A Probabilistic Strongest Neighbor Filter Algorithm for <i>m</i> Validated Measurements

Abstract

A new form of the probabilistically strongest neighbor filter (PSNF) algorithm taking into account the number of validated measurements is proposed. The probabilistic nature of the strongest neighbor (SN) measurement in a cluttered environment is shown to be varied with respect to the number of validated measurements. Incorporating the number of validated measurements into design of the PSNF produces a consistent and cost effective data association method. Simulation studies show that the new filter is less sensitive to the unknown spatial clutter density and is more reliable for practical target tracking in nonhomogeneous clutter than the existing PSNF. It has similar performances to the probabilistic data association filter amplitude information (PDAF-AI) with much less computational complexities.