Studi Komparatif Sistem Kendali Bising Aktif Umpan Maju Menggunakan Filter Adaptif Berbasis LMS

In this paper, we discuss simulation and experimental study of the application of Least Mean Square (LMS) based adaptive filter algorithm in active noise control. A number of adaptive filter algorithm based on LMS are investigated, namely Preconditioned Least Mean Square (PLMS), Filtered-reference LMS (FxLMS) and two variants of Filtered-error LMS (FeLMS) algorithms. Active noise control simulation for attenuating recorded blower noise is performed. In case of active noise control experiment, a disturbance is generated by adding three sinus waves with different amplitudes and phases using digital signal processor TMS320C6701, and passing them through primary path. Comparison of the four algorithms is carried out by measuring the performance (in terms of resulting attenuation level) and speed of convergence. Simulation and experiment results demonstrate that PLMS provides the best performance and speed of convergence