A variable step size Square Contour Algorithm based on a novel non-linear function of error signal

The Square Contour Algorithm (SCA) solves the problem caused by possible diagonal solutions produced by the Multi-Modulus Algorithm (MMA). However, it suffers from convergence problem. In this paper, we propose a variable step size approach for improving the convergence of the SCA. Lower ISI floors are achieved by a step size controlled by a novel nonlinear function of the error signal, which assigns higher step sizes in initial phases of convergence and smaller step size in the later phases and using separate step sizes for in-phase and quadrature weights depending on the magnitude of real and imaginary errors.