A Novel In-Motion Alignment Method Based on Trajectory Matching for Autonomous Vehicles

The field of automatic driving has always been a research hotspot. Based on inertial and odometer, dead reckoning is used to navigate for L4 and above level autonomous vehicles with less human intervention when GPS is blocked. The navigation accuracy is mainly determined by the alignment error and the drift accumulated over time. In this paper, we proposed a novel method to achieve initial alignment of inertial navigation for L4 and above level autonomous vehicles. The alignment is accomplished with a MEMS inertial measurement unit (MIMU) and single antenna global positioning system real-time kinematic (GPS-RTK). By matching the trajectory obtained from MIMU dead reckoning with the trajectory obtained from RTK, the initial heading angle is obtained, and the heading angle accuracy is improved through multiple iterations. In the case of RTK precision decreasing, the MIMU can still achieve lane-level navigation after the initial alignment. Because the alignment is accomplished in motion, there is no need for additional static alignment time before the vehicle starts. The field experiment results show that the method could achieve fast initial alignment in a few seconds in moving state with an alignment accuracy of 0.05 degrees.