Comprehensive analysis of LMS and NLMS algorithms using adaptive equalizers

The main emphasis in this paper lies on evaluation of adaptive algorithms which are predominantly used. They are LMS and NLMS algorithms respectively. Initially these algorithms are applied on adaptive equalizers. Then we further analysed and compared two types of LMS algorithms namely sign-sign LMS algorithm and sign-sign LMS algorithm. The overlap of coefficients is very high when step-size is very small. We found that LMS algorithm is simple and gives good results if proper step size is chosen and NLMS alternates step-size according to energy of input signal hence it is suitable for both stationary as well as non-stationary environment. Regarding sign-data and sign-sign LMS algorithms we found that the former is better than the latter as deviations and amplitude are very high in the latter case.