Kalman filter is widely taken as an optimal fusion algorithm in GNSS/INS integration, and GNSS positioning error is often simply modeled as white noise in a loosely-coupled system. But many research results have shown that GNSS positioning error has the characteristic of temporal correlation. Correct error model of Kalman filter can ensure that the level of estimated accuracy is equal to that of actual accuracy. Inaccurate model will influence the consistency between estimated accuracy and actual integrated accuracy. Moreover, quality control method based on variance-covariance with inconsistency is not reliable and not utilized to detect gross errors. So, this paper mainly researches on GNSS positioning and velocity error model and its impact on quality control in GNSS/INS integration. In this paper, state-augmentation method is applied to solve the problems of GNSS colored noise to ensure the navigation accuracy consistency, and the rationality of the improved model is verified by the effect of quality control. Simulation and field test results show that state-augmentation method can improve the consistency between actual error and estimation standard deviation, and satisfy the requirement of quality control based on variance-covariance to improve the integration reliability.

中文翻译：

---

基于一致性分析的GNSS / INS集成测量误差模型研究

卡尔曼滤波器被广泛认为是GNSS / INS集成中的最佳融合算法，并且在松耦合系统中，GNSS定位误差通常简单地建模为白噪声。但是许多研究结果表明，GNSS定位误差具有时间相关性的特征。卡尔曼滤波器的正确误差模型可以确保估计的精度水平等于实际的精度水平。不正确的模型将影响估计的准确性和实际的综合准确性之间的一致性。而且，基于具有不一致的方差-协方差的质量控制方法是不可靠的，并且没有被用于检测粗差。因此，本文主要研究GNSS定位和速度误差模型及其对GNSS / INS集成中质量控制的影响。在本文中，应用状态增强方法解决了GNSS色噪声问题，保证了导航精度的一致性，并通过质量控制的效果验证了改进模型的合理性。仿真和现场测试结果表明，状态增强方法可以提高实际误差与估计标准偏差之间的一致性，满足基于方差-协方差的质量控制要求，提高了集成可靠性。