Multitarget tracking using the Joint Multitrack Probability Density
Ángel F. García‐Fernández, Jesús Grajal · International Conference on Information Fusion · 2009
This paper addresses the problem of detecting and tracking multiple targets in a Bayesian framework. First, we introduce the definition of Joint MultitracK Probability Density (JMKPD) which is the probability of having a certain number of tracks, each one clearly identified with an ID number, and a kinematic state. We develop the a priori model needed to solve the Bayesian problem and a particle filter implementation with two layers, one that deals with the false alarms and track initiation, and another that deals with track maintenance and track end.