Target localisation for multi-static radar using ellipse intersections. Not a dating app.
3lips-contrail is one part of a multi-component passive-radar project:
- blah2-contrail — the passive radar nodes that produce bistatic range (delay) and Doppler detections, which 3lips-contrail associates and localises.
- adsb2dd-contrail — generates delay-Doppler data from ADS-B data so ADS-B targets can be plotted on the delay-Doppler map in blah2-contrail.
- Provides a JSON API for target geolocation.
- Uses a MapLibre GL JS web front-end to visualise data.
- Ability to compare a number of algorithms for target localisation.
- Install docker and docker-compose on the host machine.
- Clone this repository to some directory.
- Edit the config/config.yml file for scenario.
- Run the docker compose command.
sudo git clone http://github.com/jomosh/3lips-contrail /opt/3lips-contrail
cd /opt/3lips-contrail
sudo docker compose up -d —buildThe API front-end is available at http://localhost:49156.
The association uses the following algorithm:
- ADS-B associator will associate the closest target within some delay and Doppler around the truth.
The target localisation uses 1 of the following algorithms:
-
Ellipse parametric samples an ellipse (2D) at 0 altitude. Find intersections between 3 or more ellipses such that the distance to each point is under some threshold.
-
Ellipsoid parametric samples an ellipsoid (3D). Find intersections between 3 or more ellipsoids such that the distance to each point is under some threshold.
-
Spherical intersection a closed form solution which applies when a common receiver or transmitter are used. As described in Two Methods for Target Localization in Multistatic Passive Radar.
The system architecture is as follows:
- The API server and HTML pages are served through a Flask in Python.
- An initial API request with a new set of parameters (algorithms or radar nodes) will add these parameters to a common processing loop. This is so fair comparisons can be made on the same input data.
- A set of API parameters will continue to be processed unless there is no API call in some specified time - see main.py to update. This allows the latest geolocation to be provided, rather than adding to the processing loop and waiting for the update from the next time increment.
- Implement an association algorithm that is not reliant on ADS-B truth.
- Choose to use detection or track data from each radar.
- Long term plots to show metrics such as 2D location accuracy to ADS-B, number of aircraft tracked, etc.
- Scale number of samples in ellipse/ellipsoid to size of shape.