Navigation by Imitation in a Pedestrian-Rich Environment

Deep neural networks trained on demonstrations of human actions give robot the ability to perform self-driving on the road. However, navigation in a pedestrian-rich environment, such as a campus setup, is still challenging---one needs to take frequent interventions to the robot and take control over...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Bi, Jing, Xiao, Tianyou, Sun, Qiuyue, Xu, Chenliang
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 Bi, Jing
Xiao, Tianyou
Sun, Qiuyue
Xu, Chenliang
description Deep neural networks trained on demonstrations of human actions give robot the ability to perform self-driving on the road. However, navigation in a pedestrian-rich environment, such as a campus setup, is still challenging---one needs to take frequent interventions to the robot and take control over the robot from early steps leading to a mistake. An arduous burden is, hence, placed on the learning framework design and data acquisition. In this paper, we propose a new learning-from-intervention Dataset Aggregation (DAgger) algorithm to overcome the limitations brought by applying imitation learning to navigation in the pedestrian-rich environment. Our new learning algorithm implements an error backtrack function that is able to effectively learn from expert interventions. Combining our new learning algorithm with deep convolutional neural networks and a hierarchically-nested policy-selection mechanism, we show that our robot is able to map pixels direct to control commands and navigate successfully in real world without explicitly modeling the pedestrian behaviors or the world model.
doi_str_mv 10.48550/arxiv.1811.00506
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_1811_00506</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>1811_00506</sourcerecordid><originalsourceid>FETCH-LOGICAL-a676-86e0319a2304f8768879b362d9e802a8c14af0dd8e2b5c0fc605c17bc5142cd43</originalsourceid><addsrcrecordid>eNotj81Kw0AURmfjQqoP4Mp5gcQ7v7lZSqlaKFak-3AzM7EXmqlMQ7BvL7auDt_m8B0hHhTUFp2DJyo_PNcKlaoBHPhbge808xdNfMyyP8v1yNN1cJYkP1JMp6kw5eqTw16u8szlmMeUpztxM9DhlO7_uRC7l9Vu-VZttq_r5fOmIt_4Cn0Co1rSBuyAjUds2t54HduEoAmDsjRAjJh07wIMwYMLqumDU1aHaM1CPF61l-vdd-GRyrn7S-guCeYXR6FANQ</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Navigation by Imitation in a Pedestrian-Rich Environment</title><source>arXiv.org</source><creator>Bi, Jing ; Xiao, Tianyou ; Sun, Qiuyue ; Xu, Chenliang</creator><creatorcontrib>Bi, Jing ; Xiao, Tianyou ; Sun, Qiuyue ; Xu, Chenliang</creatorcontrib><description>Deep neural networks trained on demonstrations of human actions give robot the ability to perform self-driving on the road. However, navigation in a pedestrian-rich environment, such as a campus setup, is still challenging---one needs to take frequent interventions to the robot and take control over the robot from early steps leading to a mistake. An arduous burden is, hence, placed on the learning framework design and data acquisition. In this paper, we propose a new learning-from-intervention Dataset Aggregation (DAgger) algorithm to overcome the limitations brought by applying imitation learning to navigation in the pedestrian-rich environment. Our new learning algorithm implements an error backtrack function that is able to effectively learn from expert interventions. Combining our new learning algorithm with deep convolutional neural networks and a hierarchically-nested policy-selection mechanism, we show that our robot is able to map pixels direct to control commands and navigate successfully in real world without explicitly modeling the pedestrian behaviors or the world model.</description><identifier>DOI: 10.48550/arxiv.1811.00506</identifier><language>eng</language><subject>Computer Science - Computer Vision and Pattern Recognition</subject><creationdate>2018-11</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,781,886</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/1811.00506$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.1811.00506$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Bi, Jing</creatorcontrib><creatorcontrib>Xiao, Tianyou</creatorcontrib><creatorcontrib>Sun, Qiuyue</creatorcontrib><creatorcontrib>Xu, Chenliang</creatorcontrib><title>Navigation by Imitation in a Pedestrian-Rich Environment</title><description>Deep neural networks trained on demonstrations of human actions give robot the ability to perform self-driving on the road. However, navigation in a pedestrian-rich environment, such as a campus setup, is still challenging---one needs to take frequent interventions to the robot and take control over the robot from early steps leading to a mistake. An arduous burden is, hence, placed on the learning framework design and data acquisition. In this paper, we propose a new learning-from-intervention Dataset Aggregation (DAgger) algorithm to overcome the limitations brought by applying imitation learning to navigation in the pedestrian-rich environment. Our new learning algorithm implements an error backtrack function that is able to effectively learn from expert interventions. Combining our new learning algorithm with deep convolutional neural networks and a hierarchically-nested policy-selection mechanism, we show that our robot is able to map pixels direct to control commands and navigate successfully in real world without explicitly modeling the pedestrian behaviors or the world model.</description><subject>Computer Science - Computer Vision and Pattern Recognition</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2018</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj81Kw0AURmfjQqoP4Mp5gcQ7v7lZSqlaKFak-3AzM7EXmqlMQ7BvL7auDt_m8B0hHhTUFp2DJyo_PNcKlaoBHPhbge808xdNfMyyP8v1yNN1cJYkP1JMp6kw5eqTw16u8szlmMeUpztxM9DhlO7_uRC7l9Vu-VZttq_r5fOmIt_4Cn0Co1rSBuyAjUds2t54HduEoAmDsjRAjJh07wIMwYMLqumDU1aHaM1CPF61l-vdd-GRyrn7S-guCeYXR6FANQ</recordid><startdate>20181101</startdate><enddate>20181101</enddate><creator>Bi, Jing</creator><creator>Xiao, Tianyou</creator><creator>Sun, Qiuyue</creator><creator>Xu, Chenliang</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20181101</creationdate><title>Navigation by Imitation in a Pedestrian-Rich Environment</title><author>Bi, Jing ; Xiao, Tianyou ; Sun, Qiuyue ; Xu, Chenliang</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a676-86e0319a2304f8768879b362d9e802a8c14af0dd8e2b5c0fc605c17bc5142cd43</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2018</creationdate><topic>Computer Science - Computer Vision and Pattern Recognition</topic><toplevel>online_resources</toplevel><creatorcontrib>Bi, Jing</creatorcontrib><creatorcontrib>Xiao, Tianyou</creatorcontrib><creatorcontrib>Sun, Qiuyue</creatorcontrib><creatorcontrib>Xu, Chenliang</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Bi, Jing</au><au>Xiao, Tianyou</au><au>Sun, Qiuyue</au><au>Xu, Chenliang</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Navigation by Imitation in a Pedestrian-Rich Environment</atitle><date>2018-11-01</date><risdate>2018</risdate><abstract>Deep neural networks trained on demonstrations of human actions give robot the ability to perform self-driving on the road. However, navigation in a pedestrian-rich environment, such as a campus setup, is still challenging---one needs to take frequent interventions to the robot and take control over the robot from early steps leading to a mistake. An arduous burden is, hence, placed on the learning framework design and data acquisition. In this paper, we propose a new learning-from-intervention Dataset Aggregation (DAgger) algorithm to overcome the limitations brought by applying imitation learning to navigation in the pedestrian-rich environment. Our new learning algorithm implements an error backtrack function that is able to effectively learn from expert interventions. Combining our new learning algorithm with deep convolutional neural networks and a hierarchically-nested policy-selection mechanism, we show that our robot is able to map pixels direct to control commands and navigate successfully in real world without explicitly modeling the pedestrian behaviors or the world model.</abstract><doi>10.48550/arxiv.1811.00506</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.1811.00506
ispartof
issn
language eng
recordid cdi_arxiv_primary_1811_00506
source arXiv.org
subjects Computer Science - Computer Vision and Pattern Recognition
title Navigation by Imitation in a Pedestrian-Rich Environment
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-17T06%3A43%3A41IST&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=Navigation%20by%20Imitation%20in%20a%20Pedestrian-Rich%20Environment&rft.au=Bi,%20Jing&rft.date=2018-11-01&rft_id=info:doi/10.48550/arxiv.1811.00506&rft_dat=%3Carxiv_GOX%3E1811_00506%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