The accuracy of attitude estimation algorithms is significantly affected by the dynamic maneuvers of carriers. This paper introduces a novel Quaternion Descriptor Kalman Filter (QDKF) tailored to enhance estimation accuracy. The QDKF model integrates attitude quaternions and carrier accelerations, with the latter treated as a stochastic process. Measurements are derived from gyroscope, accelerometer, and magnetometer outputs. Unlike previous approaches that relied on weighted least square methods, this study employs a Kalman filter framework for attitude and acceleration estimation. A key innovation in the QDKF is the reduction of the state variable dimension from 8 to 7 by refining the acceleration equation. Additionally, it replaces the weighted matrix used in the Quaternion Descriptor Filter (QDF) with system and measurement noise covariance matrices, thereby eliminating the need to compute matrix inversions. By removing the historical term from the acceleration equation, the QDKF also eliminates historical error, streamlining the structure of the filter. Evaluations demonstrate that the QDKF achieves an average improvement of 52.8% in attitude estimation accuracy and 48.1% in position estimation accuracy compared to the QDF.
Attitude Estimation Algorithm Based on Quaternion Descriptor Kalman Filter
J Astronaut Sci
2025-04-11
Article (Journal)
Electronic Resource
English
Attitude Estimation Algorithm Based on Quaternion Descriptor Kalman Filter
Springer Verlag | 2025
|Attitude estimation based on quaternion square-root cubature Kalman filter
British Library Online Contents | 2013
|Quaternion Invariant Extended Kalman Filtering for Spacecraft Attitude Estimation
Online Contents | 2017
|