Answers are grounded in this paper and cite their sources.
Accurate and reliable vehicle localization in urban environments is a critical requirement for autonomous driving and intelligent transportation systems. However, achieving high-precision localization in such settings remains a significant challenge due to the limitations of individual sensors. This paper presents MC-GVIO, a novel GNSS-visual-inertial odometry system that integrates GNSS-RTK with monocular VIO, leveraging motion consistency to enhance localization accuracy and robustness. MC-GVIO incorporates three key mechanisms to improve accuracy and robustness. A smooth optical flow field is utilized to evaluate and remove abnormal feature points from system optimization. Additionally, an adaptive sliding window optimization strategy is designed based on vehicle motion states, which integrates zero-velocity constraint, adaptive keyframe selection, and edge feature compensation. Finally, a VIO-assisted GNSS outlier rejection mechanism is adopted to further suppress errors. By fully leveraging the consistency of motion information among different sensors mounted on the vehicle, MC-GVIO effectively mitigates the impact of urban challenges such as multipath effects and dynamic obstacles, ensuring reliable localization even in complex scenarios. Extensive experiments on both public and private datasets demonstrate that MC-GVIO outperforms existing methods in terms of accuracy and robustness, making it a promising solution for urban vehicle localization.
No takes yet. Share an insight, caveat, or question.
Li et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: