Blind adaptive channel equalization using the unsupervised cluster formation technique

Introduces a novel approach to decision-directed (DD) blind equalization. This approach is based on a neural network classification technique, and it is shown that the final result is closely related to previously published DD equalization algorithms. The authors briefly review the history of DD blind equalization, provide the derivation of the new algorithm, and present simulation results. In the simulations, the performance of the proposed algorithm is illustrated by applying it to a 2D digital mobile communication system where DD blind equalization algorithms find applications in equalizing and tracking the time-varying channels. A time-varying multipath fading channel model is used as the transmission medium. The proposed blind equalization algorithm is compared to the commonly used fast recursive least squares decision-feedback equalization algorithm. The comparison illustrates the improvement in performance achievable with the new decision-directed algorithm.