Bearings-only tracking of manoeuvring targets using particle filters (Q1773810)

The paper studies the application of particle filters (PF) to the problem of bearings-only tracking of manoeuvering targets and compares the results of the proposed techniques with those of traditional interacting multiple model (IMM)-based filters. The model of the target dynamics is described in the framework of a jump Markov system (JMS), namely three switching dynamics models governed by a Markov chain. The filters then provide an estimate of the target state given a sequence of bearing measurements and the particular JMS framework. Three different PFs are examined in the paper: (i) the multiple model PF (MMPF), (ii) the auxiliary MMPF (AUX-MMPF) and (iii) the JMS-PF. Their performance is compared with two conventional estimators: (i) the IMM extended Kalman filter (IMM-EKF) and (ii) the IMM unscented Kalman filter (IMM-UKF). The performance is analyzed by Monte Carlo (MC) simulations and compared with the theoretical Cramér-Rao lower bound (CRLB). The PFs are compared with the traditional IMM filters in three different contexts: (i) single-sensor case, (ii) multisensor case and (iii) tracking with hard constraints. The superiority of the PF trackers is claimed by the authors on the basis of simulation examples.