Adaptive step size sign algorithm for application to adaptive filtering in digital QAM communications

A new adaptive step size control algorithm is proposed to be combined with the sign algorithm for use in complex-valued adaptive filters for application to QAM communications. The algorithm, ASSSA, is fully analyzed to yield a set of difference equations for calculating the transient behavior, hence the steady-state performance, of the filter convergence in terms of excess mean squared error (EMSE). An approximation method for multilevel QAM is further proposed to reduce the amount of computation. The results of experiment with some examples verify the effectiveness of the proposed ASSSA in significantly improving the convergence rate, and also show that the theoretical convergence is in good agreement with that of simulations, which validates the analysis