基于一阶高斯-马尔科夫模型的IMU零偏相关时间对GNSS/INS组合导航结果的影响

GNSS(global navigation satellite system)/INS(inertial navigation system)组合导航中,Kalman滤波的最优估计性能受状态模型及其参数设定影响。IMU(inertial measurement unit)的零偏和比例因子误差通常建模为一阶高斯-马尔科夫过程,而模型中的相关时间参数比较难以获得且影响不明确。因此,以IMU零偏模型为例,通过仿真低中高三种不同精度等级的IMU数据,并对比仿真GNSS中断内的导航漂移误差,研究了相关时间参数对GNSS/INS组合导航结果的影响。结果表明:当相关时间参数接近其真值时,组合导航结果在总体上趋于最优表现;中高等级IMU的加速度计零偏相关时间的影响相对较大;但总体来说,相关时间对组合导航结果影响较弱。该研究成果可为后续IMU误差模型的参数设定和优化策略设计提供一定的实验依据和研究基础。

Effects of IMU Bias Correlation Time on GNSS/INS Integrated Navigation Results Based on the First Order Gauss-Markov Model

In GNSS(global navigation satellite system)/INS(inertial navigation system) integrated navigation, the optimal estimation performance of Kalman filtering is affected by state models and their parameter setting. The bias and scale factor of IMU(inertial measurement unit) are usually modeled as the first order Gauss-Markov process, while in which the correlation time parameter is difficult to obtain and its effects are not yet clear. Therefore, taking the IMU bias model as an example, effects of the correlation time parameter on GNSS/INS integrated navigation results were studied by simulating three different grades of IMU data and comparing the navigation drifts at GNSS outages. The results showed that when the correlation time parameter approaches its true value, the integrated navigation results tend to the best performance as a whole; correlation time of the accelerometer bias in middle and high grade IMUs has a relatively large effect; but in general, correlation time has a weak effect on the integrated navigation results. The research results can subsequently provide some experimental basis and research foundation for the parameter setting and optimization strategy design of IMU error models.