Hiroyoshi Nomura

A learning method of fuzzy inference rules by descent method

The authors propose a learning method for fuzzy inference rules by a descent method. From input-output data gathered from specialists, the inference rules expressing the input-output relation of the data are obtained automatically. The membership functions in the antecedent part and the real number in the consequent part of the inference rules are tuned by means of the descent method. The learning speed and the generalization capability of this method are higher than those of a conventional backpropagation type neural network. This method has the capability to express the knowledge acquired from input-output data in the form of fuzzy inference rules. Some numerical examples are described to show these advantages over the conventional neural network. An application of the method to a mobile robot that avoids a moving obstacle and its computer simulation are reported. >

Construction of fuzzy inference rules by NDF and NDFL

Abstract Whereas conventional fuzzy reasoning lacks determining membership functions, a neural network driven fuzzy reasoning (NDF) capable of determining membership functions uniquely by an artificial neural network is formulated. In an NDF algorithm the optimum membership function in the antecedent part of fuzzy inference rules is determined by a neural network, while in the consequent parts an amount of reasoning for each rule is determined by other plural neural networks. On the other hand, we propose a new algorithm that can adjust inference rules to compensate for a change of inference environment. We call this algorithm a neural network driven fuzzy reasoning with learning function (NDFL). NDFL can determine the optimal membership function and obtain the coefficients of linear equations in the consequent parts by the searching function of the pattern search method. In this paper, inference rules for making a pendulum stand up from its lowest suspended point ar3 determined by the NDF algorithm for verifying its effectiveness. The NDFL algorithm is formulated and applied to a simple numerical example to demonstrate its effectiveness.

Segmentation of thermal images using the fuzzy C-means algorithm

A segmentation methodology based on the fuzzy clustering algorithm is developed. The algorithm is utilized to segment a thermal image of occupants in a room taken by a thermoviewer. The purpose of segmentation is to identify the number and the positions of the occupants. Some useful applications can be realized, such as control of air-conditioning systems, security systems, and so on. The approach consists of two stages. The first stage is to distinguish occupants from a background in an image using the fuzzy C-means (FCM) algorithm. The authors have selected a suitable measure for determining the number of clusters and modified it for FCM. The purpose of the second stage is to distinguish each occupant by locating local temperature peaks in the image. A region-growing algorithm is introduced for more accurate segmentation based on the membership value determined by FCM and the number of located peaks. Some experimental results are included that relate to thermal images obtained in a meeting room.<<ETX>>

Implementation of multistep fuzzy reasoning shell

Recently, there have been many trials conducted to develop fuzzy inference systems that can handle the subjective decision of the expert. A problem in the fuzzy control system is that it is an inference composed of only a single step, and cannot be applied to the decision-making assist or diagnosis, where the inference is made in multisteps. Another point is that the rules of the inference cannot be described for each knowledge source, which has been the conventional inference shell (deterministic inference shell). This paper proposes a fuzzy inference shell which can realize the multistep inference using the blackboard model. The proposed shell uses its own rule description language to represent the knowledge, and the description can be made by combining the new method with the conventional deterministic inference. Using the divided knowledge blocks, the multi-step fuzzy inference can be realized using the blackboard model. The search function also is provided that can track the inference path in the multistep inference. This paper describes the basic configuration of such a shell.