Graduate Thesis Or Dissertation
 

Semantic embodied navigation : developing agents that navigate from language and vision

Öffentlich Deposited

Herunterladbarer Inhalt

PDF Herunterladen
https://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/n583z3248

Descriptions

Attribute NameValues
Creator
Abstract
  • Autonomous robotic agents are on their way to becoming in-home personal assistants, construction assistants, and warehouse workers. The degree of autonomy of such systems is reflected by the manner in which we specify goals to them; the abstraction of low-level commands to high-level goals goes hand-in-hand with increased autonomy. In this work, we seek to develop agents that accomplish semantic goals using the lens of semantic embodied navigation. In embodied navigation, agents act in spatial environments to follow a motion path or reach a destination. Semantic refers to how the goal is specified; instead of directly (e.g. coordinates), a goal is implied or its interpretations are constrained. We study two semantic embodied navigation tasks: navigating in response to natural language instructions and navigating to an object specified by an image. Our contributions are highlighted as follows. For both tasks, we establish benchmarks in simulation and evaluate baseline models. For language-based navigation, we develop language-conditioned waypoint prediction networks to study the impact of various action spaces. We then show that high performance on our benchmark can be achieved by transferring an agent from highly-abstract simulation to lower-level simulation. For image-based navigation, we find that a common sensors-to-action policy generalizes poorly to new environments and new goals. We then develop a modular method centered around an object re-identification sub-task that achieves strong performance on our simulation benchmark and successfully transfers to the real world.
License
Resource Type
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Committee Member
Academic Affiliation
Urheberrechts-Erklärung
Publisher
Peer Reviewed
Language

Beziehungen

Parents:

This work has no parents.

In Collection:

Artikel

Miniaturansicht Titel Hochladedatum Sichtbarkeit Aktionen
file details KrantzJacob2024.pdf 2023-12-12 Öffentlich Herunterladen