Gradient method for optimal expansions of MIMO ARX model on Laguerre bases

This paper proposes a representation of linear MIMO systems by using ARX-MIMO Laguerre model where each polynomial function of the classical ARX MIMO model associated to the inputs and the outputs is expanded on independent Laguerre basis. The characteristic pole of which should be optimized. In this context, we propose a pole optimization method based on the gradient algorithm. The proposed pole optimization is validated on numerical example.