摘要
在捷联惯导初始对准优化的研究中,由于存在非线性,很难描述误差特性。为提高捷联惯导系统初始对准的精度,缩短对准时间,根据欧拉角误差的传播规律,推导了惯导系统角度和速度误差微分方程,建立了非线性的对准模型。非线性模型的计算一直是难点问题,一般采用无迹卡尔曼滤波(UKF),为减小计算量,采用了超球面采样点变换(SSUT)的采样策略,结合线性的量测方程,对滤波算法进行了改造和简化。根据全局可观性理论,为提高系统观测度,以二位置静态对准为条件,并对提出的方法进行了仿真和实验验证。结果表明,模型具有很高的精度,并能有效估计惯性器件误差。改进方法对大失准角下的系统初始对准及惯性器件测漂具有较好的参考价值。
To improve the accuracy of strapdown inertial navigation system initial alignment and to shorten the alignment time, attitude and velocity error differential equations and a nonlinear alignment model were built based on the propagation rule of Euler angle error. The Unscented Kalman Filter is often applied with nonlinear model whose calculation is always a difficult problem. To reduce the calculation amount, the sampling strategy of Spherical Simplex Unscented Transformation was adopted to simplify the filter algorism combined with linear observation equation. To improve the system observability, simulation analysis and experiment validation were conducted in case of double-po- sition static alignment according to global observability theory. Results show that the model is still precise under large Euler angle errors and the device error can be effectively estimated. The method proposed in this paper has reference value to system initial alignment and device error estimation under the condition of large misalignment angles.
出处
《计算机仿真》
CSCD
北大核心
2013年第5期54-58,共5页
Computer Simulation
