An Approach to Online Identification of Takagi-Sugeno Fuzzy Models

Abstract

Abstract-An approach to the online learning of Takagi-Sugeno (TS) type models is proposed in the paper. It is based on a novel learning algorithm that recursively updates TS model structure and parameters by combining supervised and unsupervised learning. The rule-base and parameters of the TS model continually evolve by adding new rules with more summarization power and by modifying existing rules and parameters. In this way, the rule-base structure is inherited and up-dated when new data become available. By applying this learning concept to the TS model we arrive at a new type adaptive model called the Evolving Takagi-Sugeno model (ETS). The adaptive nature of these evolving TS models in combination with the highly transparent and compact form of fuzzy rules makes them a promising candidate for online modeling and control of complex processes, competitive to neural networks. The approach has been tested on data from an air-conditioning installation serving a real building. The results illustrate the viability and efficiency of the approach. The proposed concept, however, has significantly wider implications in a number of fields, including adaptive nonlinear control, fault detection and diagnostics, performance analysis, forecasting, knowledge extraction, robotics, behavior modeling.

Autoregressive time series prediction by means of fuzzy inference systems using nonparametric residual variance estimation

“…The second problem is addressed by data-driven techniques for identification of fuzzy systems from numerical examples [16], such as the algorithm by Wang and Mendel (W&M) [50,51] and the fuzzy identification algorithm based on clustering by Chiu [12,36], two well established methods among the many alternatives proposed throughout the years [23,19,43,34,2,26,44].…”

The Success of Smart Growth of a City

“…Here we use the Takagi-Sugeno fuzzy neural network algorithm [5] to modeling. Enter the five indexes and output the Smart Growth Status Metric.…”

A neuro-fuzzy self built system for prognostics: a way to ensure good prediction accuracy by balancing complexity and generalization

“…A first order TS model can be seen as a multi-model structure consisting of linear models that are not necessarily independent [16]. It is based on the fuzzy decomposition of the input space.…”

“…Thereby, if the behavior of input and/or output changes significantly with regards to the learning phase (like in a degradation process), predictions can suffer from the lack of representative learning data. In order to continuously integrate the dynamic of signals, evolving algorithms have finally been developed [16,19]. These algorithms are based on clustering methods and therefore, do not require the user to define the structure of the TS model.…”

“…A particular evolving TS model is that one proposed by [16,20]: the evolving extended TS system (exTS).…”