Variable step-size least mean square algorithms over adaptive networks

This paper presents two new variations for algorithms over adaptive networks using a variable step-size strategy in order to enhance the overall performance. Variable step-size least mean square (VSSLMS) algorithms for incremental as well as diffusion strategies are studied and the results are compared with existing results. A comparison is done with the recently proposed Diffusion LMS algorithm with adaptive combiners and it is shown that VSSLMS provides a simplified solution than that of the Diffusion LMS algorithm with adaptive combiners. Also, a great improvement in performance is obtained when compared with the fixed step-size Incremental and Diffusion LMS techniques.