A Technique for Deriving Multitarget Intensity Filters Using Ordinary Derivatives.
Roy L. Streit · 2014
In multitarget tracking problems based on finite point process models of targets and measurements, it is known that the distri-bution of the Bayes posterior point process is a ratio of functional derivatives of a joint probability generating functional. It is shown here that these functional derivatives can be found by evaluating or-dinary derivatives. The method is exact, not approximate. Several examples are presented, including multisensor target tracking and extended-target tracking. The method is well suited to the needs of particle filter implementations.