Optimization-Based Visual-Inertial SLAM Tightly Coupled with Raw GNSS Measurements

Fusing vision, Inertial Measurement Unit (IMU) and Global Navigation Satellite System (GNSS) information is a promising solution for accurate global positioning in complex urban scenes, because of the complementarity of the different sensors. Unlike the loose coupling approaches and the EKF-based approaches in the literature, we propose an optimization-based visual-inertial SLAM tightly coupled with raw GNSS measurements, including pseudoranges and Doppler shift, which is the first of such approaches to our knowledge. Reprojection error, IMU pre-integration error and raw GNSS measurement error are jointly optimized using bundle adjustment in a sliding window, and the asynchronism between images and raw GNSS measurements is considered. Marginalization is performed in the sliding window, and some methods dealing with noisy measurements and vulnerable situations are employed. Experimental results on public dataset in complex urban scenes prove that our proposed approach outperforms state-of-the-art visual-inertial SLAM, GNSS single point positioning, as well as a loose coupling approach, both in the scenes that mainly contain low-rise buildings and the scenes that contain urban canyons.