BiblioScholar Multiple Model Methods for Cost Function Based Multiple Hypothesis Trackers

To estimate the state of a maneuvering target in clutter, a tracking algorithm must becapable of addressing measurement noise, varying target dynamics, and clutter. Traditionally, Kalman filters have been used to reject measurement noise, and their multiple model form can accurately identify target dynamics. The Multiple Hypothesis Tracker (MHT), a Bayesian solution to the measurement association problem that retains the probability density function of the target state as a mixture of weighted Gaussians, offers the greatest potential for rejecting clutter, especially when based on an advanced mixture reduction algorithm (MRA) such as the Integral Square Error (ISE) cost function.This research seeks to incorporate multiple model filters into an ISE cost-function based MHT to increase the fidelity of target state estimation.