Land vehicle navigation using odometry/INS/vision integrated system

Abstract

Dead reckoning using odometry/INS can enhance the positioning accuracy compared with INS alone because odometry measurement errors wonpsilat grow over time. The scaling error of odometry, however, turns to be the main factor influencing on the system performance. So an improved odometry/INS integration approach is proposed that calibrate the scaling error real time at the beginning stage of integration based on the high accurate solution of INS, hence improve the positioning accuracy remarkably. But position errors are still increasing with the traveled distance since the heading error exists. Therefore, a new orientation method is proposed to re-initialize the odometry/INS system that utilizes the angle information to three known landmarks measured by the electro-optical detection system mounted on the roof of the vehicle. Simulation results show that the position and heading of land vehicles can be accurately determined only based on once measurement information of each landmark by the electro-optical detection system. The odometry/INS/vision land navigation system presented in this article is able to provide the relatively high precision navigation information over long distances at a relatively low cost, hence of great value in practice.