SINS/GPS组合导航的直接状态估计方法

Direct State Estimation Method of GPS/SINS Integrated Navigation Vectors

由全球定位系统(GPS)和捷联式惯性导航系统(SINS)构成的组合方式导航是确定载体位速和姿态的常用方法。其滤波模型一般是基于估计误差的微分方程建立,能够直接运用卡尔曼滤波求解的线性状态空间方程,即间接估计。论文提出一种基于直接估计和非线性卡尔曼滤波的GPS/SINS组合导航方法,利用惯导力学编排建立了直接估计状态模型与松组合模型,并在求解雅克比矩阵的基础上采用扩展卡尔曼滤波求解。通过数值仿真对提出模型和算法进行了对比验证,结果表明,直接估计方法相比间接估计方法在恶劣初始误差条件下具有优于后者的性能。

A coupled navigation method composed of a global positioning system(GPS)and a strapdown inertial navigation system(SINS)is now a standard for determining the position,velocity and attitude of vehicles.The filtering model is generally a lin⁃ear state space equation based on the differential equation of the estimation error,which can be directly solved by Kalman filter,namely indirect formulation.This paper proposes a GPS/SINS integrated navigation method based on direct estimation and nonlin⁃ear Kalman filter.A direct estimation state model and a loosely coupled model are established using inertial navigation mechanics,and the extended Kalman is used to solve the Jacobian matrix.Numerical simulation results suggest that the direct estimation method has better performance under severe conditions than the indirect estimation method,but its applicability will be limited for its large amount of calculation,which compromises the applications of direct formulation under precise alignment and good conditions.