Shin-ichi Horikawa

On fuzzy modeling using fuzzy neural networks with the back-propagation algorithm

A fuzzy modeling method using fuzzy neural networks with the backpropagation algorithm is presented. The method can identify the fuzzy model of a nonlinear system automatically. The feasibility of the method is examined using simple numerical data.

On identification of structures in premises of a fuzzy model using a fuzzy neural network

The fuzzy neural networks (FNNs) proposed are multilayered backpropagation (BP) models where the structures are designed to realize fuzzy reasoning and to make the connection weights of the networks correspond to the parameters of the fuzzy reasoning. By modifying the connection weights of the network through learning with the BP algorithm, the FNNs can identify the fuzzy rules and tune the membership functions of the fuzzy reasoning automatically. The authors study the tuning of the membership functions in the premises of an FNN using the input-output data for which the characteristics are known, and show that the BP algorithm realizes the appropriate tuning for representing the characteristics of teaching signals. Based on the results of this study, a method is presented to identify the fuzzy models with the minimal number of the membership functions in the premises.<<ETX>>

Composition methods of fuzzy neural networks

Fuzzy neural networks (FNNs) are systems which apply neural networks to fuzzy reasoning. Two types of FNN are presented. In the first type, the consequences of fuzzy reasoning are realized by constants. In the second type, the consequences are expressed by first-order linear equations. The FNNs can automatically identify fuzzy rules and tune membership functions. Their performance on fuzzy reasoning is examined by simulations. The features of the two types of FNNs are clarified.<<ETX>>

Knowledge acquisition of strategy and tactics using fuzzy neural networks

A knowledge acquisition method using fuzzy neural networks (FNNs) is presented. The FNNs automatically acquire control strategy and tactics from the operators manipulating data. The acquired strategy and tactics coincide well with the operators image of strategic and tactical aspects of his controls. Steering of a ship with large inertia for passing through multiple gates is simulated. The FNNs extract the strategic and tactical rules from the operators steering data. A hierarchical controller, using the FNN with strategy for generating track commands and the FNN with tactics for steering the ship to follow the commands, is developed. The hierarchical controller exhibits good performance in steering the ship. The ship is able to pass through even untrained gate patterns.<<ETX>>

On design of adaptive fuzzy controller using fuzzy neural networks and a description of its dynamical behavior

Abstract This paper presents a design method of adaptive fuzzy controller. The fuzzy controller is designed with linguistic rules of fuzzy models of controlled objects. The fuzzy model is identified with a fuzzy neural network (FNN). This paper also presents a “rule-to-rule mapping” method for describing the dynamical behavior of the designed control system. The new method uses the fuzzy rules of fuzzy control system and enables the description of the dynamical behavior using the language of the rules. Modification of control rules becomes easy with the mapping method. The designed fuzzy controller can be implemented with another FNN. An adaptive tuning of the control rules of the FNN controller is made possible utilizing the fuzzy model.

A new linguistic design method of fuzzy controller using a description of dynamical behavior of fuzzy control systems

This paper presents a new linguistic design method of fuzzy controllers using a description of the dynamical behavior of fuzzy control systems. The required response of the control systems can be given by linguistic expressions. The feasibility of the new method is examined by experiments with an inverted pendulum. The results show that the method is effective to satisfy the linguistic requirements.

A new type of fuzzy neural network based on a truth space approach for automatic acquisition of fuzzy rules with linguistic hedges

Abstract Fuzzy reasoning methods are generally classified into two approaches: the direct approach and the truth space approach. Several researches on the relationships between these approaches have been reported. There has been, however, no research which discusses their utility. The authors have previously proposed four types of fuzzy neural networks (FNNs) called Type I, II, III, and IV. The FNNs can identify the fuzzy rules and tune the membership functions of fuzzy reasoning automatically, utilizing the learning capability of neural networks. Types III and IV, which are based on the truth space approach, can acquire linguistic fuzzy rules with the fuzzy variables in the consequences labeled according to their linguistic truth values (LTVs). However, the expressions available for the linguistic labeling are limited, since the LTVs are singletons. This paper presents a new type of FNN based on the truth space approach for automatic acquisition of the fuzzy rules with linguistic hedges. The new FNN, called Type V, has the LTVs defined by fuzzy sets for fuzzy rules and can express the identified fuzzy rules linguistically using the fuzzy variables in the consequences with linguistic hedges. Two simulations are done for demonstrating the feasibility of the new method. The results show that the truth space approach makes the fuzzy rules easy to understand.

An application of fuzzy neural networks to design of adaptive fuzzy controllers

This paper presents a design method of adaptive fuzzy controllers using fuzzy neural networks (FNNs). The adaptive controller uses two FNNs. One FNN is used to identify a fuzzy model of the controlled object. The fuzzy control rules are designed with the linguistic rules of the fuzzy model and are incorporated into the other FNN. The response of the designed control system is checked with a linguistic response analysis proposed by the authors. An adaptive tuning of the control rules of the FNN controller is made possible utilizing the fuzzy model. Simulations using a controlled object with non-monotonic input-output characteristics are done to verify the proposed design method.

An application of fuzzy neural networks to a stability analysis of fuzzy control systems

The authors present a method for analyzing the stability of fuzzy control systems using fuzzy neural networks (FNNs). The FNNs are capable of acquiring fuzzy rules and tuning the membership functions automatically with the backpropagation (BP) learning algorithm. One FNN is used to obtain a fuzzy model of the controlled object. The other FNN is trained to acquire a fuzzy model of the controller. A new definition of the stability of fuzzy control systems is also presented. Using the fuzzy controller and the fuzzy model obtained for the controlled object, a stability analysis based on the proposed definition is done linguistically and is very easy to understand.<<ETX>>

On stability of fuzzy control systems using a fuzzy modeling method

The authors present an analysis method on stability of fuzzy control systems using a fuzzy modeling method. A fuzzy neural network (FNN) which is capable of acquiring fuzzy rules is used for obtaining the fuzzy model of the controller object. A new definition of the stability of fuzzy control systems is also presented. Since both the controlled object and the controller are described with linguistic rules, i.e., fuzzy if-then rules, the stability analysis can be done linguistically. Although the new definition of stability is not mathematically rigorous, the process of analysis based on the definition is very simple and easy to understand, even to those who have little knowledge of control theory. For the verification of the proposed method, simulations were carried out by using a simple nonlinear plant.<<ETX>>