An perturbational blind channel and signal estimation

Abstract

A perturbative blind channel equalization method is introduced. It is shown that through the use of perturbation analysis, both the channel and input signal can be obtained blindly in the presence of noise. The basic idea behind this development is the modelling of the received data sequence as an autoregressive moving average process and performing perturbation of a analysis over the power spectral density. Although the new approach is derived based on second order properties it can restore the channel phase even in low signal-to-noise ratios.