Accurate and fast estimation of attitude is an important performance criteria for Strapdown Inertial Navigation System (SINS). Although many algorithms for SINS attitude estimation have been proposed by related researchers, most of them take the gravity vector as a known quantity. However, the gravity vector is inaccurate or even unknown in many cases, we have to use normal gravity instead of true gravity, which will bring errors to the mechanics equation of SINS, and reduce accuracy of attitude estimation. The coupling of gravity and attitude has been widely concerned by scholars and lack of effective solutions. In this paper, a new coarse attitude estimation method for dynamic base with decoupling the gravity vector is proposed. The method does not require the participation of gravity vector in attitude estimation process. Numerical simulation based on field test data is employed to verify the effectiveness and superiority of the proposed method in land vehicle navigation. The results indicate that the method improved the error caused by inaccurate gravity vector in traditional SINS attitude estimation and it provides a new idea for the measurement of vertical deviation.

中文翻译：

---

基于解耦重力​​矢量的动态基SINS姿态估计

准确快速的姿态估计是捷联惯性导航系统（SINS）的重要性能标准。尽管相关研究人员已经提出了许多 SINS 姿态估计算法，但大多数都以重力矢量为已知量。然而，在很多情况下，重力矢量是不准确的甚至未知的，我们不得不使用法向重力而不是真重力，这会给SINS的力学方程带来误差，降低姿态估计的准确性。重力与姿态的耦合一直为学者广泛关注，缺乏有效的解决方案。本文提出了一种新的重力矢量解耦动态基体姿态粗估计方法。该方法不需要重力矢量参与姿态估计过程。基于现场试验数据的数值模拟验证了该方法在陆地车辆导航中的有效性和优越性。结果表明，该方法改善了传统SINS姿态估计中重力矢量不准确造成的误差，为垂直偏差的测量提供了一种新的思路。