Driving Data Collection Framework Using Low Cost Hardware
Abstract
Autonomous driving is driven by data. The availability of large and diverse data set from different geographies can help in maturing Autonomous driving technology faster. It is challenging to build a system to collect driving data which is cost intensive especially in emerging economies. Paradoxically these economies have chaotic driving conditions leading to a valuable data set. To address the issue of cost and scale, we have developed a data collection framework. In this paper, we’ll discuss our motive for the framework, performance bottlenecks, a two stage pipeline design and insights on how to tune the system to get maximum throughput.
Cite
Text
Jacob and Rabha. "Driving Data Collection Framework Using Low Cost Hardware." European Conference on Computer Vision Workshops, 2018. doi:10.1007/978-3-030-11021-5_38Markdown
[Jacob and Rabha. "Driving Data Collection Framework Using Low Cost Hardware." European Conference on Computer Vision Workshops, 2018.](https://mlanthology.org/eccvw/2018/jacob2018eccvw-driving/) doi:10.1007/978-3-030-11021-5_38BibTeX
@inproceedings{jacob2018eccvw-driving,
title = {{Driving Data Collection Framework Using Low Cost Hardware}},
author = {Jacob, Johnny and Rabha, Pankaj},
booktitle = {European Conference on Computer Vision Workshops},
year = {2018},
pages = {617-625},
doi = {10.1007/978-3-030-11021-5_38},
url = {https://mlanthology.org/eccvw/2018/jacob2018eccvw-driving/}
}