Binary constant modulus blind equalization for multi-level QAM signals under Gaussian and non-Gaussian noise

A binary constant modulus algorithm (BCMA) is proposed in this work for the blind equalization of multi-level quadrature amplitude modulation (QAM) systems. A novel cost function and its corresponding iteration formula are developed, which removes both the excess error and the steady state error introduced in the design of the most popular constant modulus algorithm (CMA). Theoretical analysis indicates the proposed algorithm can solve the performance degradation of CMA applying to the non-constant QAM signals. Moreover, it can combat the impulsiveness of alpha stable noise. Simulation results illustrate the effectiveness and robustness of the new algorithm under both Gaussian and non-Gaussian environments.