國立虎尾科技大學 |

Millimeter Wave Wireless Assisted Indoor Robot Navigation.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	Millimeter Wave Wireless Assisted Indoor Robot Navigation./
作者:	Yin, Mingsheng.
面頁冊數:	1 online resource (136 pages)
附註:	Source: Dissertations Abstracts International, Volume: 85-11, Section: B.
Contained By:	Dissertations Abstracts International85-11B.
標題:	Electrical engineering. -
電子資源:	click for full text (PQDT)
ISBN:	9798382757957

Millimeter Wave Wireless Assisted Indoor Robot Navigation.
Yin, Mingsheng.

Millimeter Wave Wireless Assisted Indoor Robot Navigation. - 1 online resource (136 pages)

Source: Dissertations Abstracts International, Volume: 85-11, Section: B.

Thesis (Ph.D.)--New York University Tandon School of Engineering, 2024.

Includes bibliographical references

The use of millimeter wave (mmWave) bands for high precision localization applications has garnered significant interest, particularly for indoor robot navigation, due to their high angular and temporal resolution. Traditional heuristic-based methods relying on radio frequency (RF) propagation offer intuitive solutions across simple scenarios but struggle in complex environments. This dissertation delves into mmWave-based positioning for target localization, where a stationary target emits mmWave signals and a mobile robot seeks to locate and navigate towards it. In the first part of the dissertation, a three-stage approach is introduced: First, the mobile agent uses tensor decomposition methods to detect the multipath channel components and estimate their parameters. Second, a machine-learning trained classifier is then used to predict the link state, meaning if the strongest path is line-of-sight (LOS) or non-LOS (NLOS). For the NLOS case, the link state predictor also determines if the strongest path arrived via one or more reflections. Third, based on the link state, the agent either follows the estimated angles or uses computer vision or other sensor to explore and map the environment. Demonstrated on an extensive indoor environment dataset, this method integrates path estimation and link state classification into an advanced neural simultaneous localization and mapping (SLAM) module, enhancing navigation towards the target with significant accuracy even in environments not included in the training set.Furthermore, the second part of this dissertation addresses RF propagation prediction in dynamic settings where the environment is only partially observed. A novel method leveraging partial observations to extract statistical channel models is introduced, combining simple heuristic algorithms with machine learning to estimate channel characteristics and uncertainties from partial ray tracing data.In contrast to the limitations of heuristic approaches and the potential of end-to-end (e2e) deep reinforcement learning (RL) in complex scenarios, the final part of this dissertation proposes a novel physics-informed RL (PIRL) framework. By augmenting a distance-to-target cost with physics-informed reward shaping, PIRL equips the navigation agent with efficient learning and zero-shot generalization capabilities, outperforming both heuristic and end-to-end RL methods. This approach underscores the enduring relevance of physics principles across varying environments, facilitating the agent's adaptation to unknown scenarios without the need for task-specific fine-tuning. It is shown that the PIRL significantly outperforms both e2e RL and heuristic-based solutions in terms of generalization and performance.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798382757957Subjects--Topical Terms:

596380
Electrical engineering.
Subjects--Index Terms:

Millimeter waveIndex Terms--Genre/Form:

554714
Electronic books.

Millimeter Wave Wireless Assisted Indoor Robot Navigation.
LDR:04129ntm a22003977 4500 001 1150339
005 20241028114750.5
006 m o d
007 cr bn ---uuuuu
008 250605s2024 xx obm 000 0 eng d
020 $a 9798382757957
035 $a (MiAaPQ)AAI31149184
035 $a AAI31149184
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Yin, Mingsheng. $3 1476805
245 1 0 $a Millimeter Wave Wireless Assisted Indoor Robot Navigation.
264 0 $c 2024
300 $a 1 online resource (136 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-11, Section: B.
500 $a Advisor: Rangan, Sundeep.
502 $a Thesis (Ph.D.)--New York University Tandon School of Engineering, 2024.
504 $a Includes bibliographical references
520 $a The use of millimeter wave (mmWave) bands for high precision localization applications has garnered significant interest, particularly for indoor robot navigation, due to their high angular and temporal resolution. Traditional heuristic-based methods relying on radio frequency (RF) propagation offer intuitive solutions across simple scenarios but struggle in complex environments. This dissertation delves into mmWave-based positioning for target localization, where a stationary target emits mmWave signals and a mobile robot seeks to locate and navigate towards it. In the first part of the dissertation, a three-stage approach is introduced: First, the mobile agent uses tensor decomposition methods to detect the multipath channel components and estimate their parameters. Second, a machine-learning trained classifier is then used to predict the link state, meaning if the strongest path is line-of-sight (LOS) or non-LOS (NLOS). For the NLOS case, the link state predictor also determines if the strongest path arrived via one or more reflections. Third, based on the link state, the agent either follows the estimated angles or uses computer vision or other sensor to explore and map the environment. Demonstrated on an extensive indoor environment dataset, this method integrates path estimation and link state classification into an advanced neural simultaneous localization and mapping (SLAM) module, enhancing navigation towards the target with significant accuracy even in environments not included in the training set.Furthermore, the second part of this dissertation addresses RF propagation prediction in dynamic settings where the environment is only partially observed. A novel method leveraging partial observations to extract statistical channel models is introduced, combining simple heuristic algorithms with machine learning to estimate channel characteristics and uncertainties from partial ray tracing data.In contrast to the limitations of heuristic approaches and the potential of end-to-end (e2e) deep reinforcement learning (RL) in complex scenarios, the final part of this dissertation proposes a novel physics-informed RL (PIRL) framework. By augmenting a distance-to-target cost with physics-informed reward shaping, PIRL equips the navigation agent with efficient learning and zero-shot generalization capabilities, outperforming both heuristic and end-to-end RL methods. This approach underscores the enduring relevance of physics principles across varying environments, facilitating the agent's adaptation to unknown scenarios without the need for task-specific fine-tuning. It is shown that the PIRL significantly outperforms both e2e RL and heuristic-based solutions in terms of generalization and performance.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Electrical engineering. $3 596380
650 4 $a Computer engineering. $3 569006
650 4 $a Robotics. $3 561941
653 $a Millimeter wave
653 $a Navigation
653 $a Positioning
653 $a Radio frequency
653 $a Deep reinforcement learning
655 7 $a Electronic books. $2 local $3 554714
690 $a 0544
690 $a 0464
690 $a 0771
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a New York University Tandon School of Engineering. $b Electrical & Computer Engineering. $3 1437750
773 0 $t Dissertations Abstracts International $g 85-11B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31149184 $z click for full text (PQDT)