Scholarlink

<p>Initial alignment is the prerequisite and foundation for the normal operation of the Strapdown Inertial Navigation System (SINS). Kalman filtering and various evolution algorithms are commonly employed tools for achieving initial alignment. To address the high computational load issue of the Cubature Kalman Filter in the large azimuth misalignment angle initial alignment of SINS and the odometer increment errors caused by pulse count truncation, Strong Tracking Dimensionality-Reduction CKF (STDR-CKF) algorithm is proposed. By performing differential calculations between the SINS position information and the odometer-based navigation solution within a fixed mileage segment, and using the difference for feedback correction, the proposed method effectively suppresses the odometer increment errors. Furthermore, the dimensionality of state variable in proposed algorithm is separated into 6 nonlinear states and 12 linear states based on the “linear-nonlinear” state separation framework, thus the CKF sampling points number is greatly reduced, which significantly reducing the computational load. Additionally, the Strong Tracking Filter (STF) diminishing factor is incorporated into the dimensionality-reduced CKF(DR-CKF) to achieve real-time dynamic feedback correction of system model parameters, thus enhancing the robustness of the filter. Experimental results demonstrate that, compared to the conventional CKF algorithm, the proposed method ensures significantly stronger filtering stability, higher filtering accuracy, and shorter initial alignment time under large azimuth misalignment conditions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Strong tracking dimensionality-reduction CKF algorithm for SINS initial alignment under large misalignment angle

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月3日

查看全文
  • Yong Yang,
  • Donglin Li,
  • Yujie Fu,
  • Kaijun Xu

摘要

Initial alignment is the prerequisite and foundation for the normal operation of the Strapdown Inertial Navigation System (SINS). Kalman filtering and various evolution algorithms are commonly employed tools for achieving initial alignment. To address the high computational load issue of the Cubature Kalman Filter in the large azimuth misalignment angle initial alignment of SINS and the odometer increment errors caused by pulse count truncation, Strong Tracking Dimensionality-Reduction CKF (STDR-CKF) algorithm is proposed. By performing differential calculations between the SINS position information and the odometer-based navigation solution within a fixed mileage segment, and using the difference for feedback correction, the proposed method effectively suppresses the odometer increment errors. Furthermore, the dimensionality of state variable in proposed algorithm is separated into 6 nonlinear states and 12 linear states based on the “linear-nonlinear” state separation framework, thus the CKF sampling points number is greatly reduced, which significantly reducing the computational load. Additionally, the Strong Tracking Filter (STF) diminishing factor is incorporated into the dimensionality-reduced CKF(DR-CKF) to achieve real-time dynamic feedback correction of system model parameters, thus enhancing the robustness of the filter. Experimental results demonstrate that, compared to the conventional CKF algorithm, the proposed method ensures significantly stronger filtering stability, higher filtering accuracy, and shorter initial alignment time under large azimuth misalignment conditions.