Generalized Volterra and Wiener nonlinear filters for applications in digital Communications

New developments in nonlinear dynamical systems are permitting Volterra and Wiener nonlinear functional series representations to be optimally approximated by artificial neural networks, which are capable of adaptation, learning and evolution by training with or without supervision.The framework based on Generalized Fock Space, enabling these developments is briefly reviewed. In particular, the realization of a best approximation to Wiener’s Laguerre / Hermite representation in terms of a dynamical functional artificial neural network (D-FANN) is presented. Applications of this technology to emerging telecommunications systems such as wireless, digital subscriber lines, and cable TV are discussed.