Design and performance test of relative navigation of a low cost inertial SLAM

Despite its precise positioning performance, a GPS based navigation system may require the reference or augmentation station in close boundaries and is liable to be affected by satellite observation environments. Thus, this paper presents an INS/vision sensor integrated system, which in principle uses purely unknown feature points in previous epochs in order to cope with the limited GPS/INS integration environment. For the implementation of three-dimensional navigation using feature points, the presented system takes advantage of a robust image extraction and tracking algorithm, data association, and inertial SLAM filter algorithm. Finally, experimental results verified the performance of integrated navigation system, through which the performance enhancement in estimating relative position of the vehicle is demonstrated effectively.