Publication detail
SOME NOTES ON CONTROL OF ASYNCHRONOUS ELECTROMOTOR BY IMPROVED CARLA METHOD
Turek, M. Březina, T. Pulchart, J.
Czech title
Řízení asynchroního elektromotoru pomocí metody CARLA
English title
SOME NOTES ON CONTROL OF ASYNCHRONOUS ELECTROMOTOR BY IMPROVED CARLA METHOD
Type
conference paper
Language
en
Original abstract
Modifications of reinforcement learning algorithm, so called continuous action reinforcement learning automaton (CARLA), are presented in this contribution. Automaton learning algorithm is based on rewarding that gradually evolves the set of probability densities. This set is consequently used for action set determination. Modifications consist of improving learning parameters based on learned values. Thereby higher values of probability density near the best action are reached and therefore the variance of chosen actions is lower than original. The influence of modifications is proved by simulation study describing learning and behavior of asynchronous electromotor scalar control. Standard PSD controller is used whose parameter values represent actions of three independent automata. The goal of on line learning process is to minimize the mean square of control error. Here described modifications of algorithm allow the improvement of quality of revolutions control with preserving basic algorithm characteristics.
Czech abstract
V tomto příspěvku jsou představeny modifikace metody CARLA. Učení automatu je založeno na odměnách, na jejichž základě se postupně upravuje sada rozložení hustoty pravděpodobnosti. Ty jsou poté použity pro získání sady akcí. Modifikace spočívají ve vylepšování parametrů učení na základě naučených hodnot. Modifikace vedou ke zvýšení pravděpodobnosti výběru úspěšných akcí a snížení jejich rozptylu. Vliv modifikací je ověřen simulační studií popisující učení a chování řízení asynchroního motoru. Pro řízení je použit standartní diskrétní PID regulátor. Akce automatu představují jeho parametry. Cílem on-line učení je minimalizace odchylky od požadované hodnoty. Zde popsané modifikace umožňují zlepšení kvality řízení otáček při zachování charakteristik řídící metody.
English abstract
Modifications of reinforcement learning algorithm, so called continuous action reinforcement learning automaton (CARLA), are presented in this contribution. Automaton learning algorithm is based on rewarding that gradually evolves the set of probability densities. This set is consequently used for action set determination. Modifications consist of improving learning parameters based on learned values. Thereby higher values of probability density near the best action are reached and therefore the variance of chosen actions is lower than original. The influence of modifications is proved by simulation study describing learning and behavior of asynchronous electromotor scalar control. Standard PSD controller is used whose parameter values represent actions of three independent automata. The goal of on line learning process is to minimize the mean square of control error. Here described modifications of algorithm allow the improvement of quality of revolutions control with preserving basic algorithm characteristics.
Keywords in English
active magnetic bearing, control, CARLA
RIV year
2005
Released
09.05.2005
Location
Academy of Sciences of the Czech Republic, Prague
ISBN
80-85918-93-5
Book
Book of extended abstracts
Pages from–to
315–316
Pages count
2
BIBTEX
@inproceedings{BUT20505,
author="Milan {Turek} and Tomáš {Březina} and Jaroslav {Pulchart},
title="SOME NOTES ON CONTROL OF ASYNCHRONOUS ELECTROMOTOR BY IMPROVED CARLA METHOD",
booktitle="Book of extended abstracts",
year="2005",
month="May",
pages="315--316",
address="Academy of Sciences of the Czech Republic, Prague",
isbn="80-85918-93-5"
}