Enhanced strategy to sample newborn targets within nonthresholded measurements

Recently, Random finite set theory has attracted researchers´ interest in the field of multitarget tracking time varying number of targets. Its main drawback is it is not essentially formulated for nonthresholded measurements. This paper examines the problem of multitarget tracking with time varying number of targets dealing with raw and nonthresholded measurements. Recursive equations for updating the joint multi target state posterior density are approximated by a new enhanced particle filter that includes effective strategies to tackle the challenges of effective track initialization and deletion with limited resources, as well as doing the data association step implicitly. Simulations prove efficiency of the suggested strategies under different conditions.