Diagnosis of hidden faults in the RCLL

Research output: Chapter in Book/Report/Conference proceedingConference paperpeer-review

Abstract

The importance of Artificial Intelligence and Flexible Production is increasing in industry. Factories are evolving from static automation to complex autonomous systems to better cope with the challenges introduced by globalization. The RoboCup Logistics League (RCLL) was created as a testbed for flexible production of on-demand orders. In such a flexible domain, reliable scheduling requires
execution monitoring of actions. In this paper, we propose, and experimentally evaluate, a diagnosis to deal with execution faults in the RCLL. The proposed solution is based on the monitoring of the execution state of actions of a temporal plan. Our approach consists of a simple fault model, cascade-faults and a knowledge base, followed by replanning. The experimental results support the use of this principle to face the inconsistencies occurring by a nonobservable fault during the execution of plans.
Original languageEnglish
Title of host publication32nd International Workshop on Principle of Diagnosis
Number of pages8
Publication statusAccepted/In press - 13 Sept 2021
Event32nd International Workshop on Principle of Diagnosis: DX 2021 - Hamburg, Germany
Duration: 13 Sept 202115 Sept 2021
https://www.hsu-hh.de/imb/en/dx-2021

Conference

Conference32nd International Workshop on Principle of Diagnosis
Abbreviated titleDX 2021
Country/TerritoryGermany
CityHamburg
Period13/09/2115/09/21
Internet address

Keywords

  • Robotics
  • Diagnosis
  • Plan Execution Monitoring
  • Deliberation

ASJC Scopus subject areas

  • Engineering(all)
  • Computer Science(all)

Fields of Expertise

  • Information, Communication & Computing
  • Mobility & Production

Treatment code (Nähere Zuordnung)

  • Application

Cite this