Optimal expansion of discrete Volterra models using independent Laguerre basis

In this paper, the development of Volterra kernels on independent Laguerre basis is considered to provide Laguerre Volterra model. To minimise such model truncating order, the optimal Laguerre poles have to be identified. Two procedures based on gradient and Newton-Raphson methods are proposed. These methods are tested on simulation and the resulting poles converge to their optimal values. The yielded Laguerre-Volterra models are validated and their outputs fit the process one.