Identification of Dynamic Systems by Subspace Methods in Takagi-Sugeno Fuzzy Models
System Identification, Subspace System Identification, MOESP, Fuzzy Logic, Modelo Fuzzy Takagi-Sugeno Model, Neuro-fuzzy, ANFIS.
This work proposes a new approach in the estimation of local ANFIS models using the MOESP subspace identification method. The MOESP method estimates the matrices of the system state space presentation from the extended observation matrix. The pa- per presents a survey of the state of the art of subspace identification method and fuzzy models, it explaining principal characteristics , advantages and disadvantages. In the va- lidation stage, the methodology will be applied in benchmarks known in the literature to present the practical feasibility and the quality of the models. It is expected that the esti- mates obtained will be better than those conventionally employed in fuzzy models in the application of the proposed methodology.