Research Of Application Of UKF In Integrated Navigation System

Generally the state-equations we established in the actual system are nonlinear. But the traditional Extended Kalman Filter we used makes the nonlinear system lined, so it is suboptimum filter. In this paper we used the Unscented Kalman Filter (UKF) method. It uses the nonlinear model directly, so it can avoid the nonlinear error, and improve the filter precision. Besides the computation of the UKF is the same order as that of the EKF.Firstly, the paper introduced the basic theory and characteristics of the Strapdown Inertial Navigation System (SINS), the GPS, and the DVL system. Because of the difference between them we can use these systems together. And it established the nonlinear error model of the SINS.Secondly, it introduced the basic theory of the Kalman Filter of the linear system, and the EKF of the nonlinear system. And gived their arithmetics.Thirdly, the paper deeply researched the UKF method. And introduced a new UT dots which extended the UKF. Researched the UKF in the mixed system. For the system composed of linear state-equation and nonlinear observation-equation or nonlinear state-equation and linear observation-equation, we can use SUKF to simply the arithmetic farther.Finally, we done some emulation by some vehicle test data, and validated the superior performance of the UKF.