WOMD-Reasoning: A Large-Scale Dataset and Benchmark for Interaction and Intention Reasoning in Driving

We propose Waymo Open Motion Dataset-Reasoning (WOMD-Reasoning), a comprehensive large-scale dataset with 3 million Q&As built on WOMD focusing on describing and reasoning interactions and intentions in driving scenarios. Existing language datasets for driving primarily capture interactions caus...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Li, Yiheng, Fan, Cunxin, Ge, Chongjian, Zhao, Zhihao, Li, Chenran, Xu, Chenfeng, Yao, Huaxiu, Tomizuka, Masayoshi, Zhou, Bolei, Tang, Chen, Ding, Mingyu, Zhan, Wei
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Li, Yiheng
Fan, Cunxin
Ge, Chongjian
Zhao, Zhihao
Li, Chenran
Xu, Chenfeng
Yao, Huaxiu
Tomizuka, Masayoshi
Zhou, Bolei
Tang, Chen
Ding, Mingyu
Zhan, Wei
description We propose Waymo Open Motion Dataset-Reasoning (WOMD-Reasoning), a comprehensive large-scale dataset with 3 million Q&As built on WOMD focusing on describing and reasoning interactions and intentions in driving scenarios. Existing language datasets for driving primarily capture interactions caused by close distances. However, interactions induced by traffic rules and human intentions, which can occur over long distances, are yet sufficiently covered. To address this, WOMD-Reasoning presents by far the largest multi-modal Q&A dataset on real-world driving scenarios, covering a wide range of driving topics from map descriptions and motion status descriptions to narratives and analyses of agents' interactions, behaviors, and intentions. We further introduce Motion-LLaVA, a motion-language model fine-tuned on the proposed dataset with robust interaction reasoning capabilities. We benchmark its performance across various configurations including different input modalities, reasoning techniques, and network architectures. The robust, diverse, and multi-modal nature of WOMD-Reasoning highlights its potential to advance future autonomous driving research and enable a broad range of applications. The dataset and its vision modal extension are available at https://waymo.com/open/download, and the codes & prompts to build it are available at https://github.com/yhli123/WOMD-Reasoning.
doi_str_mv 10.48550/arxiv.2407.04281
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2407_04281</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2407_04281</sourcerecordid><originalsourceid>FETCH-arxiv_primary_2407_042813</originalsourceid><addsrcrecordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMjEw1zMwMbIw5GRIC_f3ddENSk0szs_LzEu3UnBU8EksSk_VDU5OzElVcEksSSxOLVFIzEtRcErNS87ITSzKVkjLL1LwzCtJLUpMLsnMzwPLgvh5YB7cMIXMPAWXoswyIJOHgTUtMac4lRdKczPIu7mGOHvogl0UX1CUCTS4Mh7ksniwy4wJqwAAS2tDeQ</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>WOMD-Reasoning: A Large-Scale Dataset and Benchmark for Interaction and Intention Reasoning in Driving</title><source>arXiv.org</source><creator>Li, Yiheng ; Fan, Cunxin ; Ge, Chongjian ; Zhao, Zhihao ; Li, Chenran ; Xu, Chenfeng ; Yao, Huaxiu ; Tomizuka, Masayoshi ; Zhou, Bolei ; Tang, Chen ; Ding, Mingyu ; Zhan, Wei</creator><creatorcontrib>Li, Yiheng ; Fan, Cunxin ; Ge, Chongjian ; Zhao, Zhihao ; Li, Chenran ; Xu, Chenfeng ; Yao, Huaxiu ; Tomizuka, Masayoshi ; Zhou, Bolei ; Tang, Chen ; Ding, Mingyu ; Zhan, Wei</creatorcontrib><description>We propose Waymo Open Motion Dataset-Reasoning (WOMD-Reasoning), a comprehensive large-scale dataset with 3 million Q&amp;As built on WOMD focusing on describing and reasoning interactions and intentions in driving scenarios. Existing language datasets for driving primarily capture interactions caused by close distances. However, interactions induced by traffic rules and human intentions, which can occur over long distances, are yet sufficiently covered. To address this, WOMD-Reasoning presents by far the largest multi-modal Q&amp;A dataset on real-world driving scenarios, covering a wide range of driving topics from map descriptions and motion status descriptions to narratives and analyses of agents' interactions, behaviors, and intentions. We further introduce Motion-LLaVA, a motion-language model fine-tuned on the proposed dataset with robust interaction reasoning capabilities. We benchmark its performance across various configurations including different input modalities, reasoning techniques, and network architectures. The robust, diverse, and multi-modal nature of WOMD-Reasoning highlights its potential to advance future autonomous driving research and enable a broad range of applications. The dataset and its vision modal extension are available at https://waymo.com/open/download, and the codes &amp; prompts to build it are available at https://github.com/yhli123/WOMD-Reasoning.</description><identifier>DOI: 10.48550/arxiv.2407.04281</identifier><language>eng</language><subject>Computer Science - Robotics</subject><creationdate>2024-07</creationdate><rights>http://creativecommons.org/licenses/by-nc-nd/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2407.04281$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2407.04281$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Li, Yiheng</creatorcontrib><creatorcontrib>Fan, Cunxin</creatorcontrib><creatorcontrib>Ge, Chongjian</creatorcontrib><creatorcontrib>Zhao, Zhihao</creatorcontrib><creatorcontrib>Li, Chenran</creatorcontrib><creatorcontrib>Xu, Chenfeng</creatorcontrib><creatorcontrib>Yao, Huaxiu</creatorcontrib><creatorcontrib>Tomizuka, Masayoshi</creatorcontrib><creatorcontrib>Zhou, Bolei</creatorcontrib><creatorcontrib>Tang, Chen</creatorcontrib><creatorcontrib>Ding, Mingyu</creatorcontrib><creatorcontrib>Zhan, Wei</creatorcontrib><title>WOMD-Reasoning: A Large-Scale Dataset and Benchmark for Interaction and Intention Reasoning in Driving</title><description>We propose Waymo Open Motion Dataset-Reasoning (WOMD-Reasoning), a comprehensive large-scale dataset with 3 million Q&amp;As built on WOMD focusing on describing and reasoning interactions and intentions in driving scenarios. Existing language datasets for driving primarily capture interactions caused by close distances. However, interactions induced by traffic rules and human intentions, which can occur over long distances, are yet sufficiently covered. To address this, WOMD-Reasoning presents by far the largest multi-modal Q&amp;A dataset on real-world driving scenarios, covering a wide range of driving topics from map descriptions and motion status descriptions to narratives and analyses of agents' interactions, behaviors, and intentions. We further introduce Motion-LLaVA, a motion-language model fine-tuned on the proposed dataset with robust interaction reasoning capabilities. We benchmark its performance across various configurations including different input modalities, reasoning techniques, and network architectures. The robust, diverse, and multi-modal nature of WOMD-Reasoning highlights its potential to advance future autonomous driving research and enable a broad range of applications. The dataset and its vision modal extension are available at https://waymo.com/open/download, and the codes &amp; prompts to build it are available at https://github.com/yhli123/WOMD-Reasoning.</description><subject>Computer Science - Robotics</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMjEw1zMwMbIw5GRIC_f3ddENSk0szs_LzEu3UnBU8EksSk_VDU5OzElVcEksSSxOLVFIzEtRcErNS87ITSzKVkjLL1LwzCtJLUpMLsnMzwPLgvh5YB7cMIXMPAWXoswyIJOHgTUtMac4lRdKczPIu7mGOHvogl0UX1CUCTS4Mh7ksniwy4wJqwAAS2tDeQ</recordid><startdate>20240705</startdate><enddate>20240705</enddate><creator>Li, Yiheng</creator><creator>Fan, Cunxin</creator><creator>Ge, Chongjian</creator><creator>Zhao, Zhihao</creator><creator>Li, Chenran</creator><creator>Xu, Chenfeng</creator><creator>Yao, Huaxiu</creator><creator>Tomizuka, Masayoshi</creator><creator>Zhou, Bolei</creator><creator>Tang, Chen</creator><creator>Ding, Mingyu</creator><creator>Zhan, Wei</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20240705</creationdate><title>WOMD-Reasoning: A Large-Scale Dataset and Benchmark for Interaction and Intention Reasoning in Driving</title><author>Li, Yiheng ; Fan, Cunxin ; Ge, Chongjian ; Zhao, Zhihao ; Li, Chenran ; Xu, Chenfeng ; Yao, Huaxiu ; Tomizuka, Masayoshi ; Zhou, Bolei ; Tang, Chen ; Ding, Mingyu ; Zhan, Wei</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-arxiv_primary_2407_042813</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Robotics</topic><toplevel>online_resources</toplevel><creatorcontrib>Li, Yiheng</creatorcontrib><creatorcontrib>Fan, Cunxin</creatorcontrib><creatorcontrib>Ge, Chongjian</creatorcontrib><creatorcontrib>Zhao, Zhihao</creatorcontrib><creatorcontrib>Li, Chenran</creatorcontrib><creatorcontrib>Xu, Chenfeng</creatorcontrib><creatorcontrib>Yao, Huaxiu</creatorcontrib><creatorcontrib>Tomizuka, Masayoshi</creatorcontrib><creatorcontrib>Zhou, Bolei</creatorcontrib><creatorcontrib>Tang, Chen</creatorcontrib><creatorcontrib>Ding, Mingyu</creatorcontrib><creatorcontrib>Zhan, Wei</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Li, Yiheng</au><au>Fan, Cunxin</au><au>Ge, Chongjian</au><au>Zhao, Zhihao</au><au>Li, Chenran</au><au>Xu, Chenfeng</au><au>Yao, Huaxiu</au><au>Tomizuka, Masayoshi</au><au>Zhou, Bolei</au><au>Tang, Chen</au><au>Ding, Mingyu</au><au>Zhan, Wei</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>WOMD-Reasoning: A Large-Scale Dataset and Benchmark for Interaction and Intention Reasoning in Driving</atitle><date>2024-07-05</date><risdate>2024</risdate><abstract>We propose Waymo Open Motion Dataset-Reasoning (WOMD-Reasoning), a comprehensive large-scale dataset with 3 million Q&amp;As built on WOMD focusing on describing and reasoning interactions and intentions in driving scenarios. Existing language datasets for driving primarily capture interactions caused by close distances. However, interactions induced by traffic rules and human intentions, which can occur over long distances, are yet sufficiently covered. To address this, WOMD-Reasoning presents by far the largest multi-modal Q&amp;A dataset on real-world driving scenarios, covering a wide range of driving topics from map descriptions and motion status descriptions to narratives and analyses of agents' interactions, behaviors, and intentions. We further introduce Motion-LLaVA, a motion-language model fine-tuned on the proposed dataset with robust interaction reasoning capabilities. We benchmark its performance across various configurations including different input modalities, reasoning techniques, and network architectures. The robust, diverse, and multi-modal nature of WOMD-Reasoning highlights its potential to advance future autonomous driving research and enable a broad range of applications. The dataset and its vision modal extension are available at https://waymo.com/open/download, and the codes &amp; prompts to build it are available at https://github.com/yhli123/WOMD-Reasoning.</abstract><doi>10.48550/arxiv.2407.04281</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2407.04281
ispartof
issn
language eng
recordid cdi_arxiv_primary_2407_04281
source arXiv.org
subjects Computer Science - Robotics
title WOMD-Reasoning: A Large-Scale Dataset and Benchmark for Interaction and Intention Reasoning in Driving
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-06T18%3A50%3A53IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=WOMD-Reasoning:%20A%20Large-Scale%20Dataset%20and%20Benchmark%20for%20Interaction%20and%20Intention%20Reasoning%20in%20Driving&rft.au=Li,%20Yiheng&rft.date=2024-07-05&rft_id=info:doi/10.48550/arxiv.2407.04281&rft_dat=%3Carxiv_GOX%3E2407_04281%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true