基于未补偿偏置滤波器的惯导动态初始对准

Initial alignment based on uncompensated bias filter for inertial navigation systems under dynamic conditions

动态初始对准是惯性导航系统(惯导)工程应用的重要功能之一。针对动态环境下随机干扰和弱可观惯性仪表误差导致对准滤波器性能下降的问题,采用未补偿偏置滤波器实现惯导系统的初始对准。给出了带高度阻尼的惯导水平通道误差模型,根据最小二乘估计原理定量分析了陀螺漂移误差对降维滤波器精度的影响,进而推导出带偏置结构的对准误差模型,设计出基于水平位置误差观测的7维未补偿偏置Kalman滤波器。动态试验结果表明,未补偿偏置滤波器能有效提高惯导动态初始对准性能,仅需40 min对准精度即达标,比直接降维滤波器快一倍以上,具有较强的工程应用价值。

The initial alignment of inertial navigation system（INS） under dynamic conditions plays a key role in the engineering applications. In this paper, an uncompensated bias filter is adopted in INS initial alignment to overcom the problem of alignment filter＇s performance degradation due to random disturbances and inertial instrument errors of weak observability under dynamic conditions. An INS horizontal error model with altitude damped is presented, and the effect of gyro drift errors on accuracy of reduced alignment filter is analyzed quantitatively according to the principle of least squares estimation. Consequently, an alignment error model with bias structure is derived and a seven-dimension Kalman filter with uncompensated bias based on observations of horizontal position errors is designed. Results of dynamic experiments show that performance of INS initial alignment under dynamic conditions can be improved effectively by using the Kalman filter with uncompensated bias, in which the expected alignment accuracy is achived within 40 min, and the speed is at least one fold faster than that of direct dimension-reduction filters, showing great value in engineering applications.