Nonlinear-Filtering with Interacting Multiple-Model Algorithm for Coastal Radar Target Tracking System
Rika Sustika, Joko Suryana
This paper presents a performance evaluation of nonlinear filtering with Interacting Multiple Model (IMM) algorithm for implementation on Indonesian coastal radar target tracking system. On this radar, target motion is modeled using Cartesian coordinate but target position measurements are provided in polar coordinate (range and azimuth). For this implementation, we investigated two types of nonlinear filtering, Converted Measurement Kalman Filter (CMKF) and Unscented Kalman Filter (UKF). IMM algorithm is used to anticipate target motion uncertainty. Many simulations on radar target tracking are developed under assumption that noise characteristic is known. In this paper, the performance of IMM-CMKF and IMM-UKF is evaluated for condition that radar doesn’t know noise characteristic and there is mismatch on noise modeling. Results from simulation show that IMM-CMKF has better performance than IMM-UKF when tracking maneuvering trajectory. Results also show that IMM-CMKF is more robust than IMM-UKF when there is mismatch on noise modeling.