Particle Filter for INS In-Motion Alignment

This paper presents a nonlinear dynamical model for the in-motion alignment of the inertial navigation system (INS) in the case that the observation variable is the velocity information. It allows the initial misalignment uncertainty. Therefore, this model is also suitable for the transfer alignment based on the velocity matching algorithm. Then the Gaussian particle filter (GPF) is analyzed and utilized for the nonlinear filtering. Under the turn maneuver, this paper analyzes and compares the misalignment estimation error and convergence rate of GPF with the unscented Kalman filter (UKF) when the initial misalignment is uncertain. The results of the simulation show that GPF is robust for the initial misalignment, but UKF is influenced badly. When the misalignment is large, the convergence rate of UKF is very slow, but GPF is not. Therefore, GPF is suitable for INS in-motion alignment