Decision feedback equalization for Volterra systems-a root method

The need to recover the original input from the output of a nonlinear system is a problem common to many applications. In this paper, we propose a Volterra decision feedback equalizer that is based on finding the roots of a polynomial function of the input. It is capable of equalizing severely nonlinear systems under the assumption that the input comes from a finite symbol set. A modification is also proposed which allows for the extension of the root method to baseband Volterra systems with complex inputs and the analysis of the effects of noise on the current symbol estimate. Simulations demonstrate the performance of the proposed algorithm.