Blind channel equalization and ϵ-approximation algorithms

We show that a blind equalizer can be obtained without estimating any statistics based on the received data if the constellation alphabet is known. The idea is to formulate the blind channel equalization problem into a quadratic optimization with binary constraints. Then, efficient ε-approximation algorithms are presented and applied to find the equalizer coefficients. The proposed approach blindly equalizes the channel at symbol baud rates, does not rely on estimating any statistics, applies to nonminimum phase channels, and converges to an ε neighborhood of a global optimum. Simulation results support the theoretical analysis