• Journal of Industrial Technology
• /
• v.30 no.A
• /
• pp.111-117
• /
• 2010

To improve the performance of the prediction system, the system should reflect well the uncertainty of nonlinear data. Thus, this paper presents multiple prediction systems based on Type-2 fuzzy sets. To construct each prediction system, an Interval Type-2 TSK Fuzzy Logic System and difference data were used, because, in general, it has been known that the Type-2 Fuzzy Logic System can deal with the uncertainty of nonlinear data better than the Type-1 Fuzzy Logic System, and the difference data can provide more steady information than that of original data. Also, to improve each rule base of the fuzzy prediction systems, the HCBKA (Hierarchical Correlation Based K-means clustering Algorithm) was applied because it can consider correlationship and statistical characteristics between data at a time. Subsequently, to alleviate complexity of the proposed prediction system, a system selection method was used. Finally, this paper analyzed and compared the performances between the Type-1 prediction system and the Interval Type-2 prediction system using simulations of three typical time series examples.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.34S no.3
• /
• pp.46-54
• /
• 1997

In this paper, we psropose the feedback linearization technique for a nonlinear system using genetic algorithms (GAs) and fuzzy logic system. The proposed control scheme approximates the nonlinear term of a nonlinear system using the fuzzy logic system and computes the control input for cancelling the nonlinear term. Then in the fuzzy logic system, the number and shape of membership function of the premise aprt will be tuned to minimize the control error boundary using GAs. And the parameters of the consequence of fuzzy rule will be tuned by the adaptive laws based on lyapunov stability theory in order to guarantee the closed loop stability of control system. The evolution of fuzzy logic system is processed during the on-line adaptive control. The effectiveness of proposed method will be demonstrated by computer simulation of simple nonlinear sytem.

• Nuclear Engineering and Technology
• /
• v.44 no.4
• /
• pp.363-368
• /
• 2012

A feedwater controller targeted for an ABWR has been implemented using a modern field programmable gate array (FPGA), and verified using the full scope simulator at Taipower's Lungmen nuclear power station. The adopted control algorithm is a rule-based fuzzy logic. Point to point validation of the FPGA circuit board has been executed using a digital pattern generator. The simulation model of the simulator was employed for verification and validation of the controller design under various plant initial conditions. The transient response and the steady state tracking ability were evaluated and showed satisfactory results. The present work has demonstrated that the FPGA based approach incorporated with a rule-based fuzzy logic control algorithm is a flexible yet feasible approach for feedwater controller design in nuclear power plant applications.

• Transactions on Control, Automation and Systems Engineering
• /
• v.3 no.3
• /
• pp.170-175
• /
• 2001

Q-learning is a kind of reinforcement learning where the agent solves the given task based on rewards received from the environment. Most research done in the field of Q-learning has focused on discrete domains, although the environment with which the agent must interact is generally continuous. Thus we need to devise some methods that enable Q-learning to be applicable to the continuous problem domain. In this paper, an extended fuzzy rule is proposed so that it can incorporate Q-learning. The interpolation technique, which is widely used in memory-based learning, is adopted to represent the appropriate Q value for current state and action pair in each extended fuzzy rule. The resulting structure based on the fuzzy inference system has the capability of solving the continuous state about the environment. The effectiveness of the proposed structure is shown through simulation on the cart-pole system.

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.48 no.2
• /
• pp.145-152
• /
• 1999

This paper presents an optimal design method for fuzzy logic controllers using genetic algorithms. In general, the design of fuzzy logic controllers has difficulties in the acquisition of exper's knowledge and relies to a great extent on empirical and heuristic knowledge which, in many cases, cannot be objectively justified. So, the performance of the controller can be degraded in the case of plant parameter variations or unpredictable incident which the designer may have ignored, and parameters of the fuzzy logic controller obtained by expert's control action may not be global. To solve these problems, the proposed method using genetic algorithms in this paper, can tune the parameters of fuzzy logic controller including scaling factors and determine the appropriate number of fuzzy reles systematically and automatically. We provide the second drder dead time plant and inverted pendulum system to evaluate the feasibility and generality of our proposed method. Comparison shows that the proposed controller can producd higher accuracy and a smaller number of fuzzy rules than manually tuned fuzzy logic controller.

• Korean Journal of Cognitive Science
• /
• v.13 no.1
• /
• pp.13-21
• /
• 2002

Fuzzy logic ignores some information in the reasoning process. Neural network is powerful tools for the pattern processing, but, not appropriate for the logical reasoning. To model human knowledge, besides pattern processing capability, the logical reasoning capability is equally important. Another new neural network called neural logic network is able to do the logical reasoning. Because the fuzzy inference is a fuzzy logical reasoning, we construct fuzzy inference network based on the neural logic network, extending the existing rule-inference network. And the traditional propagation rule is modified. Experiments are performed to compare search costs by sequential searching and searching by priority. The experimental results show that the searching by priority is more efficient than the sequential searching as the size of the fuzzy inference network becomes larder and an the number of searching increases.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.06a
• /
• pp.268-271
• /
• 1998

this paper presents the potential application of fuzzy logic to the automatic incident detection system. While the conventional incident detection algorithms are based on a binary decision process, the algorithm using fuzzy logic can incorporate ambiguity which occurs in determining incidents. Since collecting good amount of data to construct data base for incidents is pretty expensive, a traffic simulator called FRESIM is used to simulate traffic condition in a freeway. Incident data are obtained by changing input parameters of the simulator and the fuzzy algorithm generates fuzzy rule for determining normal and incident traffic conditions. In this paper, various steps are described to test the algorithm and its results are summarized.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.1 no.1
• /
• pp.44-49
• /
• 2001

For minimum phase systems, the conventional fuzzy logic controllers (FLCs) use the error and the change-of-error as fuzzy input variables. Then the control rule table is a skew symmetric type, that is, it has UNLP (Upper Negative and Lower Positive) or UPLN property. This property allowed to design a single-input FLC (SFLC) that has many advantages. But its control parameters are not automatically adjusted to the situation of the controlled plant. That is, the adaptability is still deficient. We here design a single-input direct adaptive FLC (SDAFLC). In the AFLC, some parameters of the membership functions characterizing the linguistic terms of the fuzzy rules are adjusted by an adaptive law. The SDAFLC is designed by a stable error dynamics. We prove that its closed-loop system is globally stable in the sense that all signals involved are bounded and its tracking error converges to zero asymptotically. We perform computer simulations using a nonlinear plant and compare the control performance between the SFLC and the SDAFLC.

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.2
• /
• pp.13-17
• /
• 2024

In this paper, union-based rule antecedent neuro-fuzzy controller, which can guarantee a parsimonious knowledge base with reduced number of rules, is proposed. The proposed neuro-fuzzy controller allows union operation of input fuzzy sets in the antecedents to cover bigger input domain compared with the complete structure rule which consists of AND combination of all input variables in its premise. To construct the proposed neuro-fuzzy controller, we consider the multiple-term unified logic processor (MULP) which consists of OR and AND fuzzy neurons. The fuzzy neurons exhibit learning abilities as they come with a collection of adjustable connection weights. In the development stage, the genetic algorithm (GA) constructs a Boolean skeleton of the proposed neuro-fuzzy controller, while the stochastic reinforcement learning refines the binary connections of the GA-optimized controller for further improvement of the performance index. An inverted pendulum system is considered to verify the effectiveness of the proposed method by simulation and experiment.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.3 no.1
• /
• pp.44-51
• /
• 2003

Knowledge acquisition is a bottleneck in knowledge-based system implementation. Decision tree induction is a useful machine learning approach for extracting classification knowledge from a set of training examples. Many real-world data contain fuzziness due to observation error, uncertainty, subjective judgement, and so on. To cope with this problem of real-world data, there have been some works on fuzzy classification rule learning. This paper makes a survey for the kinds of fuzzy classification rules. In addition, it presents a fuzzy classification rule learning method based on decision tree induction, and shows some experiment results for the method.