KalmanFormer
A Python implementation of a published method that learns the Kalman gain of a Kalman filter directly from data using a transformer, rather than computing it from a hand-derived model of the system.
KalmanFormer implements a technique proposed in a research paper for state estimation, where a transformer model learns to predict the Kalman gain that a traditional Kalman filter would otherwise compute analytically. The stated goal is to work around the limitations of classical Kalman filters when the underlying system is non-linear, letting the learned component adapt where a fixed mathematical model would not.
The package installs from PyPI and is meant to be used as a library within a larger state-estimation pipeline rather than as a standalone application.
It is written in Python and released under the MIT license; the accompanying paper was published in Frontiers in Neurorobotics.