一种基于因果关系的惯性制导工具误差系数分离方法

A Method for Separating Error Coefficients of Inertial Guidance Tools Based on Causality

由于惯性测量系统误差模型的环境函数矩阵存在复共线性问题,所以无法采用最小二乘方法求解出误差系数。针对该问题,首先给出了几种传统的工具误差系数分离方法并进行了验证,但都无法准确估计结果。然后,提出了一种基于因果关系的惯性制导工具误差系数分离方法,引入相关系数对相互关联的系数进行重新整合,构造出新的误差模型。最后,进行了方法验证,不仅大幅度降低了结构维数,还克服了传统方法不能精确求解的缺点,有助于提升惯性制导精度。

Due to the complex collinearity of the environment function matrix of inertial measurement system error model,the least square method can not be used to solve the error coefficient.To solve this problem,several traditional separation methods of tool error coefficients are given and verified firstly,but they can not accurately estimate the results.Then,an error coefficient separation method of inertial guidance tools based on causality is proposed.The correlation coefficient is introduced to reintegrate the interrelated coefficients,and a new error model is constructed.Finally,the method is verified,which not only greatly reduces the structural dimension,but also overcomes the disadvantage that the traditional method can not solve accurately,which is helpful to improve the accuracy of inertial guidance.