Leveraging Egocentric and Allocentric Representations for Navigation (LEARN)

Abstract

Current autonomous vehicle systems -- service robots, autonomous cars, and unmanned drones -- cannot navigate well in complex, unknowns or changing environments. Most autonomous systems simple divide the navigation problem into two largely independent sub-problems: local navigation is based on sensor measurements that are prone to fail in cluttered environments; global navigation is based on 3D maps acquired previously. In contract, mammals navigate flexibly in the 3D natural world. Mammalian navigation is sub-served by a network of brain structures that represent the world both from the animals point of view (egocentric) and in a world-centered coordinate frame (allocentric). During natural navigation these representations are updated continuously through active perception. We propose an integrated program of experimental and theoretical work aimed at developing a new bio-inspired computational framework for autonomous vehicle navigation. Our framework will exploit dynamic egocentric information obtained as an agent moves through the environment in order to form, update and maintain allocentric representation of the world. The resulting framework will provide a platform for ground and aerial robots to explore and navigate in novel, dynamic environments, perform complex tasks such as searching and chasing, gather important geo-spatial information and share it with other agents.

Document Details

Document Type
DoD Grant Award
Publication Date
Dec 16, 2019
Source ID
N000142012002

Entities

People

  • Jack L. Gallant

Organizations

  • Office of Naval Research
  • United States Navy
  • University of California Regents

Tags

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Computer Vision.

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • Autonomy
  • Autonomy - Autonomous System Control