Provably-Correct, Perception-based Composition and Repair for High-level Tasks

Abstract

To enable long term safe deployment of autonomous systems in partially known andunstructured environments, they must be able to rea son about, compose and repair theiractions based on their perception of their environment, capabilities and task. Researchregardin g perception-based action composition and repair must be grounded in physicalsystems; both to identify implicit assumptions that ma y or may not be realistic, and toexploit physical constraints that make the problem scope more tractable.The proposed testbed will enable and ground research on correct, perception-based compositionand repair of robot actions. It consists of a static manipulato r, two mobile manipulatorswith different motion capabilities, a motion capture system, and four depth cameras.The testbed provides a rich set of robot capabilities (motion in the environment, object andenvironment manipulation) together with extensive onboard p erception (robot sensors) andground truth (motion capture and depth cameras) information. It will enable significantprogress and i nsight on creating autonomous systems exhibiting safe, robust and explainablecomplex behaviors in unstructured environments.The te stbed will directly support experimental validation and grounding of research aspart of an ONR Multi-University Research Initiative (MURI) grant focused on perceptionin-the-loop abstraction, specifications, verification, learning and repair techniques thatinteg rate perception and action of autonomous systems operating in unstructured andchanging environments. The testbed will be located in a recently renovated robotics spaceat Cornell, shared by over 20 PhD students, and will be used for education, in addition torese arch.

Document Details

Document Type
DoD Grant Award
Publication Date
Sep 08, 2021
Source ID
N000142112854

Entities

People

  • Hadas Kress-Gazit

Organizations

  • Cornell University
  • Office of Naval Research
  • United States Navy

Tags

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Distributed Systems and Data Platform Development
  • Sensor Fusion and Tracking Systems.

Technology Areas

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