一种组合导航数据的自适应卡尔曼滤波方法、装置和系统
By detecting the innovation sign and deviation magnitude in the observation vector of the integrated navigation data and adjusting the process noise variance Q matrix, the problem of Kalman gain error affecting the convergence speed of navigation data filtering is solved, and high-precision, high-dynamic-characteristic navigation applications are realized.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HUAZHONG UNIV OF SCI & TECH
- Filing Date
- 2025-10-21
- Publication Date
- 2026-07-17
AI Technical Summary
The deterministic error of the Kalman gain affects the convergence speed of navigation data filtering calculations, which is detrimental to navigation applications requiring high precision and high dynamic characteristics.
By detecting the innovation sign and deviation magnitude of each dimension in the observation vector of the integrated navigation data, it is determined whether there is a deterministic deviation, and the process noise variance Q is adjusted by trial and error to replace the Q matrix in the standard Kalman filter method for filtering.
It improves the robustness and adaptability of the Kalman filter algorithm, enabling it to quickly enter the unbiased estimation state and enhancing the accuracy and convergence speed of navigation data.
Smart Images
Figure CN121461930B_ABST