Scribd
Upload
34 views2 pages

Type 2 ANFIS

This document discusses using a hybrid learning algorithm called recursive least squares - backpropagation (RLS-BP) to train an interval type-1 non-singleton type-2 Takagi-Sugeno-Kang fuzzy logic system (IT2 NS-1 ANFIS). The proposed approach estimates antecedent and consequent parameters during training, unlike previous methods that exclusively used backpropagation. As a case study, the model is used to predict transfer bar surface temperatures at a steel mill. Comparative results show the performance of the hybrid RLS-BP method versus only using backpropagation.

Uploaded by

BALASUNDAR C
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
34 views2 pages

Type 2 ANFIS

This document discusses using a hybrid learning algorithm called recursive least squares - backpropagation (RLS-BP) to train an interval type-1 non-singleton type-2 Takagi-Sugeno-Kang fuzzy logic system (IT2 NS-1 ANFIS). The proposed approach estimates antecedent and consequent parameters during training, unlike previous methods that exclusively used backpropagation. As a case study, the model is used to predict transfer bar surface temperatures at a steel mill. Comparative results show the performance of the hybrid RLS-BP method versus only using backpropagation.

Uploaded by

BALASUNDAR C
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Interval Type-2 ANFIS

Gerardo M. Mendez1 and Ma. De Los Angeles Hernandez2


1
Instituto Tecnologico de Nuevo Leon
Department of Electronics and Electrical Engineering
Av. Eloy Cavazos 2001, Cd. Guadalupe, NL, Mexico, CP 67170
[email protected]
2
Instituto Tecnologico de Nuevo Leon
Department of Economic Sciences
Av. Eloy Cavazos 2001, Cd. Guadalupe, NL, Mexico, CP 67170
[email protected]

Abstract. This article presents a new learning methodology based on a hybrid algorithm for
interval type-1 non-singleton type-2 TSK fuzzy logic systems (FLS). Using input-output data
pairs during the forward pass of the training process, the interval type-1 non-singleton type-2
TSK FLS output is calculated and the consequent parameters are estimated by the recursive
least-squares (RLS) method. In the backward pass, the error propagates backward, and the
antecedent parameters are estimated by the back-propagation (BP) method. The proposed
hybrid methodology was used to construct an interval type-1 non-singleton type-2 TSK fuzzy
model capable of approximating the behavior of the steel strip temperature as it is being rolled
in an industrial Hot Strip Mill (HSM) and used to predict the transfer bar surface temperature at
finishing Scale Breaker (SB) entry zone. Comparative results show the performance of the
hybrid learning method (RLS-BP) against the only BP learning.

Keywords: Interval type-2 fuzzy logic systems, ANFIS, neuro fuzzy systems, hybrid learning.

1 Introduction

Interval type-2 (IT2) fuzzy logic systems (FLS) constitute an emerging technology. In
[1] both, one-pass and back-propagation (BP) methods are presented as IT2 Mamdani
FLS learning methods, but only BP is presented for IT2 Takagi-Sugeno-Kang (TSK)
FLS systems. One-pass method generates a set of IF-THEN rules by using the given
training data one time, and combines the rules to construct the final FLS. When BP
method is used in both Mamdani and TSK FLS, none of antecedent and consequent
parameters of the IT2 FLS are fixed at starting of training process; they are tuned
using exclusively steepest descent method. In [1] recursive least-squares (RLS) and
recursive filter (REFIL) algorithms are not presented as IT2 FLS learning methods.
The hybrid algorithm for IT2 Mamdani FLS has been already presented elsewhere
[2, 3, 4] with three combinations of learning methods: RLS-BP, REFIL-BP and
orthogonal least-squares-BP (OLS-BP), whilst the hybrid algorithm for singleton IT2
TSK FLS (IT2 ANFIS) has been presented elsewhere [5] with two combinations of
learning methods: RLS-BP and REFIL-BP.

E. Corchado et al. (Eds.): Innovations in Hybrid Intelligent Systems, ASC 44, pp. 64–71, 2007.
springerlink.com © Springer-Verlag Berlin Heidelberg 2007
Interval Type-2 ANFIS 65

The aim of this work is to present and discuss a new hybrid learning algorithm for
interval type-1 non-singleton type-2 TSK FLS (IT2 NSFLS-1 or IT2 NS-1 ANFIS)
using RLS-BP combination in order to estimate the antecedent and consequent
parameters during the training process. The proposed IT2 NS-1 ANFIS system is
evaluated making transfer bar surface temperature predictions at Hot Strip Mill
(HSM) Finishing Scale Breaker (SB) entry zone.

2 Problem Formulation

Most of the industrial processes are highly uncertain, non-linear, time varying and
non-stationary [2, 6], having very complex mathematical representations. IT2 NS-1
ANFIS takes easily the random and systematic components of type A or B standard
uncertainty [7] of industrial measurements. The non-linearities are handled by FLS as
identifiers and universal approximators of nonlinear dynamic systems [8, 9, 10, 11].
Stationary and non-stationary additive noise is modeled as a Gaussian function
centred at the measurement value. In stationary additive noise the standard deviation
takes a single value, whereas in non-stationary additive noise the standard deviation
varies over an interval of values [1]. Such characteristics make IT2 NS-1 ANFIS a
powerful inference system to model and control industrial processes.
Only the BP learning method for IT2 TSK SFLS has been proposed in the
literature and it is used as a benchmark algorithm for parameter estimation or systems
identification [1]. To the best knowledge of the authors, IT2 NS-1 ANFIS has not
been reported in the literature [1, 12].
One of the main contributions of this work is to implement an application of the
IT2 NS-1 ANFIS using the hybrid REFIL-BP learning algorithm, capable of
compensates for uncertain measurements.

3 Problem Solution

3.1 Using Hybrid RLS-BP Method in IT2 NS-1 ANFIS Training

The Table 1 shows the activities of the one pass learning algorithm of BP method.
Both, IT2 TSK SFLS (BP) and IT2 TSK NSFLS-1 (BP) outputs are calculated
during forward pass. During the backward pass, the error propagates backward
and the antecedent and consequent parameters are estimated using only the BP
method.

Table 1. One pass in learning procedure for IT2 TSK (BP)

Forward Pass Backward Pass


Antecedent Parameters Fixed BP
Consequent Parameters Fixed BP

You might also like