Reflex-augmented reinforcement learning for electrical energy management in vehicles

Andreas Heimrath, Joachim Froeschl, Uwe Baumgarten

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

2 Scopus citations

Abstract

This paper presents reflex-augmented reinforcement learning (RARL) and its application to electrical energy management in vehicles. RARL complements reinforcement learning (RL) with an organically-inspired reflex to pave the way for the application ofRL in safety-critical systems which were previously limited to rule-based decision systems. RARL also makes fast training directly in complex technical systems possible to avoid the use of a simulator. A realization ofRARL based on the cybernetic viable system model is introduced. In case of electrical energy management, RARL can be expected to outperform rule-based decision systems in terms of efficiency. This is to be shown by simulations and experiments in the real vehicle.

Original languageEnglish
Title of host publication2018 World Congress in Computer Science, Computer Engineering and Applied Computing, CSCE 2018 - Proceedings of the 2018 International Conference on Artificial Intelligence, ICAI 2018
EditorsHamid R. Arabnia, David de la Fuente, Elena B. Kozerenko, Jose A. Olivas, Fernando G. Tinetti
PublisherCSREA Press
Pages429-430
Number of pages2
ISBN (Electronic)1601324804, 9781601324801
StatePublished - 2018
Event2018 International Conference on Artificial Intelligence, ICAI 2018 at 2018 World Congress in Computer Science, Computer Engineering and Applied Computing, CSCE 2018 - Las Vegas, United States
Duration: 30 Jul 20182 Aug 2018

Publication series

Name2018 World Congress in Computer Science, Computer Engineering and Applied Computing, CSCE 2018 - Proceedings of the 2018 International Conference on Artificial Intelligence, ICAI 2018

Conference

Conference2018 International Conference on Artificial Intelligence, ICAI 2018 at 2018 World Congress in Computer Science, Computer Engineering and Applied Computing, CSCE 2018
Country/TerritoryUnited States
CityLas Vegas
Period30/07/182/08/18

Keywords

  • Cybernetics
  • Electrical energy management
  • Reinforcement learning
  • Rule-based decision system
  • Safety-critical system

Fingerprint

Dive into the research topics of 'Reflex-augmented reinforcement learning for electrical energy management in vehicles'. Together they form a unique fingerprint.

Cite this