A combination of two adaptive algorithms SMI and CMA
Rumiko Yonezawa, Keiji Hirata, T. Kirimoto, Isamu Chiba · 2002
A combination of two major adaptive algorithms, the SMI (sample matrix inversion) method and the CMA (constant modulus algorithm), is presented. The idea is to use the SMI method to determine the initial weights for the CMA operation. Although the merit of the CMA, being a blind algorithm, is not fully taken advantage of, the good aspects of both the SMI and CMA can be introduced. By this approach, two major problems of the CMA can be solved, the ill convergence with false solution when the interfering signal is stronger than the desired signal and its slow convergence time. Simulation results are shown to confirm the theory.