Kilometer Sign Positioning-Aided INS/Odometer Integration for Land Vehicle Autonomous Navigation

Abstract

Land vehicles generally rely on inertial navigation system (INS) and odometer integration system to realize high-precision autonomous navigation in a global navigation satellite system (GNSS)-denied environment. However, the navigation errors of the INS and odometer integration system will still accumulate with time due to the existence of odometer scale factor error and installation deviation angles between the INS and the odometer. To solve the problem, this article proposed an integration method, which utilizes kilometer signs to aid in correcting the navigation errors of the INS/odometer integration system. First, the absolute global positions of the kilometer signs are obtained based on image recognition and data matching. Second, the vehicle visual positioning results are obtained based on monocular camera after compensating lever arm between the kilometer sign and INS. Afterward, the vehicle positions from kilometer sign are fused by a filter to reduce the visual positioning uncertainty of the land vehicle. Finally, two Kalman filters are established to integrate the information from visual positioning, INS, and odometer. Simulation and vehicle experiment results verify the performance of the proposed method, which can improve the autonomous navigation accuracy of the land vehicle.