组合导航初始对准无迹粒子滤波算法

UPF Algorithm to Initial Alignment of Integrated Navigation System

针对GPS/SINS组合导航系统,提出了一种初始对准方案,依靠GPS测量信息进行速度匹配,完成初始对准。粒子滤波虽然不受非高斯、非线性限制,但存在粒子数匮乏、实时性差等问题。该方案采用粒子滤波的改进方法——无迹粒子滤波,来解决对准过程中的非线性问题。仿真结果表明:该方案的对准精度(1σ)可以达到东向失准角误差为5′,北向失准角误差为2′,方位失准角误差为6′。

A method of initial alignment is presented for the integrated system of GPS with a strapdown inertial navigation system （SINS）. It uses the GPS measurements for the velocity matching to achieve initial alignment. Particle filter can handle nonlinear and non-Gaussian problems without any limits, but it still has some existing problems,such as sample impoverishment and bad real-time performance. To solve nonlinear problem caused by large attitude errors, the unscented particle filtering （UPF）, a improved particle filter algorithm, is used. Simulation result shows that the alignment accuracy is 5′for east direction maladjustment angle error, 2′ north direction maladjustment angle error,and 6′for azimuth maladjustment angle error.