Realistic Estimation of Model Parameters

Pavel Ettler

Abstract

Most often, the normal distribution N plays the key role in the process modelling and parameter estimation. The paper deals with 'realistic' estimation of model parameters which takes into account limitations on parameters which arise in industrial applications of the model-based adaptive control. Here the limitation of a normally distributed random variable is being modelled by specific distribution - the probabilistic mixture D. It is shown that relationship between distributions N and D coincides with properties of the generalized normal distribution G and that relations between their first and second statistical moments can be adequately approximated by G's cumulative distribution function and probability density function, respectively. The derived method is then applied to estimation of bounded parameters. In combination with the idea of parallel identification of the full and reduced models of the process, a working algorithm is derived. Performance of the algorithm is illustrated by examples on both simulated and real data.

Download


Paper Citation


in Harvard Style

Ettler P. (2017). Realistic Estimation of Model Parameters . In Proceedings of the 14th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO, ISBN 978-989-758-263-9, pages 527-534. DOI: 10.5220/0006395705270534


in Bibtex Style

@conference{icinco17,
author={Pavel Ettler},
title={Realistic Estimation of Model Parameters},
booktitle={Proceedings of the 14th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO,},
year={2017},
pages={527-534},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0006395705270534},
isbn={978-989-758-263-9},
}


in EndNote Style

TY - CONF
JO - Proceedings of the 14th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO,
TI - Realistic Estimation of Model Parameters
SN - 978-989-758-263-9
AU - Ettler P.
PY - 2017
SP - 527
EP - 534
DO - 10.5220/0006395705270534