Título: Nonlinear System Identification Using Local ARX Models Based On The Self-Organazing Map
Autores: Souza, Luís Gustavo M.; Barreto, Guilherme A.
Resumo: In this paper, we build multiple local ARX (Auto-Regressive model with eXogenous variables) models for nonlinear system identification based on Kohonen’s Self-Organizing Map. These models are evaluated in the task of approximating the inverse dynamics of two benchmarking input-output data sets: a hydraulic actuator and a flexible robotic arm. Simulation results demonstrate that the proposed multiple local ARX modeling approach outperforms both single ARX model and MLP-based NARX model approaches in the aforementioned tasks.
Palavras-chave: Kohonen map; local linear models; NARX models; system identification; hydraulic actuator; robot arm
Páginas: 12
Código DOI: 10.21528/lmln-vol4-no2-art2
Artigo em PDF: vol4-no2-art2.pdf
Arquivo BibTex: vol4-no2-art2.bib