Analysis of MNLMS and KLMS Algorithm for Underwater Acoustic Communications

Abstract

Communicated by Laszlo B. KishThe use of adaptive filters to alleviate the degradation caused by wind driven ambient noise in shallow water is considered in this paper. Since, underwater acoustic signals are greatly affected by the ocean interference and ambient noise disturbances when propagating through underwater channels, an effective adaptive filtering system is necessary for denoising the signal which are degraded by noise. Least mean square (LMS), normalized LMS (NLMS), Modified New LMS (MNLMS) and Kalman LMS (KLMS) based adaptive algorithms are analyzed in terms of their performance with the aid of performance measure characteristics such as signal to noise ratio (SNR) and mean square error (MSE). The MNLMS is developed by calculating an optimum learning parameter that best suits for the acoustic signal used. The analysis is carried out for a range of 100 Hz to 10 KHz source signals and the algorithm proves that any ambient noise signals against the source signal in this range can be eliminated and the source signal can be reconstructed. Our simulation results show that KLMS and MNLMS algorithms achieve remarkable performance even in the very low SNR region as compared to LMS and NMLS algorithms. Moreover, it is observed that the output convergence is also very fast for MNLMS and KLMS.