Integrating Vision and Audition within a Cognitive Architecture to Track Conversations

Abstract

We describe a computational cognitive architecture for robots which we call ACT-R/E (ACT-R/Embodied). ACT-R/E is based on ACT-R, but uses different visual, auditory and movement modules. We describe a model that uses ACT-R/E to integrate visual and auditory information to perform conversation tracking in a dynamic environment. We also performed an empirical evaluation study which shows that people see our conversational tracking system as extremely natural.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 01, 2008
Accession Number
ADA480061

Entities

People

  • Benjamin R. Fransen
  • J. Gregory Trafton
  • Magdalena D. Bugajska
  • Raj M. Ratwani

Organizations

  • United States Naval Research Laboratory

Tags

Communities of Interest

  • Autonomy

DTIC Thesaurus Topics

  • Abstracts
  • Ambient Noise
  • Background Noise
  • Computers
  • Detectors
  • Directional
  • Frequency
  • Governments
  • Human-Machine Interaction
  • Human-Robot Interaction
  • Intervals
  • Military Research
  • Person Tracking
  • Robots
  • Test And Evaluation
  • United States
  • Urban Areas

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Auditory Neuroscience/Auditory Physiology.

Technology Areas

  • AI & ML
  • AI & ML - Machine Translation
  • Autonomy