Recurrent fuzzy logic using neural network

Khan, E.; Ünal, Fatih

doi:10.1007/3-540-60607-6_4

Cited by 11 publications

(11 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A variety of implementations of FFA have been proposed, some in digital systems [39], [40]. However, here we give a proof that such implementations in sigmoid activation RNN's are stable, i.e., guaranteed to converge to the correct prespecified membership.…”

Section: B Backgroundmentioning

confidence: 87%

“…There has been much work on the learning, synthesis, and extraction of finite state automata in recurrent neural networks (see, for example, [46]- [53]). A variety of neural network implementations of FFA have been proposed [39], [40], [54], [55]. We have previously shown how fuzzy finite state automata can be mapped into recurrent neural networks with second-order weights using a crisp representation 2 of FFA states [56].…”

Section: B Backgroundmentioning

confidence: 99%

See 1 more Smart Citation

Equivalence in knowledge representation: automata, recurrent neural networks, and dynamical fuzzy systems

1999

View full text Add to dashboard Cite

show abstract

Section: B Backgroundmentioning

confidence: 87%

Section: B Backgroundmentioning

confidence: 99%

Equivalence in knowledge representation: automata, recurrent neural networks, and dynamical fuzzy systems

1999

View full text Add to dashboard Cite

show abstract

“…Two examples of such a synergy are the adaptive network-based fuzzy inference system (ANFIS) [22], which is a feedforward network representation of the fuzzy reasoning process (see [34] for an extension to RNNs), and the fuzzy-MLP, which is a feedforward network with fuzzified inputs [40], [46]. Neuro-fuzzy systems have been intensively discussed in the literature (see the survey [39]), but rarely do they contain feedback connections [23].…”

Section: Introductionmentioning

confidence: 99%

A New Approach to Knowledge-Based Design of Recurrent Neural Networks

Kolman

Margaliot

2008

IEEE Trans. Neural Netw.

View full text Add to dashboard Cite

Abstract-A major drawback of artificial neural networks (ANNs) is their black-box character. This is especially true for recurrent neural networks (RNNs) because of their intricate feedback connections. In particular, given a problem and some initial information concerning its solution, it is not at all clear how to design an RNN that is suitable for solving this problem.In this paper, we consider a fuzzy rule-base with a special structure, referred to as the fuzzy all-permutations rulebase (FARB). Inferring the FARB yields an input-output mapping that is mathematically equivalent to that of an RNN.We use this equivalence to develop two new knowledge-based design methods for RNNs. The first method, referred to as the direct approach, is based on stating the desired functioning of the RNN in terms of several sets of symbolic rules, each one corresponding to a sub-network. Each set is then transformed into a suitable FARB.The second method is based on first using the direct approach to design a library of simple modules, such as counters or comparators, and realize them using RNNs. Once designed, the correctness of each RNN can be verified. Then, the initial design problem is solved by using these basic modules as building blocks. This yields a modular and systematic approach for knowledge-based design of RNNs. We demonstrate the efficiency of these approaches by designing RNNs that recognize both regular and non-regular formal languages.

show abstract

“…Parameter tuning concerns with adjustments of the position and the shape of membership functions (fuzzy weights). Other existing feedback fuzzy neural approaches (e.g., [18,26]) concentrated on parameter tuning, selecting the structure on a trial-and-error basis. However, there are no simple ways to determine in advance the minimal size of the partition of the input's universe of discourse or the minimal size of the hidden (rule) layer necessary to achieve the desired performance.…”

Section: Introductionmentioning

confidence: 99%

A dynamically-constructed fuzzy neural controller for direct model reference adaptive control of multi-input-multi-output nonlinear processes

Frayman¹,

Wang²

2002

Soft Computing - A Fusion of Foundations, Methodologies and App

View full text Add to dashboard Cite

Conventional industrial control systems are in majority based on the single-input-single-output design principle with linearized models of the processes. However, most industrial processes are nonlinear and multivariable with strong mutual interactions between process variables that often results in large robustness margins, and in some cases, extremely poor performance of the controller. To improve control accuracy and robustness to disturbances and noise, new design strategies are necessary to overcome problems caused by nonlinearity and mutual interactions. We propose to use a dynamicallyconstructed, feedback fuzzy neural controller (DCF-FNC) from the input±output data of the process and a reference model, for direct model reference adaptive control (MRAC) to deal with such problems. The effectiveness of our approach is demonstrated by simulation results on a real-world example of cold mill thickness control and is compared with the performances of the conventional PID controller and the cascade correlation neural network (CCN). Exploiting the advantage of intelligent adaptive control, both the CCN and our DCF-FNC signi®cantly increases the control precision and robustness, compared to the linear PID controller, with our DCF-FNC giving the best results in terms of both accuracy and compactness of the controller, as well as being less computationally intensive than the CCN. We argue that our DCF-FNC feedback controller with both structure and parameter learning can provide a computationally ef®cient solution to control of many real-world multivariable nonlinear processes in presence of disturbances and noise.Keywords Feedback fuzzy neural controller, Dynamic controller structure, Direct MRAC, Cold mill thickness control IntroductionIndustrial processes are often multivariable and have strongly nonlinear and time varying behavior. There also exist strong mutual interactions between process variables. Conventional control of industrial processes usually uses simple linear or linearized models to approximate the processes. For multi-input-multi-output (MIMO) processes it is normally very dif®cult to derive accurate models, due to complex nonlinear relationships among variables, time dependent changes in process dynamics, and model uncertainties when dealing with some physical phenomena. The severe nonlinearity and complexity of the processes thus result in large robustness margins, and in some cases, extremely poor performance of the controller. It is therefore necessary to develop solid control methodologies that are capable of coping with both nonlinearities and interactions, as well as time varying processes under the strong in¯uence of disturbances and noise. In addition, nonlinear control schemes that employ more realistic and more complex descriptions for nonlinear processes require process models in the form of nonlinear differential equations. This limits its industrial applications, since such ®rst principles models are not readily available in industrial practice due to a chronic lack of detailed and ext...

show abstract

Recurrent fuzzy logic using neural network

Cited by 11 publications

References 5 publications

Equivalence in knowledge representation: automata, recurrent neural networks, and dynamical fuzzy systems

Equivalence in knowledge representation: automata, recurrent neural networks, and dynamical fuzzy systems

A New Approach to Knowledge-Based Design of Recurrent Neural Networks

A dynamically-constructed fuzzy neural controller for direct model reference adaptive control of multi-input-multi-output nonlinear processes

Contact Info

Product

Resources

About