Multitarget detection and tracking in dynamic quadratic clutter

In previous publications the author introduced CPHD filters designed to detect and track multiple targets in unknown, dynamically changing clutter backgrounds. The first such filters employed Poisson clutter generators and resulted in combinatorially complex algorithms. Subsequent CPHD filters achieved computational tractability by replacing Poisson clutter generators with Bernoulli clutter generators. Because they are statistically first-degree, Bernoulli generators are insufficiently complex to model real-world clutter with high accuracy. This paper describes CPHD filters based on second-degree quadratic clutter generators. CPHD filters based on quadratic generators are combinatorially second-order and therefore more amenable to approximation than those based on Poisson clutter generators. They can also be implemented in exact closed form using beta-Gaussian mixture (BGM) or Dirichlet-Gaussian mixture (DGM) techniques.