Abstract: This paper establishes the model of one-step randomly delayed observations, develops Unscented Filtering(UF) algorithm. The nonlinear alignment model of SINS with large heading misalignment angle on moving bases is established and the Extended Filtering(EF) algorithm and UF algorithm are applied and simulated for initial alignment of SINS with different delay probabilities respectively. Simulation results show that the UF algorithm has faster convergent rate and better alignment precision than EF algorithm.

Key words: randomly delayed observations, Extended Filtering(EK), Unscented Filtering(UF), Strapdown Inertial Navigation System(SINS), initial alignment

摘要： 建立观测信号随机一步延时模型，在此基础上给出Unscented滤波(UF)算法。建立捷联惯导系统(SINS)动基座大方位失准角下的对准模型，将扩展滤波(EF)算法和UF算法应用于SINS动基座初始对准中，分别进行不同延时概率条件下的实验仿真，结果表明，UF算法相对于EF算法具有更高的对准精度和更快的收敛速度。

关键词: 观测值随机延时, 扩展滤波, Unscented滤波, 捷联惯导系统, 初始对准

CLC Number: 

• U666.1