Extended least square unbiased FIR filter for target tracking using the constant velocity motion model

Abstract

This paper proposes a new nonlinear state estimator that has a finite impulse response (FIR) structure. The proposed state estimator is called the extended least square unbiased FIR filter (ELSUFF) because it is derived using a least square criterion and has an unbiasedness property. The ELSUFF is a special FIR filter designed for the constant velocity motion model and does not require noise information, such as covariance of Gaussian noise. In situations where noise information is highly uncertain, the ELSUFF can provide consistent performance, while existing nonlinear state estimators, such as the extended Kalman filter (EKF) and the particle filter (PF), often exhibit degraded performance under the same condition. Through simulations, we demonstrate the robustness of the ELSUFF against noise model uncertainty.