Multitarget tracking via joint PHD filtering and multiscan association
Francesco Papi, Giorgio Battistelli, Luigi Chisci, Stefano Morrocchi., Alfonso Maria Farina, A. Graziano · Florence Research (University of Florence) · 2009
A PHD (probability hypothesis density) filter and multiscan association are combined in a feedback fashion in order to provide robust and efficient multitarget tracking. The resulting hybrid tracker, thanks to the feedback connection, provides remarkable performance improvements with respect to both an open-loop PHD filter with estimate extraction via clustering and a traditional tracker equipped with a track formation logic.