MOPA: Modular Object Navigation with PointGoal Agents

We propose a simple but effective modular approach MOPA (Modular ObjectNav with PointGoal agents) to systematically investigate the inherent modularity of the object navigation task in Embodied AI. MOPA consists of four modules: (a) an object detection module trained to identify objects from RGB ima...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2024-01
Main Authors: Raychaudhuri, Sonia, Campari, Tommaso, Jain, Unnat, Savva, Manolis, Chang, Angel X
Format: Article
Language:English
Subjects:
Color imagery
Modularity
Modules
Navigation
Object recognition
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We propose a simple but effective modular approach MOPA (Modular ObjectNav with PointGoal agents) to systematically investigate the inherent modularity of the object navigation task in Embodied AI. MOPA consists of four modules: (a) an object detection module trained to identify objects from RGB images, (b) a map building module to build a semantic map of the observed objects, (c) an exploration module enabling the agent to explore the environment, and (d) a navigation module to move to identified target objects. We show that we can effectively reuse a pretrained PointGoal agent as the navigation model instead of learning to navigate from scratch, thus saving time and compute. We also compare various exploration strategies for MOPA and find that a simple uniform strategy significantly outperforms more advanced exploration methods.
ISSN:2331-8422