Hierarchical target-driven visual navigation in indoor environment
Navigation problem plays an important role in the field of artificial intelligence. In order to navigate the agent to a specific location efficiently, most researches take map information, relative distance towards the target or other prior information as input. Instead, this dissertation is target-...
Main Author: | |
---|---|
Other Authors: | |
Format: | Thesis-Master by Coursework |
Language: | English |
Published: |
Nanyang Technological University
2022
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/155435 |
_version_ | 1811677747887472640 |
---|---|
author | Zhu, Hongrun |
author2 | Wang Dan Wei |
author_facet | Wang Dan Wei Zhu, Hongrun |
author_sort | Zhu, Hongrun |
collection | NTU |
description | Navigation problem plays an important role in the field of artificial intelligence. In order to navigate the agent to a specific location efficiently, most researches take map information, relative distance towards the target or other prior information as input. Instead, this dissertation is target-driven navigation with only RGB images as input. To address this challenge, we decompose it into two subtasks: object-search and point-navigation. The agent first executes object-search task until the target object is detected, and then executes point-navigation task to reach the target. |
first_indexed | 2024-10-01T02:42:17Z |
format | Thesis-Master by Coursework |
id | ntu-10356/155435 |
institution | Nanyang Technological University |
language | English |
last_indexed | 2024-10-01T02:42:17Z |
publishDate | 2022 |
publisher | Nanyang Technological University |
record_format | dspace |
spelling | ntu-10356/1554352023-07-04T16:14:32Z Hierarchical target-driven visual navigation in indoor environment Zhu, Hongrun Wang Dan Wei School of Electrical and Electronic Engineering EDWWANG@ntu.edu.sg Engineering::Electrical and electronic engineering::Control and instrumentation::Robotics Navigation problem plays an important role in the field of artificial intelligence. In order to navigate the agent to a specific location efficiently, most researches take map information, relative distance towards the target or other prior information as input. Instead, this dissertation is target-driven navigation with only RGB images as input. To address this challenge, we decompose it into two subtasks: object-search and point-navigation. The agent first executes object-search task until the target object is detected, and then executes point-navigation task to reach the target. Master of Science (Computer Control and Automation) 2022-02-25T02:18:55Z 2022-02-25T02:18:55Z 2022 Thesis-Master by Coursework Zhu, H. (2022). Hierarchical target-driven visual navigation in indoor environment. Master's thesis, Nanyang Technological University, Singapore. https://hdl.handle.net/10356/155435 https://hdl.handle.net/10356/155435 en application/pdf Nanyang Technological University |
spellingShingle | Engineering::Electrical and electronic engineering::Control and instrumentation::Robotics Zhu, Hongrun Hierarchical target-driven visual navigation in indoor environment |
title | Hierarchical target-driven visual navigation in indoor environment |
title_full | Hierarchical target-driven visual navigation in indoor environment |
title_fullStr | Hierarchical target-driven visual navigation in indoor environment |
title_full_unstemmed | Hierarchical target-driven visual navigation in indoor environment |
title_short | Hierarchical target-driven visual navigation in indoor environment |
title_sort | hierarchical target driven visual navigation in indoor environment |
topic | Engineering::Electrical and electronic engineering::Control and instrumentation::Robotics |
url | https://hdl.handle.net/10356/155435 |
work_keys_str_mv | AT zhuhongrun hierarchicaltargetdrivenvisualnavigationinindoorenvironment |