The Study Of Methods Of Adaptive Noise Cancellation In Communication System

Signal transmitted in the communication system will be exposed to various noise or interference, how to effectively extract useful signal from the interference environment is always an urgent communication technology needed to be resolved. The usual approach is to use a filter aimed at curbing noise or interference while retaining the useful signal. These filters are divided into classic filters and modern filters. The characteristic of classic filters is that the frequency band of useful signal and that of noise are different components. The classic filter realizes purpose of filtering by using frequency selecting filter. The characteristic of modern filters such as the Wiener filter, Kalman filter, adaptive filter, is that when the frequency band of useful signal and that of noise are overlapping the useful signal can be effectively extracted.In this paper, the noise in the communication system is simply introduced. The principle of adaptive filter and adaptive noise cancellation system, the performance of ANC, the influence of the noise correlation and adaptive filter on the algorithm are studied and analyzed. The fixed step-size LMS algorithm, Normalized LMS algorithm, LMS algorithm with momentum term and LMS algorithm based on Tongue-Like Curve are deeply studied and analyzed. Based on the study and analysis, the improved algorithms are given. The improved algorithms are a variable step-size LMS algorithm applied to adaptive noise cancellation, a variable step-size ELMS algorithm adapted for adaptive noise cancellation and an improved MLMS algorithm. Theory analysis shows that the improved algorithms can better solve the contradiction between convergence rate and tracking speed and steady-state error. In other words, the improved algorithms have stronger capability of noise cancellation and faster convergence rate and littler steady-state error. Based on the MATLAB platforms, simulation and analysis of the above-mentioned adaptive algorithms is carried out. The simulation study primarily include: (1) the ability of noise cancellation of the algorithms; (2) the influence of the adaptive filter tap number and the step-size of the algorithm on the ability of noise cancellation of the algorithms; (3) the convergence speed, steady-state error; (4) the parameters choice of the improved algorithms. The simulation confirmed the theory conclusion.