• The KIPS Transactions:PartB
• /
• v.8B no.2
• /
• pp.123-129
• /
• 2001

본 연구에서는 전문가와 운전자의 제어 지식을 더 정확하게 표현하여 퍼지 논리 제어기의 성능을 향상시킬 수 있는 소속함수 수정 알고리즘을 제안한다. 제안된 알고리즘은 제어지식을 더 정확히 표현할 수 있도록 직관적인 지식과 경험으로부터 유추된 대략적인 제어지식을 평가기준으로 하고 입출력 데이터 클러스터링에 의해 소속함수의 형태와 위치를 수정한다. 제안된 방법을 수위 조절 모델과 교통신호 제어 모델에 적용한 실험을 통해서, 제안된 알고리즘이 기존 제어기의 성능을 향상시킬 수 있고, 퍼지 제어기에서 언어적 변수에 대한 구간 설정의 어려움을 해결할 수 있음을 알 수 있었다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2001.04a
• /
• pp.309-312
• /
• 2001

본 연구에서는 상류부 교차로에서 발생하는 교차로 막힘 현상으로 인해 진행방향의 녹색시간의 손실이라는 장애가 발생하게되는 상황을 고려하기 위해 진행차선의 정체도를 도입하여 교통 혼잡상황에 적절히 대응할 수 있는 퍼지 교통신호 제어기를 제안한다. 먼저 입출력 공간을 균등 분할한 퍼지 교통신호 제어기를 구성하고, 소속함수 수정알고리즘에 의해 제어기를 수정한다. 실험을 통해 고정식 제어기, 균등 분할한 제어기와 수정된 제어기의 성능을 교차로 지체시간, 진입율과 통과율 면에서 비교하였다. 실험 결과는 수정된 제어기가 다른 제어기들에 비해 향상된 성능을 보여주었다.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2001.05a
• /
• pp.43-46
• /
• 2001

We, in this paper, design the Sugeno-models fuzzy controller by using the membership function modification algorithm and ANFIS, which are clustering and learning the input-output data. The membership function modification algorithm constructs the more concrete fuzzy controller by clustering the input-output data from the fuzzy inference system. ANFIS construct the Sugeno-models fuzzy controller by learning the input-output data from the above controller. We showed that the fuzzy controller designed by our method could have the stable learning and the enhanced performance.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.11 no.3
• /
• pp.241-245
• /
• 2001

The existing fuzzy traffic controllers construct the rule-base based on the intuitive knowledge and experience or the standard rule-base, but the rule-base constructed by the above methods has difficulty in representing exactly and detailedly the control knowledge of the export and the operator. Therefore, in this paper, we propose a method that can improve the performance of the fuzzy traffic control by designing the fuzzy traffic controller which represents the control knowledge more exactly. The proposed method so modifies the position and shape of the fuzzy membership function based on the input-output data clustering that the fuzzy traffic controller can represent the control knowledge more exactly. Our method use the rough control knowledge based on intuitive knowledge and experience as the evaluation function for clustering the input-output data. The fuzzy traffic controller designed by the our method could represent the control knowledge of the expert and the operator more exactly, and it outperformed the existing controller in terms of the number of passed vehicles and the wasted green-time.

• The Transactions of the Korea Information Processing Society
• /
• v.1 no.3
• /
• pp.319-326
• /
• 1994

Automatic control has been for the most part applied to linear systems where ti can be approximately formalized. In case that it is not definitely established the mathematical modelling to control objects, it requires manual control strategies which put under the human rule. In this paper, it constructs an FLC (Fuzzy Logic Controller) in order to turn a hand control into an automatic control in the domain of swimming pool that has been almost absolutely dependant on a skilled worker's experience. Genetic algorithms upgrade the knowledge which is acquired from human expert, using by FLC, so as to maintain knowledge in the very optimal way. It also designs an algorithm that modifies the rule base and the membership function at the same time, and ultimately will show that it can get better result than human controllers.

• Journal of KIISE:Software and Applications
• /
• v.33 no.8
• /
• pp.692-702
• /
• 2006

In this study, a weighted fuzzy min-max (WFMM) neural network model for pattern classification is proposed. The model has a modified structure of FMM neural network in which the weight concept is added to represent the frequency factor of feature values in a learning data set. First we present in this paper a new activation function of the network which is defined as a hyperbox membership function. Then we introduce a new learning algorithm for the model that consists of three kinds of processes: hyperbox creation/expansion, hyperbox overlap test, and hyperbox contraction. A weight adaptation rule considering the frequency factors is defined for the learning process. Finally we describe a feature analysis technique using the proposed model. Four kinds of relevance factors among feature values, feature types, hyperboxes and patterns classes are proposed to analyze relative importance of each feature in a given problem. Two types of practical applications, Fisher's Iris data and Cleveland medical data, have been used for the experiments. Through the experimental results, the effectiveness of the proposed method is discussed.