A Robust In-Motion Coarse Alignment Method for Low-Accuracy SINS and GPS Integrated System

Xiaoren Zhou,Meng Zhang,Jianchen Hu,Lin Li,Xiaohong Guan

DOI: https://doi.org/10.1109/tmech.2024.3397244

2024-01-01

Abstract:Many in-motion coarse alignment methods have been proposed and have achieved good results for the strapdown inertial navigation system (SINS) aided by the global positioning system (GPS). However, most existing methods are only applicable to the high-accuracy SINS and the potential influence of the outliers in the GPS outputs are usually ignored. In this article, we consider two practical difficulties: first, the accuracy of the vector observations is degraded by the gyroscope bias; second, the GPS outputs are affected by outliers. To address these issues, we propose a new robust quaternion unscented Kalman filter (RQUKF) method for the low-accuracy SINS and GPS integrated system. First, the RQUKF method is used to estimate the gyroscope bias with the aid of the GPS. Second, to obtain the reliable GPS outputs, the magnitude matching method based on the Huber's robust theory is utilized to detect and eliminate the outliers. The simulation and field tests are conducted and the results illustrate that even if the GPS outputs are affected by outliers, the proposed method still has the higher accuracy in terms of the yaw angle, than traditional algorithms without outliers in the GPS outputs.