Sky-GVIO: an enhanced GNSS/INS/Vision navigation with FCN-based sky-segmentation in urban canyon

TL;DR

This work tackles precise localization in urban canyons where GNSS signals suffer severe NLOS effects. It introduces Sky-GVIO, a tightly coupled GNSS/INS/Vision system augmented by S-NDM, which leverages FCN-based sky-view segmentation to identify LOS/NLOS satellites and adapt GNSS observation models accordingly. The approach demonstrates meter-level accuracy for SPP and sub-decimeter accuracy for RTK in challenging environments, outperforming baselines like VINS-mono and GVINS, and it releases a sky-view image dataset to support further research. Overall, Sky-GVIO enhances robustness and accuracy for autonomous driving and mobile navigation in urban canyons by integrating semantic sky information into multi-sensor fusion.

Abstract

Accurate, continuous, and reliable positioning is a critical component of achieving autonomous driving. However, in complex urban canyon environments, the vulnerability of a stand-alone sensor and non-line-of-sight (NLOS) caused by high buildings, trees, and elevated structures seriously affect positioning results. To address these challenges, a sky-view images segmentation algorithm based on Fully Convolutional Network (FCN) is proposed for GNSS NLOS detection. Building upon this, a novel NLOS detection and mitigation algorithm (named S-NDM) is extended to the tightly coupled Global Navigation Satellite Systems (GNSS), Inertial Measurement Units (IMU), and visual feature system which is called Sky-GVIO, with the aim of achieving continuous and accurate positioning in urban canyon environments. Furthermore, the system harmonizes Single Point Positioning (SPP) with Real-Time Kinematic (RTK) methodologies to bolster its operational versatility and resilience. In urban canyon environments, the positioning performance of S-NDM algorithm proposed in this paper is evaluated under different tightly coupled SPP-related and RTK-related models. The results exhibit that Sky-GVIO system achieves meter-level accuracy under SPP mode and sub-decimeter precision with RTK, surpassing the performance of GNSS/INS/Vision frameworks devoid of S-NDM. Additionally, the sky-view image dataset, inclusive of training and evaluation subsets, has been made publicly accessible for scholarly exploration at https://github.com/whuwangjr/sky-view-images .

Figures (13)

• Figure 1: The system structure of the proposed Sky-GVIO.
• Figure 2: The sky-view images segmentation algorithm based on FCN.
• Figure 3: The overall flow of S-NDM algorithm.
• Figure 4: Illustration of experimental hardware platform.
• Figure 5: The experimental route and scene in the urban canyon. (A, B, C and D on the right correspond to the sky-view images of the four scenes in the trajectory, respectively. In the sky-view images on the right, the red dots represent the NLOS satellite, the blue dots represent the LOS satellite.)