Application of unscented Kalman filter to novel terrain passive integrated navigation system;

Based on combination of sampling importance resampling(SIR) and unscented Kalman filter(UKF),a novel particle filter is proposed,possessing the merits of high utility efficiency of particles in unscented particle filter(UPF) and of simple operation in SIR,and overcoming the drawback of the rate of increase of computational cost being faster than that of state dimension in UPF.

This paper introduces the newly proposed unscented Kalman filter (UKF).

The unscented Kalman filter(UKF) model for the system is built up,and a numerical simulation is performed with the software Matlab.

We aim to eliminate these shortcomings as much as possible with a different and we believe better method by using the unscented Kalman filter(UKF) based SLAM(simultaneous localization and mapping) technique.

Its core consists of:(1) we design a variable structure sliding-mode speed controller and a variable structure sliding -mode current controller to replace the traditional speed PI controller and two current PI controllers;(2) we design the unscented Kalman filter(UKF) observer to estimate direct-axis current,quadrature-axis current,load torque,rotor position and speed simultaneously.

In view of the problem that determination of relative attitude between formation satellites is difficult,a modified Unscented Kalman Filter(UKF)was adopted to design the Filter of system in the paper.

The Standard Kalman Filter can not solve the nonlinear mathematic model,and the Extending Kalman Filter(EKF) will consume much more calculation,so Unscented Kalman Filter(UKF) is adopted in this system.

In view of the influence of non-linear measurement equation on precision and stability of filtering in target tracking application,this paper especially analyzes the basic principle,characteristics and adaptive conditions of model-linearization filtering algorithm,unscented Kalman filtering(UKF)algorithm and particle filtering(PF)algorithm.

In view of control difficulty caused by severe nonlinear performance of brushless DC motor(BLDCM),an observer was designed for estimating the rotor position and velocity of BLDCM by using unscented Kalman filter(UKF) algorithm.

The psi-angle model of nonlinear inertial navigation system(INS) alignment for large misalignment error was discussed,and the principle of the unscented Kalman filter(UKF) was analyzed.

For verifying the performances of extended Kalman filter(EKF) and unscented Kalman filter(UKF) in tightly-coupled integrated navigation system,an INS system error model in the earth-centered earth-fixed(ECEF) frame and a GPS pseudo range/range rate model are proposed to form the system equation and the measurement equation respectively,and then EKF and UKF filtering equations are derived.