Overcoming Unknown Measurement Noise Powers in Multistatic Target Localization: A Cyclic Minimization and Joint Estimation Algorithm

Abstract

This paper investigates the issue of multistatic target localization using measurements including angle of arrival (AOA), time delay (TD), and Doppler shift (DS). We delve into a practically driven nonideal localization scenario where the measurement noise powers remain unknown. An algorithm that jointly estimates target position-velocity and measurement noise powers is proposed. Initially, an optimization model for the joint estimation is developed following the maximum likelihood estimation criterion. Subsequently, we cyclically minimize the optimization model to yield estimates for target position-velocity and measurement noise powers. The Cramer-Rao lower bound (CRLB) for this joint estimation is also derived. Contrary to existing algorithms, our proposed method eliminates the need for prior knowledge of measurement noise powers, simultaneously estimating the target position-velocity and measurement noise powers. Simulation results indicate superior localization accuracy with our algorithm, particularly in scenarios with unknown measurement noise powers. Furthermore, at moderate noise levels, the algorithm's estimation accuracy for target position-velocity and measurement noise powers meets the CRLB.