Detail publikace

On-line Fuzzy Parameters Setting of Predictive Controller

Originální název

On-line Fuzzy Parameters Setting of Predictive Controller

Anglický název

On-line Fuzzy Parameters Setting of Predictive Controller

Jazyk

en

Originální abstrakt

Classical model–based predictive regulation sets the action signal magnitude in a way that ensures, that the chosen cost function will reach its minimum, which means, that the proper cost function parameters setting is crucial for the shape of transient. These parameters are usually constants that do not change their value in the course of regulation. However, given the actual condition of the regulation, their change can improve the quality of the transient. This solution reduces the adverse effects caused by the inexact model of the plant or short predictive horizon (short sample period) with regard to the length of the transient. This can occur when we use adaptive predictive control, where the dynamics and consequently also the duration of the transient are not constant. With respect to the fact, that the relation between the state of regulation and the suitable cost function parameters setting is strongly non-linear, fuzzy rules can bring an asset to the description of this. So the fuzzy system will operate as a supervisor of a predictive controller.

Anglický abstrakt

Classical model–based predictive regulation sets the action signal magnitude in a way that ensures, that the chosen cost function will reach its minimum, which means, that the proper cost function parameters setting is crucial for the shape of transient. These parameters are usually constants that do not change their value in the course of regulation. However, given the actual condition of the regulation, their change can improve the quality of the transient. This solution reduces the adverse effects caused by the inexact model of the plant or short predictive horizon (short sample period) with regard to the length of the transient. This can occur when we use adaptive predictive control, where the dynamics and consequently also the duration of the transient are not constant. With respect to the fact, that the relation between the state of regulation and the suitable cost function parameters setting is strongly non-linear, fuzzy rules can bring an asset to the description of this. So the fuzzy system will operate as a supervisor of a predictive controller.

BibTex


@inproceedings{BUT2535,
  author="Petr {Pivoňka} and Petr {Halva}",
  title="On-line Fuzzy Parameters Setting of Predictive Controller",
  annote="Classical model–based predictive regulation sets the action signal magnitude in a way that ensures, that the chosen cost function will reach its minimum, which means, that the proper cost function parameters setting is crucial for the shape of transient. These parameters are usually constants that do not change their value in the course of regulation. However, given the actual condition of the regulation, their change can improve the quality of the transient. This solution reduces the adverse effects caused by the inexact model of the plant or short predictive horizon (short sample period) with regard to the length of the transient. This can occur when we use adaptive predictive control, where the dynamics and consequently also the duration of the transient are not constant. With respect to the fact, that the relation between the state of regulation and the suitable cost function parameters setting is strongly non-linear, fuzzy rules can bring an asset to the description of this. So the fuzzy system will operate as a supervisor of a predictive controller.",
  address="IPM, University of Applied Science, Zittau",
  booktitle="Proceedings of East West Fuzzy Colloquium 2000, 8th Zittau Fuzzy Colloquium",
  chapter="2535",
  institution="IPM, University of Applied Science, Zittau",
  year="2000",
  month="august",
  pages="89",
  publisher="IPM, University of Applied Science, Zittau",
  type="conference paper"
}