Faculty of Mechanics and Mathematics, Lomonosov Moscow State University, 119991 Moscow, Russia.
Sensors (Basel). 2022 Dec 21;23(1):77. doi: 10.3390/s23010077.
We address the angular misalignment calibration problem, which arises when a multi-antenna GNSS serves as a source of aiding information for inertial sensors in an integrated navigation system. Antennas usually occupy some outside structure of the moving carrier object, whilst an inertial measurement unit typically remains inside. Especially when using low- or mid-grade MEMS gyroscopes and accelerometers, it is either impossible or impractical to physically align IMU-sensitive axes and GNSS antenna baselines within some 1-3 degrees due to the micromechanical nature of the inertial sensors: they are just too small to have any physical reference features to align to. However, in some applications, it is desirable to line up all sensors within a fraction-of-a-degree level of accuracy. One may imagine solving this problem via the long-term averaging of sensor signals in different positions to ensure observability and then using angle differences for analytical compensation. We suggest faster calibration in special rotations using sensor fusion. Apart from quicker convergence, this method also accounts for run-to-run inertial sensor bias instability. In addition, it allows further on-the-fly finer calibration in the background when the navigation system performs its regular operation, and carrier objects may undergo gradual deformations of its structure over the years.
我们解决了多天线 GNSS 作为惯性传感器集成导航系统辅助信息源时出现的角度失准校准问题。天线通常位于移动载体对象的外部结构中,而惯性测量单元通常位于内部。特别是在使用低等级或中等级的 MEMS 陀螺仪和加速度计时,由于惯性传感器的微机械性质,实际上不可能或不切实际地在 1-3 度内物理对准 IMU 敏感轴和 GNSS 天线基线:它们太小了,没有任何物理参考特征可以对齐。然而,在某些应用中,希望将所有传感器以亚度级的精度对齐。人们可能会想象通过在不同位置对传感器信号进行长期平均来解决这个问题,以确保可观测性,然后使用角度差异进行分析补偿。我们建议在特殊旋转中使用传感器融合进行更快的校准。除了更快的收敛速度外,这种方法还考虑了惯性传感器偏置随时间变化的不稳定性。此外,当导航系统执行其常规操作时,它还允许在后台进行进一步的实时精细校准,并且载体对象可能会随着时间的推移而逐渐变形。