Simak Vojtech, Andel Jan, Nemec Dusan, Kekelak Juraj
Department of Control and Information Systems, Faculty of Electrical Engineering and Information Technology, University of Žilina, 010 26 Žilina, Slovakia.
Sensors (Basel). 2024 Nov 25;24(23):7525. doi: 10.3390/s24237525.
Global satellite navigation systems (GNSSs) are the most-used technology for the localization of vehicles in the outdoor environment, but in the case of a densely built-up area or during passage through a tunnel, the satellite signal is not available or has poor quality. Inertial navigation systems (INSs) allow localization dead reckoning, but they have an integration error that grows over time. Inexpensive inertial measurement units (IMUs) are subject to thermal-dependent error and must be calibrated almost continuously. This article proposes a novel method of online (continuous) calibration of inertial sensors with the aid of the data from the GNSS receiver during the vehicle's route. We performed data fusion using an extended Kalman filter (EKF) and calibrated the input sensors through error backpropagation. The algorithm thus calibrates the INS sensors while the GNSS receiver signal is good, and after a GNSS failure, for example in tunnels, the position is predicted only by low-cost inertial sensors. Such an approach significantly improved the localization precision in comparison with offline calibrated inertial localization with the same sensors.
全球卫星导航系统（GNSS）是在户外环境中用于车辆定位的最常用技术，但在密集建成区或通过隧道时，卫星信号不可用或质量较差。惯性导航系统（INS）允许进行航位推算定位，但它们存在随时间增长的积分误差。廉价的惯性测量单元（IMU）会受到与温度相关的误差影响，并且几乎必须持续进行校准。本文提出了一种在车辆行驶过程中借助GNSS接收器的数据对惯性传感器进行在线（连续）校准的新方法。我们使用扩展卡尔曼滤波器（EKF）进行数据融合，并通过误差反向传播对输入传感器进行校准。该算法在GNSS接收器信号良好时校准INS传感器，在GNSS失效后，例如在隧道中，仅由低成本惯性传感器预测位置。与使用相同传感器的离线校准惯性定位相比，这种方法显著提高了定位精度。
