惯导测量发射位置和速度误差的迭代计算方法

Iterative Computation Method of Origin Position and Velocity Error Measured by Inertial Navigation

为了辨识惯性导航测量的发射原点位置和初始速度的误差,利用外弹道起始时刻数据和遥测视加速度,建立前推的非线性微分方程计算发射原点和初始速度。利用迭代计算降低模型误差源对辨识精度的影响,建立了仿真数据;利用仿真数据给出模型的逼近精度,利用蒙特卡洛方法给出模型辨识精度。实验发现,微分方程的起始点应该选择在时间靠前且加速度变化平稳的位置;待辨识原点误差对辨识精度没有影响,辨识精度只与外测起始点误差和遥测视加速度误差相关;如果在遥测上面添加系统性偏差,那么初始点选择越靠前越好;发射方位角偏差不大于0.001rad对原点误差辨识的影响可以忽略。

To recognize launch origin-position and initial velocity error measured by inertial guidance,a backward non-linear differential equation was built to calculate origin position and initial velocity by using exterior measurement start-point and telemetry apparent acceleration.Iterative computation was used to decrease the influence of model error source on error recognition precision.Artificial data was built.Approximation accuracy of this model was obtained by artificial data.The recognition precision of origin error was given by Monte-Carlo method.The simulation result shows that the start-point of differential equation should has stationary acceleration and be close to the front.Origin error to be recognition has no impact on recognition accuracy,and recognition accuracy only depends on exterior measurement start-point error and telemetry apparent acceleration error.If telemetry has systematic error,exterior measurement start-point closing to the front is better.If launch azimuth angle error is less than or equal to 0.001 rad,the influence on model can be ignored.