Recurrent fuzzy CMAC in hierarchical form for dynamic system identification

Floriberto Ortiz Rodriguez, Wen Yu, Marco A. Moreno-Armendariz

Research output: Contribution to conferencePaperResearch

3 Citations (Scopus)

Abstract

The conventional fuzzy CMAC neural networks perform well in terms of their fast learning speed and local generalization capability for approximating nonlinear functions. However, it requires an enormous memory and the dimension increase exponentially with the input number. In this paper, we use two techniques to overcome these problems: recurrent and hierarchical structures and propose a new CMAC, named Hierarchical Recurrent Fuzzy CMAC (HRFCMAC). Since the structure of HRFCMAC is very complex, the normal training methods are difficult to be applied. A new simple algorithm is given, we can train each sub-block of the hierarchical CMAC independently. A time-varying learning rate assures the learning algorithm is stable. © 2007 IEEE.
Original languageAmerican English
Pages5706-5711
Number of pages5134
DOIs
StatePublished - 1 Dec 2007
EventProceedings of the American Control Conference -
Duration: 1 Jan 2015 → …

Conference

ConferenceProceedings of the American Control Conference
Period1/01/15 → …

Fingerprint

Fuzzy neural networks
system identification
Learning algorithms
learning
Identification (control systems)
Dynamical systems
Data storage equipment
education

Cite this

Rodriguez, F. O., Yu, W., & Moreno-Armendariz, M. A. (2007). Recurrent fuzzy CMAC in hierarchical form for dynamic system identification. 5706-5711. Paper presented at Proceedings of the American Control Conference, . https://doi.org/10.1109/ACC.2007.4282705
Rodriguez, Floriberto Ortiz ; Yu, Wen ; Moreno-Armendariz, Marco A. / Recurrent fuzzy CMAC in hierarchical form for dynamic system identification. Paper presented at Proceedings of the American Control Conference, .5134 p.
@conference{2c9e77b1251e4bf0960d6d0a880ca09c,
title = "Recurrent fuzzy CMAC in hierarchical form for dynamic system identification",
abstract = "The conventional fuzzy CMAC neural networks perform well in terms of their fast learning speed and local generalization capability for approximating nonlinear functions. However, it requires an enormous memory and the dimension increase exponentially with the input number. In this paper, we use two techniques to overcome these problems: recurrent and hierarchical structures and propose a new CMAC, named Hierarchical Recurrent Fuzzy CMAC (HRFCMAC). Since the structure of HRFCMAC is very complex, the normal training methods are difficult to be applied. A new simple algorithm is given, we can train each sub-block of the hierarchical CMAC independently. A time-varying learning rate assures the learning algorithm is stable. {\circledC} 2007 IEEE.",
author = "Rodriguez, {Floriberto Ortiz} and Wen Yu and Moreno-Armendariz, {Marco A.}",
year = "2007",
month = "12",
day = "1",
doi = "10.1109/ACC.2007.4282705",
language = "American English",
pages = "5706--5711",
note = "Proceedings of the American Control Conference ; Conference date: 01-01-2015",

}

Rodriguez, FO, Yu, W & Moreno-Armendariz, MA 2007, 'Recurrent fuzzy CMAC in hierarchical form for dynamic system identification' Paper presented at Proceedings of the American Control Conference, 1/01/15, pp. 5706-5711. https://doi.org/10.1109/ACC.2007.4282705

Recurrent fuzzy CMAC in hierarchical form for dynamic system identification. / Rodriguez, Floriberto Ortiz; Yu, Wen; Moreno-Armendariz, Marco A.

2007. 5706-5711 Paper presented at Proceedings of the American Control Conference, .

Research output: Contribution to conferencePaperResearch

TY - CONF

T1 - Recurrent fuzzy CMAC in hierarchical form for dynamic system identification

AU - Rodriguez, Floriberto Ortiz

AU - Yu, Wen

AU - Moreno-Armendariz, Marco A.

PY - 2007/12/1

Y1 - 2007/12/1

N2 - The conventional fuzzy CMAC neural networks perform well in terms of their fast learning speed and local generalization capability for approximating nonlinear functions. However, it requires an enormous memory and the dimension increase exponentially with the input number. In this paper, we use two techniques to overcome these problems: recurrent and hierarchical structures and propose a new CMAC, named Hierarchical Recurrent Fuzzy CMAC (HRFCMAC). Since the structure of HRFCMAC is very complex, the normal training methods are difficult to be applied. A new simple algorithm is given, we can train each sub-block of the hierarchical CMAC independently. A time-varying learning rate assures the learning algorithm is stable. © 2007 IEEE.

AB - The conventional fuzzy CMAC neural networks perform well in terms of their fast learning speed and local generalization capability for approximating nonlinear functions. However, it requires an enormous memory and the dimension increase exponentially with the input number. In this paper, we use two techniques to overcome these problems: recurrent and hierarchical structures and propose a new CMAC, named Hierarchical Recurrent Fuzzy CMAC (HRFCMAC). Since the structure of HRFCMAC is very complex, the normal training methods are difficult to be applied. A new simple algorithm is given, we can train each sub-block of the hierarchical CMAC independently. A time-varying learning rate assures the learning algorithm is stable. © 2007 IEEE.

UR - https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=46449127839&origin=inward

UR - https://www.scopus.com/inward/citedby.uri?partnerID=HzOxMe3b&scp=46449127839&origin=inward

U2 - 10.1109/ACC.2007.4282705

DO - 10.1109/ACC.2007.4282705

M3 - Paper

SP - 5706

EP - 5711

ER -

Rodriguez FO, Yu W, Moreno-Armendariz MA. Recurrent fuzzy CMAC in hierarchical form for dynamic system identification. 2007. Paper presented at Proceedings of the American Control Conference, . https://doi.org/10.1109/ACC.2007.4282705