Distributed peer-to-peer multitarget tracking with association-based track fusion

Giorgio Battistelli, Luigi Chisci, Claudio Fantacci, Nicola Forti, Alfonso Farina, Antonio Graziano

Abstract: The paper addresses distributed multitarget tracking over a peer-to-peer network of sensors with limited target observability. It is shown how to extend the Cheap Joint Probabilistic Data Association (CJPDA) filter to this setting by devising suitable distributed, consensus-based, procedures for track initiation and track fusion. The resulting, association-based, Distributed-CJPDA (D-CJPDA) tracker is then compared to a previously proposed random set approach for the same problem based on the Gaussian Mixture Cardinalized Probability Hypothesis Density (CPHD) filter.