BLIND EQUALIZATION USING HIGHER-ORDER CUMULANTS AND NEURAL-NETWORK

This paper develops a new method to achieve blind equalization in digi tal communication for linear finite impulse response (FIR) systems, wh ether the systems are minimum phase or not. This new approach divides the problem into two parts. First, it employs the characteristic of th e linear system to estimate the original channel based on the fourth-o rder cumulants instead of time samples of the channel output. Thus, no nminimum phase channels can be handled. Second, it utilizes the nonlin ear characteristics of the neural network to build an inverse system ( equalizer) for the original channel. This is done by using the estimat ed channel as a reference system to train the neural network. The neur al network helps the equalizer to reduce the degree of model uncertain ty and makes the equalizer resistant to additive noise. Taking the adv antages of both linear and nonlinear systems, this new scheme works we ll for both stationary and nonstationary cases and leads to good equal ization results.