Neural HD Map Generation from Multiple Vectorized Tiles Locally Produced by Autonomous Vehicles

Miao Fan,Yi Yao,Jianping Zhang,Xiangbo Song,Daihui Wu

2024-09-05

Abstract:High-definition (HD) map is a fundamental component of autonomous driving systems, as it can provide precise environmental information about driving scenes. Recent work on vectorized map generation could produce merely 65% local map elements around the ego-vehicle at runtime by one tour with onboard sensors, leaving a puzzle of how to construct a global HD map projected in the world coordinate system under high-quality standards. To address the issue, we present GNMap as an end-to-end generative neural network to automatically construct HD maps with multiple vectorized tiles which are locally produced by autonomous vehicles through several tours. It leverages a multi-layer and attention-based autoencoder as the shared network, of which parameters are learned from two different tasks (i.e., pretraining and finetuning, respectively) to ensure both the completeness of generated maps and the correctness of element categories. Abundant qualitative evaluations are conducted on a real-world dataset and experimental results show that GNMap can surpass the SOTA method by more than 5% F1 score, reaching the level of industrial usage with a small amount of manual modification. We have already deployed it at Navinfo Co., Ltd., serving as an indispensable software to automatically build HD maps for autonomous driving systems.

Robotics

What problem does this paper attempt to address?

The paper aims to address the problem of high-definition (HD) map construction in autonomous driving systems. Specifically, existing methods based on onboard sensors can only generate about 65% of the map elements around the vehicle during a single drive, and how to integrate these local map elements into a high-quality, globally consistent HD map remains an unsolved challenge. To solve this problem, the authors propose GNMap, an end-to-end generative neural network that can automatically generate a complete HD map in a global coordinate system from local vector maps produced by multiple autonomous vehicles over multiple drives. GNMap employs a multi-layer attention mechanism autoencoder as a shared network and ensures the completeness of the generated map and the accuracy of map element classification through a two-stage learning process of pre-training and fine-tuning. Experimental results show that GNMap outperforms the current state-of-the-art methods by more than 5% in F1 score on real datasets and has been deployed at FourDimensional Tuner Company to automatically construct HD maps for autonomous driving systems in mainland China.