A Vision/Inertial Navigation/Global Navigation Satellite Integrated System for Relative and Absolute Localization in Land Vehicles

Yao Zhang,Liang Chu,Yabin Mao,Xintong Yu,Jiawei Wang,Chong Guo

DOI: https://doi.org/10.3390/s24103079

IF: 3.9

2024-05-13

Sensors

Abstract:This paper presents an enhanced ground vehicle localization method designed to address the challenges associated with state estimation for autonomous vehicles operating in diverse environments. The focus is specifically on the precise localization of position and orientation in both local and global coordinate systems. The proposed approach integrates local estimates generated by existing visual–inertial odometry (VIO) methods into global position information obtained from the Global Navigation Satellite System (GNSS). This integration is achieved through optimizing fusion in a pose graph, ensuring precise local estimation and drift-free global position estimation. Considering the inherent complexities in autonomous driving scenarios, such as the potential failures of a visual–inertial navigation system (VINS) and restrictions on GNSS signals in urban canyons, leading to disruptions in localization outcomes, we introduce an adaptive fusion mechanism. This mechanism allows seamless switching between three modes: utilizing only VINS, using only GNSS, and normal fusion. The effectiveness of the proposed algorithm is demonstrated through rigorous testing in the Carla simulation environment and challenging UrbanNav scenarios. The evaluation includes both qualitative and quantitative analyses, revealing that the method exhibits robustness and accuracy.

engineering, electrical & electronic,instruments & instrumentation,chemistry, analytical

What problem does this paper attempt to address?

The paper aims to address the challenges of state estimation for autonomous vehicles in different environments, particularly the issues of precise position and orientation localization. Specifically, the paper proposes an enhanced ground vehicle localization method to tackle the problems that existing Visual-Inertial Navigation Systems (VINS) and Global Navigation Satellite Systems (GNSS) may encounter in complex environments such as urban canyons.

The main objectives of the paper include:
1. **Accurate local and global localization**: Achieving high-precision local and global localization by fusing local estimates generated by existing Visual-Inertial Odometry (VIO) methods with global position information provided by GNSS.
2. **Adaptive fusion mechanism**: Introducing an adaptive fusion mechanism that allows seamless switching between three modes: using only VINS, using only GNSS, and normal fusion. This mechanism can handle potential failures of VINS or GNSS, ensuring the continuity and robustness of the system.
3. **Two-stage initialization method**: Proposing a two-stage initialization method that uses inertial sensors, cameras, and asynchronous GPS measurement data for coarse initialization, and further optimizes the initialization parameters after the vehicle starts moving.

The effectiveness of the proposed algorithm is validated through rigorous testing in the Carla simulation environment and challenging UrbanNav scenarios. Experimental results demonstrate that the method exhibits good robustness and accuracy in different scenarios.