An NLMS algorithm with variable step-size updates through selective inputs

This paper presents a variable step-size normalized least-mean-square (VSS-NLMS) algorithm using only one parameter. Numerous VSS-NLMS algorithms lead to fast convergence rate and low misadjustment. However, they depend on several parameters that are not simple to tune on practice. The proposed VSS-NLMS algorithm leads to fast convergence rate and low misadjustment although the algorithm uses only one parameter. Further, the parameter can be easily tuned through input data.