College of Automation, Harbin Engineering University, Harbin 150001, China.
Department of Electrical and Computer Engineering, University of Calgary, Calgary, AB T2N1N4, Canada.
Sensors (Basel). 2018 Sep 5;18(9):2952. doi: 10.3390/s18092952.
The strapdown inertial navigation system (SINS) is widely used in autonomous vehicles. However, the random drift error of gyroscope leads to serious accumulated navigation errors during long continuous operation of SINS alone. In this paper, we propose to combine the Inertial Measurement Unit (IMU) data with the line feature parameters from a camera to improve the navigation accuracy. The proposed method can also maintain the autonomy of the navigation system. Experimental results show that the proposed inertial-visual navigation system can mitigate the SINS drift and improve the accuracy, stability, and reliability of the navigation system.
捷联惯性导航系统（SINS）在自动驾驶车辆中得到了广泛应用。然而，由于陀螺仪的随机漂移误差，SINS 在长时间连续运行时会导致严重的累积导航误差。在本文中，我们提出将惯性测量单元（IMU）数据与相机的线特征参数相结合，以提高导航精度。所提出的方法还可以保持导航系统的自主性。实验结果表明，所提出的惯性-视觉导航系统可以减轻 SINS 漂移并提高导航系统的精度、稳定性和可靠性。
