Stands for Bernoulli Random Finite Set. It is a rust implementation with python bindings of Ristic et.al. tutorial on Bernoulli filters.

A development environment may be activated using

nix develop .

followed by

maturin develop

This will construct a virtual environment in your current directory at .venv. Using maturin, development loop is quick. Change some rust or python code, rerun maturin develop and you should have an incremental debug build of the project. Running maturin develop --release will give that same environment, but the rust library is compiled in release mode, i.e., speed!

After running maturin develop, you may see a target tracking example using the particle-filter-based bernoulli random finite set on detection measurements using

python -m berrfs.example

Entering a python shell should now give you access to the python module. An example instance of a BerRFS-struct (the Gaussian Sum Filter implementation) can be constructed using

# Python 
import berrfs
b = berrfs.example_setup()

on which you can perform prediction steps

b.predict()

and measurement updates

b.update([np.array([1.]), np.array([2.])])

Note that the measurement update performs a prediction step before applying the information from the measurements.

A shell with a python environment with this module compiled in release mode is constructed through instead by running

nix develop .#pyrelease

Now, there is no need to use maturin. The library should already be installed in your python environment. The build is likely to take a few minutes.