Real-time optimal planning for redirected walking using deep q-learning

Dang Yang Lee; Yang Hun Cho; In Kwan Lee

doi:10.1109/VR.2019.8798121

Real-time optimal planning for redirected walking using deep q-learning

Dang Yang Lee, Yang Hun Cho, In Kwan Lee

College of Computing

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

37 Citations (Scopus)

Abstract

This work presents a novel control algorithm of redirected walking called steer-to-optimal-target (S2OT) for effective real-time planning in redirected walking. S2OT is a method of redirection estimating the optimal steering target that can avoid the collision on the future path based on the user's virtual and physical paths. We design and train the machine learning model for estimating optimal steering target through reinforcement learning, especially, using the technique called Deep Q-Learning. S2OT significantly reduces the number of resets caused by collisions between user and physical space boundaries compared to well-known algorithms such as steer-to-center (S2C) and Model Predictive Control Redirection (MPCred). The results are consistent for any combinations of room-scale and large-scale physical spaces and virtual maps with or without predefined paths. S2OT also has a fast computation time of 0.763 msec per redirection, which is sufficient for redirected walking in real-time environments.

Original language	English
Title of host publication	26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	63-71
Number of pages	9
ISBN (Electronic)	9781728113777
DOIs	https://doi.org/10.1109/VR.2019.8798121
Publication status	Published - 2019 Mar
Event	26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Osaka, Japan Duration: 2019 Mar 23 → 2019 Mar 27

Publication series

Name	26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings

Conference

Conference	26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019
Country/Territory	Japan
City	Osaka
Period	19/3/23 → 19/3/27

Bibliographical note

Funding Information:
This work has supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT)(No. NRF-2017R1A2B4005469), and the MSIT(Ministry of Science and ICT), Korea, under the ITRC(Information Technology Research Center) support program(IITP-2018-2018-0-01419) supervised by the IITP(Institute for Information communications Technology Promotion.)

Publisher Copyright:
© 2019 IEEE.

All Science Journal Classification (ASJC) codes

Human-Computer Interaction
Media Technology

Access to Document

10.1109/VR.2019.8798121

Cite this

Lee, D. Y., Cho, Y. H., & Lee, I. K. (2019). Real-time optimal planning for redirected walking using deep q-learning. In 26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings (pp. 63-71). [8798121] (26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/VR.2019.8798121

Lee, Dang Yang ; Cho, Yang Hun ; Lee, In Kwan. / Real-time optimal planning for redirected walking using deep q-learning. 26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 63-71 (26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings).

@inproceedings{bd59022a167842c1b453540d10389e7b,

title = "Real-time optimal planning for redirected walking using deep q-learning",

abstract = "This work presents a novel control algorithm of redirected walking called steer-to-optimal-target (S2OT) for effective real-time planning in redirected walking. S2OT is a method of redirection estimating the optimal steering target that can avoid the collision on the future path based on the user's virtual and physical paths. We design and train the machine learning model for estimating optimal steering target through reinforcement learning, especially, using the technique called Deep Q-Learning. S2OT significantly reduces the number of resets caused by collisions between user and physical space boundaries compared to well-known algorithms such as steer-to-center (S2C) and Model Predictive Control Redirection (MPCred). The results are consistent for any combinations of room-scale and large-scale physical spaces and virtual maps with or without predefined paths. S2OT also has a fast computation time of 0.763 msec per redirection, which is sufficient for redirected walking in real-time environments.",

author = "Lee, {Dang Yang} and Cho, {Yang Hun} and Lee, {In Kwan}",

note = "Funding Information: This work has supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT)(No. NRF-2017R1A2B4005469), and the MSIT(Ministry of Science and ICT), Korea, under the ITRC(Information Technology Research Center) support program(IITP-2018-2018-0-01419) supervised by the IITP(Institute for Information communications Technology Promotion.) Publisher Copyright: {\textcopyright} 2019 IEEE.; 26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 ; Conference date: 23-03-2019 Through 27-03-2019",

year = "2019",

month = mar,

doi = "10.1109/VR.2019.8798121",

language = "English",

series = "26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "63--71",

booktitle = "26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings",

address = "United States",

}

Lee, DY, Cho, YH & Lee, IK 2019, Real-time optimal planning for redirected walking using deep q-learning. in 26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings., 8798121, 26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 63-71, 26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019, Osaka, Japan, 19/3/23. https://doi.org/10.1109/VR.2019.8798121

Real-time optimal planning for redirected walking using deep q-learning. / Lee, Dang Yang; Cho, Yang Hun; Lee, In Kwan.
26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. p. 63-71 8798121 (26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Real-time optimal planning for redirected walking using deep q-learning

AU - Lee, Dang Yang

AU - Cho, Yang Hun

AU - Lee, In Kwan

N1 - Funding Information: This work has supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT)(No. NRF-2017R1A2B4005469), and the MSIT(Ministry of Science and ICT), Korea, under the ITRC(Information Technology Research Center) support program(IITP-2018-2018-0-01419) supervised by the IITP(Institute for Information communications Technology Promotion.) Publisher Copyright: © 2019 IEEE.

PY - 2019/3

Y1 - 2019/3

N2 - This work presents a novel control algorithm of redirected walking called steer-to-optimal-target (S2OT) for effective real-time planning in redirected walking. S2OT is a method of redirection estimating the optimal steering target that can avoid the collision on the future path based on the user's virtual and physical paths. We design and train the machine learning model for estimating optimal steering target through reinforcement learning, especially, using the technique called Deep Q-Learning. S2OT significantly reduces the number of resets caused by collisions between user and physical space boundaries compared to well-known algorithms such as steer-to-center (S2C) and Model Predictive Control Redirection (MPCred). The results are consistent for any combinations of room-scale and large-scale physical spaces and virtual maps with or without predefined paths. S2OT also has a fast computation time of 0.763 msec per redirection, which is sufficient for redirected walking in real-time environments.

AB - This work presents a novel control algorithm of redirected walking called steer-to-optimal-target (S2OT) for effective real-time planning in redirected walking. S2OT is a method of redirection estimating the optimal steering target that can avoid the collision on the future path based on the user's virtual and physical paths. We design and train the machine learning model for estimating optimal steering target through reinforcement learning, especially, using the technique called Deep Q-Learning. S2OT significantly reduces the number of resets caused by collisions between user and physical space boundaries compared to well-known algorithms such as steer-to-center (S2C) and Model Predictive Control Redirection (MPCred). The results are consistent for any combinations of room-scale and large-scale physical spaces and virtual maps with or without predefined paths. S2OT also has a fast computation time of 0.763 msec per redirection, which is sufficient for redirected walking in real-time environments.

UR - http://www.scopus.com/inward/record.url?scp=85071863253&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85071863253&partnerID=8YFLogxK

U2 - 10.1109/VR.2019.8798121

DO - 10.1109/VR.2019.8798121

M3 - Conference contribution

AN - SCOPUS:85071863253

T3 - 26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings

SP - 63

EP - 71

BT - 26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019

Y2 - 23 March 2019 through 27 March 2019

ER -

Lee DY, Cho YH, Lee IK. Real-time optimal planning for redirected walking using deep q-learning. In 26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2019. p. 63-71. 8798121. (26th IEEE Conference on Virtual Reality and 3D User Interfaces, VR 2019 - Proceedings). doi: 10.1109/VR.2019.8798121

Real-time optimal planning for redirected walking using deep q-learning

Abstract

Publication series

Conference

Bibliographical note

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this