2D LIDAR Aided INS for vehicle positioning in urban environments

Abstract

This paper presents a novel method to utilize 2D LIDAR for INS (Inertial Navigation System) aiding to improve 3D vehicle position estimation accuracy, especially when GNSS signals are shadowed. In the proposed framework, 2D LIDAR aiding is carried out without imposing any assumptions on the vehicle motion (e.g. we allow full six degree-of freedom motion). To achieve this, a closed-form formula is derived to predict the line measurement in the LIDAR's frame. This makes the feature association, residual formation and GUI display possible. With this formula, the Extended Kalman Filter (EKF) can be employed in a straightforward manner to fuse the LIDAR and IMU data to estimate the full state of the vehicle. Preliminary experimental results show the effectiveness of the LIDAR aiding in reducing the state estimation uncertainty along certain directions, when GNSS signals are shadowed.