Huang Yongjiang, Liu Xixiang, Zhang Yupeng, Zhao Miaomiao, Yan Jie
School of Instrument Science and Engineering, Southeast University, Nanjing 210096, China.
Beijing Institute of Electronic System Engineering, Beijing 100854, China.
Rev Sci Instrum. 2020 Dec 1;91(12):125102. doi: 10.1063/5.0029584.
The initial alignment method, including the identification of inertial device error parameters, has always been a key issue in an inertial navigation system (INS). This study focuses on the error caused by the random noise of inertial devices that can be compensated by the reconstruction of gravitational apparent motion in an inertial frame under the condition of swinging motion. Attitude angles and accelerometer bias can also be estimated. However, the analysis and simulation results indicate that the existing methods cannot estimate the gyroscope bias. The accelerometer and the gyroscope bias will change over a long time, which will lead to long-term parameter identification accuracy decline or even failure. In this paper, a parameter identification algorithm based on Newton iterative optimization combined with a window loop calculation is designed to solve these problems. Simulation and turntable tests indicate that the proposed new algorithm can fulfill the initial alignment of strapdown INS under the swinging condition and estimate accelerometer bias effectively. Moreover, the new algorithm improves data utilization, which also has better time sensitivity, and the calculated alignment errors can nearly approach zero.
初始对准方法，包括惯性器件误差参数的辨识，一直是惯性导航系统（INS）中的关键问题。本研究聚焦于惯性器件随机噪声所引起的误差，在摆动运动条件下，该误差可通过惯性系中重力视运动的重构来补偿。姿态角和加速度计偏置也能够被估计。然而，分析和仿真结果表明，现有方法无法估计陀螺仪偏置。加速度计和陀螺仪偏置会随时间变化，这将导致长期参数辨识精度下降甚至失效。本文设计了一种基于牛顿迭代优化并结合窗口循环计算的参数辨识算法来解决这些问题。仿真和转台测试表明，所提出的新算法能够在摆动条件下完成捷联惯性导航系统的初始对准，并有效估计加速度计偏置。此外，新算法提高了数据利用率，具有更好的时间敏感性，且计算得到的对准误差几乎可趋近于零。
