spots - dump1090 in Python

An implementation to detect and decode Mode-S messages modulated on 1090MHz. The implementation is implemented fully in Python (2.7.9), tested on raspberry pi 2 model B hardware using an RTL-SDR USB dongle.

Try http://www.viltstigen.se/spots/spots for live demo. No guarantee however that it is up and running.

Focus is on using Python idioms and readability, not optimizations. Other implementations in C are likely more efficient.

The implementation is multi-threaded:

• Tuner: sample the signal
• Squitter: decoded messages
• Radar: main application that displays messages
• Basic: fundamentals

It tries to be as complete and accurate as possible but with no guarantees of being correct.

The following message will be decoded:

• Short air-to-air surveillance (Downlink format: 0)
• Surveillance altitude reply (Downlink format: 4)
• Surveillande identity reply (Downlink format: 5)
• Long air-to-air surveillance (Downlink format: 16)
• ADS-B, Extended Squitter (Downlink format:17)
• Comm BDS altitude reply (Downlink format: 20)
• Comm BDS identity reply (Downlink format: 21)

Messages decoded are displayed either in a serialised format on standard output or in a tabular format depending on preference. An inbuilt server is listening on port 5051 (configurable) and enables a client to access decoded messages in json format.

Some statistics is collected, this data is also accessible through the server

Dependencies

Spots uses pyrtlsdr v0.2.0 to read samples. Use the installation description to install.

pyrtlsdr is wrapper for rtlsdr library, so this needs to be installed.

See the following references:

• http://sdr.osmocom.org/trac/wiki/rtl-sdr
• http://zr6aic.blogspot.se/2013/02/setting-up-my-raspberry-pi-as-sdr-server.html

The following worked for me

$ sudo apt-get update
$ sudo apt-get install cmake
$ sudo apt-get install libusb-1.0-0.dev
$ git clone git://git.osmocom.org/rtl-sdr.git
$ cd rtl-sdr/
$ mkdir build
$ cd build
$ cmake ../
$ make
$ sudo make install
$ sudo ldconfig

Edit the blacklist

$ sudo nano /etc/modprobe.d/raspi-blacklist.conf

Add these lines

blacklist dvb_usb_rtl28xxu
blacklist rtl2832
blacklist rtl2830

Finally

$ sudo cp ../rtl-sdr.rules /etc/udev/rules.d/
$ pip install pyrtlsdr
$ sudo shutdown -r 0

References

As the implementation have no access to specifications the following implementations serves as references for spots

• dump1090 by antirez, the original
• dump1090 by Malcolm Robb, a fork of antirez
• dump1090 by flighaware, another fork
• java adsb at OpenSky, a java implementation
• ADS-B decoding guide
• Software Defined Radio, lab at Berkley

Usage

Simply try

$ python radar.py

Configuration options

Configuration for spots is in spots_config.json. Follows json syntax with no error checks so be careful.

• verbose logging (true/false): writes messages to spots logfile
• check crc (true/false): whether to check crc (recommended) or not
• check phase (true/false): simple check if there is a phase shift and correction
• use metric (true/false): show values in metric system or not (altitude and velocity)
• apply bit error correction (true/false): whether to try to correct bit errors or not (CPU demanding if true)
• run as daemon (true/false): if true and read from file is true, do not terminate the program when file read is done
• read from file (true/false): if true, read samples from a file rather than from the USB dongle
• file name (string): if "read from file" is true, this is the file to read from
• use text display (true/false): if true, show data in table format, if false show in serialised way
• max blip ttl (integer or float): how many seconds to keep an identified aircraft in the table display
• user latitude (float): your latitude position (not used)
• user longitude (float): and your longitude (not used)
• log file (string): The name of the log file
• log max bytes (integer): How many bytes to log before the log file is rotated
• log backup count (integer): How many roted log files to keep
• spots server address (localhost or ip-address): the address for the server
• spots server port (5051): the server port
• flight db name (string): name of database file to store flight counts, if the value is "" this function is skipped
• statistics file name (string): name of file to store statistics, which are read during start
• config file (string): name of file where personal email configuration data is stored, if the value is "" no emailing is done
• email recipient (string): name of email recipient, used when rare errors occur for notification. SMTP information is stored in the config file (above)

config file for SMTP info use the following syntax: { "//": "CAUTION: Stored in plain text!", "SMTP_server": "smtp.gmail.com", "SMTP_port": 587, "MAIL_username": "mats.melander@gmail.com", "MAIL_pw": "YourSercretPassword" }

Client/Server

Use nginx as proxy server with the following added to the nginx conf file

location /spots {
try_files $uri $uri/ $uri/index.html $uri.html @spots;
}

location @spots {
    proxy_pass http://rpi3.local:8080;
    proxy_redirect     off;
    proxy_set_header   Host $host;
    proxy_set_header   X-Real-IP $remote_addr;
    proxy_set_header   X-Forwarded-For $proxy_add_x_forwarded_for;
    proxy_set_header   X-Forwarded-Host $server_name;
}

So, nginx will forward any http requests to spots (e.g. http://www.viltstigen.se/spots) to http://rpi3.local:8080 (spots runs on rpi3-node). Restart nginx to capture updates through

$ sudo /etc/init.d/nginx restart
$ /etc/init.d/nginx status

Flask is running using Gunicorn, listening on port 8080, see emitter.py and spots_emitter.conf for details. Use supervisor to control processes running as daemons.

$ sudo ln -s /home/pi/app/spots/spots_emitter.conf /etc/supervisor/conf.d/spots_emitter.conf
$ sudo ln -s /home/pi/app/spots/radar.conf /etc/supervisor/conf.d/radar.conf
$ sudo supervisorctl reread
$ sudo supervisorctl update
$ sudo supervisorctl status

The flask application communicates with the radar application (that listen on port 5051) through a simple text protocol, see files emitter.py and server.py

Using some html, bootstrap css and javascripts (jQuery and Highcharts), see files spots.html and spots.js, it is possible to get this view in a web browser.

What's next?

There are probably inconsistencies, bugs, optimizations, documentation etc etc to make. If you find something, let me know but be aware that this is a leisure thing for me.