Techniques for birth-particle placement in the probability hypothesis density particle filter applied to passive radar

An initial investigation found Ronald Mahler´s probability hypothesis density (PHD) filter to be a promising tool for the passive coherent location (PCL) of targets observed via multiple bistatic radar measurements. Two significant improvements to the particle-filter implementation of the PHD-based multitarget, multisensor tracker for PCL are introduced. The first is a simple, yet novel, peak-extraction technique that exploits the integral property of the PHD and provides a faster and more accurate method over currently used techniques to extract the target states from the PHD. The second is an improved method of placing birth particles, whereby a precomputed range-variance-based grid and a least-squares iterative technique is used to place birth particles at the bistatic range and Doppler observation intersections. This leads to greater reliability in detecting targets and a significant reduction in the number of particles required for tracking.