Apart of the linear Kalman filter, SkyRadar also presents an implementation of the interacting multiple model (iMM) filter for object tracking.
The main difference between the linear Kalman filter, is the extension of object prediction and correction of the object positions also on the lateral positions. The linear Kalman filter is mainly taking as track updates the longitudinal object positions, where with the iMM also the track updates on the lateral positions are considered, and in this way more robustness on object tracking is added.
As main future work on this feature, we will add also the extended Kalman filter for non-linear environments, alpha-beta tracker, and a combination of all these tracking filter based on deep learning concepts.
In the following figure, the difference between linear Kalman filter and iMMTracking for the same scene is shown in PPI scope. As expected, the object movement tracks are better and clearer detected/estimated.
The Kalman filter is part of FreeScopes ATC II. It integrates into a suite of modular radars and modules for digital signal processing. Talk to us to see how it fits into your aviation or defense academy or your university.
