A Cognitive Robot Platform for Scene Understanding and Expeditionary Maneuver

Abstract

The objective of this DURIP proposal is to develop a cognitive robot platform with the following capabilities: (i) Understanding 3D scenes for robot navigation and manipulation, (ii) Reasoning about physical and social commonsense for human-robot interaction, and (iii) Learning and generalizing knowledge obtained by interacting with the environment, introducing a higher level of autonomy to a robot. To achieve these goals, the infrastructures of (i) a general-purpose state-of-the-art dual-arm mobile manipulator with depth sensors, Lidar, and tactile sensors, and (ii) a cutting-edge GPU server as the cloud service, are required. The platform will be used by the Center for Vision, Cognition, Learning, and Autonomy (VCLA) at the University of California, Los Angeles (UCLA), directed by PI Song-Chun Zhu. In particular, the mobile manipulator is used to develop, examine, and validate the cognitive robot platform with real-life applications. The GPU server can support large-scale physics-based simulations in a virtual environment for robot learning, as well as promote real-time computing when the robotic system is deployed.The proposed effort is of importance for several DoD tasks, such as (i) intelligence and surveillance with ground and aerial sensors by constructing a 3D scene from a network of ground and tower cameras, which requires a deep understanding of the scenes and activities to support queries and retrieval from analytics, and (ii) autonomous robotssearch and rescue missions when a robot needs to understand the physical and social scenes in order to perform its tasks and interact with human agents. The necessary infrastructures also benefitseveral on-going DoD projects at VCLA as well as various research communities in computer vision, machine learning, cognition, and artificial intelligence.

Document Details

Document Type
DoD Grant Award
Publication Date
Aug 31, 2020
Source ID
N000142012812

Entities

People

  • Song-Chun Zhu

Organizations

  • Office of Naval Research
  • United States Navy
  • University of California, Los Angeles

Tags

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Research Science/Academic Research

Technology Areas

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