Bootstrapping Autonomous Lane Changes with Self-Supervised Augmented Runs
Abstract
In this paper, we want to strengthen an autonomous vehicle’s lane-change ability with limited lane changes performed by the autonomous system. In other words, our task is bootstrapping the predictability of lane-change feasibility for the autonomous vehicle. Unfortunately, autonomous lane changes happen much less frequently in autonomous runs than in manual-driving runs. Augmented runs serve well in terms of data augmentation: the number of samples generated from augmented runs in a single one is comparable with that of samples retrieved from real runs in a month. In this paper, we formulate the Lane-Change Feasibility Prediction problem and also propose a data-driven learning approach to solve it. Experimental results are also presented to show the effectiveness of learned lane-change patterns for the decision making.
Cite
Text
Xiang. "Bootstrapping Autonomous Lane Changes with Self-Supervised Augmented Runs." European Conference on Computer Vision Workshops, 2022. doi:10.1007/978-3-031-25069-9_9Markdown
[Xiang. "Bootstrapping Autonomous Lane Changes with Self-Supervised Augmented Runs." European Conference on Computer Vision Workshops, 2022.](https://mlanthology.org/eccvw/2022/xiang2022eccvw-bootstrapping/) doi:10.1007/978-3-031-25069-9_9BibTeX
@inproceedings{xiang2022eccvw-bootstrapping,
title = {{Bootstrapping Autonomous Lane Changes with Self-Supervised Augmented Runs}},
author = {Xiang, Xiang},
booktitle = {European Conference on Computer Vision Workshops},
year = {2022},
pages = {118-130},
doi = {10.1007/978-3-031-25069-9_9},
url = {https://mlanthology.org/eccvw/2022/xiang2022eccvw-bootstrapping/}
}