A Robot for Research in Language Guided Dexterous Mobile Manipulation

Abstract

The Defense University Research Instrumentation Program (DURIP) is designed to improve the research capabilities of U.S. Universities and to educate scientists and engineers in selected technical areas of importance to national defense. DURIP funding provides for the acquisition of research equipment and instrumentation for this purpose. For autonomous systems to become effective parts of human-robot teams, they must demonstrate the capacity to perceive the world, learn from experience, adapt to changes in the environment, and naturally and efficiently share information with human partners. Mobile manipulators are particularly important because they offer the flexibility to interact with the environment in meaningful ways, such as transporting materials, assembling structures, or sampling in-situ. The complexity and dimensionality of such systems however limits their ability to achieve robust performance with a fast operational tempo, and contemporary approaches for natural language understanding that do not address the diversity of physical interactions that a mobile manipulator may undertake. This proposal concerns the acquisition of a robotic mobile manipulator to investigate algorithms and models for efficient language-guided dexterous manipulation, motion planning, and mobile manipulation. PI Thomas Howard will use the mobile manipulator to augment and enhance research capabilities in the area of robotics, motion planning, and model learning at the Robotics and Artificial Intelligence Laboratory at the University of Rochester. Specifically, PI Howard will use the platform for investigations into adaptive probabilistic graphical models for efficient natural language understanding and adaptive recombinant search spaces for optimal robot motion planning. We will use the platform as a base of experiments for learning how we can adapt the structure of probabilistic graphical models for natural language understanding based on information acquired from on-board perception and for experiments related to adapting the effective resolution of recombinant state lattices for optimal robot motion planning.

Document Details

Document Type
DoD Grant Award
Publication Date
Jan 12, 2017
Source ID
W911NF1510402

Entities

People

  • Thomas P Howard

Organizations

  • Army Contracting Command
  • United States Army
  • University of Rochester

Tags

Readers

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

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • AI & ML - DoD AI Strategy
  • Autonomy
  • Space
  • Space - Spacecraft Maneuvers