Detection and Diagnosis of Deviations in Distributed Systems of Autonomous Agents - Cyber

Abstract

Detection and Diagnosis of Deviations in Distributed Systems of Autonomous Agents Project Summary Autonomous agents such as mobile robots, drones, intelligent buoys, underwater autonomous vehicles, and smart devices have great potential for providing services such as event security, search and rescue, managing hazardous materials, or tracking entities. Such agents must operate in unpredictable, possibly hostile environments where communication can be disrupted and only partial information is available. Agents may need to adapt in the face of failures and threats. SRI proposes to develop a formal framework for specifying and reasoning about autonomous agent systems. The framework will provide tools for describing which system executions are within a safety envelop and achieving goals within tolerance. It will also provide mechanisms for defining and detecting deviations from acceptable system executions, and algorithms and techniques for determinng the cause of deviations (diagnosis). Formal threat models will be developed to enable reasoning about how the environment can impact the system. Here environment includes natural phenomena, accidental interference, evil agents, and physical faults. Agents will monitor system execution to build models of system execution events and status from a local perspective. Agents collaborate by sharing and combining local models to develop partial views of a systems activity. These models will be used to detect deviations, predict possible deviations, and hypothesize unobserved events to explain the deviations. The framework will support inference under uncertainty (partial information) with precision of results improving as more information becomes available. Early detection can improve the precision of diagnosis and increase the ability to recover.

Document Details

Document Type
DoD Grant Award
Publication Date
Aug 12, 2016
Source ID
N000141512202

Entities

People

  • Carolyn Taclott

Organizations

  • Office of Naval Research
  • SRI International
  • United States Navy

Tags

Fields of Study

  • Computer science
  • Engineering

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Artificial Intelligence

Technology Areas

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