利用硬样本挖掘提高非结构化环境中的多车导航性能

Yining Ma, Ang Li, Qadeer Khan, Daniel Cremers
{"title":"利用硬样本挖掘提高非结构化环境中的多车导航性能","authors":"Yining Ma, Ang Li, Qadeer Khan, Daniel Cremers","doi":"arxiv-2409.05119","DOIUrl":null,"url":null,"abstract":"Contemporary research in autonomous driving has demonstrated tremendous\npotential in emulating the traits of human driving. However, they primarily\ncater to areas with well built road infrastructure and appropriate traffic\nmanagement systems. Therefore, in the absence of traffic signals or in\nunstructured environments, these self-driving algorithms are expected to fail.\nThis paper proposes a strategy for autonomously navigating multiple vehicles in\nclose proximity to their desired destinations without traffic rules in\nunstructured environments. Graphical Neural Networks (GNNs) have demonstrated good utility for this task\nof multi-vehicle control. Among the different alternatives of training GNNs,\nsupervised methods have proven to be most data-efficient, albeit require ground\ntruth labels. However, these labels may not always be available, particularly\nin unstructured environments without traffic regulations. Therefore, a tedious\noptimization process may be required to determine them while ensuring that the\nvehicles reach their desired destination and do not collide with each other or\nany obstacles. Therefore, in order to expedite the training process, it is\nessential to reduce the optimization time and select only those samples for\nlabeling that add most value to the training. In this paper, we propose a warm\nstart method that first uses a pre-trained model trained on a simpler subset of\ndata. Inference is then done on more complicated scenarios, to determine the\nhard samples wherein the model faces the greatest predicament. This is measured\nby the difficulty vehicles encounter in reaching their desired destination\nwithout collision. Experimental results demonstrate that mining for hard\nsamples in this manner reduces the requirement for supervised training data by\n10 fold. Videos and code can be found here:\n\\url{https://yininghase.github.io/multiagent-collision-mining/}.","PeriodicalId":501315,"journal":{"name":"arXiv - CS - Multiagent Systems","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Enhancing the Performance of Multi-Vehicle Navigation in Unstructured Environments using Hard Sample Mining\",\"authors\":\"Yining Ma, Ang Li, Qadeer Khan, Daniel Cremers\",\"doi\":\"arxiv-2409.05119\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Contemporary research in autonomous driving has demonstrated tremendous\\npotential in emulating the traits of human driving. However, they primarily\\ncater to areas with well built road infrastructure and appropriate traffic\\nmanagement systems. Therefore, in the absence of traffic signals or in\\nunstructured environments, these self-driving algorithms are expected to fail.\\nThis paper proposes a strategy for autonomously navigating multiple vehicles in\\nclose proximity to their desired destinations without traffic rules in\\nunstructured environments. Graphical Neural Networks (GNNs) have demonstrated good utility for this task\\nof multi-vehicle control. Among the different alternatives of training GNNs,\\nsupervised methods have proven to be most data-efficient, albeit require ground\\ntruth labels. However, these labels may not always be available, particularly\\nin unstructured environments without traffic regulations. Therefore, a tedious\\noptimization process may be required to determine them while ensuring that the\\nvehicles reach their desired destination and do not collide with each other or\\nany obstacles. Therefore, in order to expedite the training process, it is\\nessential to reduce the optimization time and select only those samples for\\nlabeling that add most value to the training. In this paper, we propose a warm\\nstart method that first uses a pre-trained model trained on a simpler subset of\\ndata. Inference is then done on more complicated scenarios, to determine the\\nhard samples wherein the model faces the greatest predicament. This is measured\\nby the difficulty vehicles encounter in reaching their desired destination\\nwithout collision. Experimental results demonstrate that mining for hard\\nsamples in this manner reduces the requirement for supervised training data by\\n10 fold. Videos and code can be found here:\\n\\\\url{https://yininghase.github.io/multiagent-collision-mining/}.\",\"PeriodicalId\":501315,\"journal\":{\"name\":\"arXiv - CS - Multiagent Systems\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-08\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - CS - Multiagent Systems\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.05119\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - CS - Multiagent Systems","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.05119","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

当代自动驾驶研究在模仿人类驾驶特征方面展现出巨大潜力。然而,它们主要适用于道路基础设施完备、交通管理系统完善的地区。因此,在没有交通信号或非结构化环境中,这些自动驾驶算法预计会失败。本文提出了一种策略,用于在非结构化环境中,在没有交通规则的情况下,自动导航多辆车,使其接近所需的目的地。图形神经网络(GNN)在多车控制任务中表现出良好的实用性。在训练 GNN 的各种方法中,有监督的方法被证明是数据效率最高的,尽管需要地面实况标签。然而,这些标签并不总是可用的,尤其是在没有交通规则的非结构化环境中。因此,可能需要一个繁琐的优化过程来确定这些标签,同时确保车辆到达预期目的地,并且不会相互碰撞或遇到任何障碍物。因此,为了加快训练过程,必须缩短优化时间,只选择那些对训练最有价值的样本进行标记。在本文中,我们提出了一种热启动方法,首先使用在较简单数据子集上训练的预训练模型。然后在更复杂的场景中进行推理,以确定模型面临最大困境的困难样本。这是以车辆在到达预期目的地时遇到的不碰撞困难来衡量的。实验结果表明,通过这种方式挖掘困难样本,对有监督训练数据的要求降低了 10 倍。视频和代码请点击这里:url{https://yininghase.github.io/multiagent-collision-mining/}.
本文章由计算机程序翻译,如有差异,请以英文原文为准。
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
Enhancing the Performance of Multi-Vehicle Navigation in Unstructured Environments using Hard Sample Mining
Contemporary research in autonomous driving has demonstrated tremendous potential in emulating the traits of human driving. However, they primarily cater to areas with well built road infrastructure and appropriate traffic management systems. Therefore, in the absence of traffic signals or in unstructured environments, these self-driving algorithms are expected to fail. This paper proposes a strategy for autonomously navigating multiple vehicles in close proximity to their desired destinations without traffic rules in unstructured environments. Graphical Neural Networks (GNNs) have demonstrated good utility for this task of multi-vehicle control. Among the different alternatives of training GNNs, supervised methods have proven to be most data-efficient, albeit require ground truth labels. However, these labels may not always be available, particularly in unstructured environments without traffic regulations. Therefore, a tedious optimization process may be required to determine them while ensuring that the vehicles reach their desired destination and do not collide with each other or any obstacles. Therefore, in order to expedite the training process, it is essential to reduce the optimization time and select only those samples for labeling that add most value to the training. In this paper, we propose a warm start method that first uses a pre-trained model trained on a simpler subset of data. Inference is then done on more complicated scenarios, to determine the hard samples wherein the model faces the greatest predicament. This is measured by the difficulty vehicles encounter in reaching their desired destination without collision. Experimental results demonstrate that mining for hard samples in this manner reduces the requirement for supervised training data by 10 fold. Videos and code can be found here: \url{https://yininghase.github.io/multiagent-collision-mining/}.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
期刊最新文献
Putting Data at the Centre of Offline Multi-Agent Reinforcement Learning HARP: Human-Assisted Regrouping with Permutation Invariant Critic for Multi-Agent Reinforcement Learning On-policy Actor-Critic Reinforcement Learning for Multi-UAV Exploration CORE-Bench: Fostering the Credibility of Published Research Through a Computational Reproducibility Agent Benchmark Multi-agent Path Finding in Continuous Environment
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1