Showing papers by "Université du Québec à Montréal published in 1970"

Une contribution au filtrage linéaire optimal non stationnaire par la méthode de Wiener-Booton

Abstract: We show how it is possible to systematize the statistical linear optimization process which uses the Wiener-Booton method. Starting from the differential equation of the linear filter, we obtain its representation by ‘ state-transition ’ matrix and use thia for establishing tho Booton equation. In order to solve this one, we assume the verified Shinbrot hypothosis and we notice that the true identity of the integral equation kernel is an operator instead of a function. We use also the fact that this kernel is separable. We construct a set of equations which puts forward clearly the noise effect in the optimization equation and gives the minimum order of the filter. We give the optimization process for obtaining the filter equation from the Booton equation solution.