Tracking targets in K clutter via particle filter

Abstract

In this paper the radar detection and tracking of targets in sea clutter is examined through the use of particle filter (PF). The clutter is modeled by the K-distribution, which is usually used to model radar sea clutter returns received from small grazing angle or high resolution radars. Two target models are considered, i.e. non-fluctuating target model and Swerling type 1 target model, to capture the effect of radar crosssection changes that a target would present to the radar over time. For the above two target models, the probability density functions of clutter-only and clutter-plus-target case are given, and then the likelihood ratios (LLRs) are derived and used in the filtering process of PF. The performance is examined using various simulations.