The paper presents a comprehensive approach to enhance vehicle positioning in complex urban canyon environments. Key highlights include:
Adaptive Sky-view Image Segmentation: The authors introduce an FCN-based sky-view image segmentation algorithm that can adapt to varying lighting conditions, addressing limitations of traditional methods.
Tightly Coupled GNSS/INS/Vision Integration: The proposed Sky-GVIO model integrates GNSS, inertial navigation system (INS), and vision sensors in a tightly coupled framework. It extends the authors' previous NLOS detection and mitigation (S-NDM) algorithm to this integrated system.
Performance Evaluation: The paper conducts a thorough evaluation of the S-NDM algorithm's effectiveness within both GNSS pseudorange (single point positioning, SPP) and carrier phase (real-time kinematic, RTK) positioning frameworks. This sheds light on the algorithm's applicability across different GNSS-related integration techniques.
Open-Source Sky-view Image Dataset: The authors provide an open-source repository of sky-view images, including training and testing data, to contribute to the research community and address the lack of available resources in this field.
The experimental results demonstrate that the Sky-GVIO system can achieve meter-level accuracy under SPP mode and sub-decimeter precision with RTK, outperforming GNSS/INS/Vision frameworks without the S-NDM algorithm. This highlights the effectiveness of the proposed approach in enhancing vehicle positioning performance in challenging urban canyon environments.
To Another Language
from source content
arxiv.org
Дополнительные вопросы