Model-free reinforcement-learning-based control methodology for power electronic converters

Dajr Alfred, Dariusz Czarkowski, Jiaxin Teng

Research output: Chapter in Book/Report/Conference proceedingConference contribution


This paper presents a novel reinforcement learning (RL) based discrete-time closed-loop control methodology for switch-mode, pulse-width-modulated (PWM) power electronic converters. This method of closed-loop optimal output regulation is achieved by utilizing measured data to approximate system dynamics, thus obviating the need for prior knowledge of system/plant dynamics. The underlying RL algorithm is then utilized to obtain the optimal feedback controller. The derived controller is obtained in a manner akin to that of a Linear Quadratic Regulator (LQR) and involves the iterative solution of an algebraic Riccati equation (ARE). This closed-loop control methodology is implemented on both buck and boost converters and its robustness to load and line variation is tested. A Type-III compensator was also developed in order to compare its performance with that of the proposed controller. Simulation results are provided to verify the effectiveness and examine the limitations of the proposed control strategy.

Original languageEnglish (US)
Title of host publicationProceedings - 2021 13th Annual IEEE Green Technologies Conference, GREENTECH 2021
PublisherIEEE Computer Society
Number of pages8
ISBN (Electronic)9781728191393
StatePublished - Apr 2021
Event13th Annual IEEE Green Technologies Conference, GREENTECH 2021 - Denver, United States
Duration: Apr 7 2021Apr 9 2021

Publication series

NameIEEE Green Technologies Conference
ISSN (Electronic)2166-5478


Conference13th Annual IEEE Green Technologies Conference, GREENTECH 2021
Country/TerritoryUnited States


  • Algebraic Riccati equation
  • Closed-loop control
  • Linear quadratic regulator
  • Model-free control
  • Off-policy reinforcement learning
  • Optimal output regulation
  • Power electronic converters
  • Type-III compensator

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Ecological Modeling
  • Environmental Engineering


Dive into the research topics of 'Model-free reinforcement-learning-based control methodology for power electronic converters'. Together they form a unique fingerprint.

Cite this