GPS辅助的SINS系统快速动基座初始对准

Fast initial alignment of GPS-assisted SINS system on moving base

摘要:

为实现基于优化的动基座对准算法(OBA)对陀螺仪误差的估计,并使其能够应用于低精度SINS系统中,将自适应无迹卡尔曼滤波算法与OBA算法相结合,提出一种新的由GPS辅助的SINS系统快速动基座对准（FIMA）算法.该算法首先推导了陀螺仪常值漂移与失准角之间的关系,并以此构建非线性系统状态方程,然后用重力加速度和GPS输出速度的积分构建量测方程；由于系统存在非线性,提出使用UKF算法对失准角以及陀螺常值漂移进行估计；由于量测方程由速度和重力加速度的积分构成,量测噪声协方差难以确定,引入自适应滤波算法对量测噪声实时估计. 跑车实验结果表明:对于低精度SINS系统,该算法可在15 s左右将航向角误差收敛到3°以内,在3 min以后航向角误差可收敛到1°以内；与传统非线性动基座对准算法以及OBA算法相比,该算法可在无任何初始姿态信息的条件下快速对准,且能够对陀螺常值漂移进行在线估计和载体系失准角补偿,提高了动基座对准的精度和收敛性能.

Abstract:

To realize the estimation of gyro error by optimization-based in-motion alignment algorithm (OBA) and apply it to low precision SINS systems, a new fast in-motion alignment (FIMA) algorithm was proposed by combining the adaptive unscented Kalman filter algorithm with the OBA algorithm for SINS systems assisted by GPS. In the proposed algorithm, the relationship between the gyro constant drift and the misalignment angle was used to build the state equation, and the measurement equation was constructed by integrating the gravity acceleration and the ground speed. Since the system was nonlinear, the UKF algorithm was applied to estimate the misalignment angle and the gyro constant drift. Due to the uncertainty of measurement noise, an adaptive filtering algorithm was introduced to estimate the noise in real time. Results show that for low precision SINS systems, the proposed algorithm could converge the heading angle error to less than 3 degrees in about 15 s, and within 3 min, the heading angle error could be converged to less than 1 degree. Compared with the traditional nonlinear moving base alignment algorithm and the OBA algorithm, the proposed algorithm could realize rapid alignment under any misalignment angle. In addition, it could estimate the gyro constant drift online and compensate the misalignment angle of the system, which improved the alignment accuracy and convergence performance.