International Conferences
| Vehicle Localization Network using Simultaneous Coarse and Fine Visual Map Matching |
| Year |
2023 |
| Month |
|
| Journal |
2023 / IEEE Conference on Intelligent Transportation Systems (ITSC) |
| Author |
Jeong-Hoon Kim, Seung-Hyun Kong*, Min-Hyeok Sun |
|
Positioning is crucial for safe and efficient navigation for vehicle. Global Positioning System (GPS) is the most widely used positioning system by low-cost receiver, but suffers from poor performance in signal-shadowed areas such as urban environments. Additionally, the GPS receiver requires online external correction information to achieve lanelevel position accuracy(<1.5m). Simultaneous Localization and Mapping (SLAM) is a common localization technique which operates without satellite signals. However, even SLAM has a drawback of accumulating position estimation error over time. In this paper, we propose a deep learning-based position correcting system that can operate offline using the Visual Map Matching Network(VMMNet) which is introduced as an endto-end network for matching high-definition map (HD map) and Bird-Eye View (BEV) images from camera. The system can also take role of an independent positioning system by continuously correcting the previous corrected position. Our results demonstrate that the proposed system can improve position accuracy of GPS receiver to the lane-level in urban areas, and can conduct stable positioning even in the situation where the GPS is not working. |
https://ieeexplore.ieee.org/document/10422635
22회 연결
