PERBANDINGAN KINERJA ALGORITMA LMS DAN NLMS UNTUK PEREDAMAN DERAU SECARA ADAPTIF

Abstract

We often feel disturbed by noise caused by our surrounding environment when we are in communication. An example of the noise is very loud vehicle sound in highway that interfere our communication. We may lose our stated information if the noise is not cancelled. This noise problem could be solved by adaptive filter. In noise cancelling, filter parameters are set to optimize information signal so it will be freed from noise. In this final project, performances of two algorithms were compared, namely Least Mean Square (LMS) and Normalized Least Mean Square (NLMS). Both algorithms will be used to renew adaptive filter coefficient for noise cancelling. Two performance parameters that were used are Mean Square Error (MSE) and correlation coefficient. Error quadratic curves were graphed to explain the adaptation process of both algorithms. Test results show that filter reaches optimal condition at certain step size and filter length. Filtering with NLMS algorithm is better than LMS algorithm at same step size and filter length. It is proved by lower MSE value, higher correlation coefficient and faster adaptation process than filtering with LMS algorithm.