S. Nabiha (France) and M. Hassani (Tunisia)
Modelling, identification, non linear system, SIMO,Volterra, PARAFAC decomposition
This paper extends the results obtained in modeling Single Input Single Output non linear process using tensor decomposition to the case when a Single Input Multi output (SIMO) process is addressed. Despite its non linearity the process is handled by a model linear with respect to its parameters derived by using Volterra series. To alleviate the drawbacks raised from the model parameter number increase, the high order Volterra kernels are described by tensor which can be decomposed into simples matrices to provide a reduced complexity models.
Important Links:
Go Back
