基于ASYM的激光陀螺捷联惯性系统温度误差模型辨识

Asymptotical identification for temperature error model of laser SINS

为减小温度对导航精度的影响,实现系统级的温度补偿,在实验中采用静态条件下的标定方法;基于激光陀螺捷联惯性系统的误差模型方程,用广义逆算法顺利分离求得陀螺各零偏及标度因数值;根据以往温度误差模型的结构特点,运用渐近辨识方法(ASYM)中的最终输出误差准则(FOE)对温度误差模型中非线性部分的阶次进行准确的计算,确定了合理的温度误差模型结构。为了解决用最小二乘法辨识模型结构的系数时,信息矩阵求逆容易溢出的问题,采用了自适应的岭估计算法确定陀螺零偏温度误差模型的系数,实现了系统级的温度误差建模。所得到的温度误差模型补偿效果比定阶前明显提高。

To improve the accuracy of Laser Strapdown Inertial Navigation System（LSINS）, the experiment was carried out under static state, and the error model of LSINS was derived to calibrate the gyro＇s bias and scaling factor using generalized inverse algorithm. Based on the past temperature error models and using Final Output Error criteria of Asymptotical Identification theory, the rank of nonlinear part of temperature error models was calculated, and the best structure of temperature error compensation model was determined. The paper applies Self-adapt Ridge Regression estimator to calculate the coefficients of the temperature error models to avoid overflowing of least square algorithm, and establishes a system-level temperature error compensation. The experiment results show that the method is applicable, and the acquisition precision of the system has been improved obviously.