]

I also want to get RA data and do the work you did. I don't know what to do because I don't have knowledge like decoding. Could you provide me with your modified raw, pyopensky, and openskyimplawrapper? Also, I would appreciate it if you could tell me how I can demonstrate and view the materials I received like that. I really want data, but I don't know how to look at them. I'm desperate for your help