Comparative performance analysis and hardware implementation of adaptive filter algorithms for acoustic noise cancellation

Speech has always been one of the most important carriers of information for people and has become a challenge to maintain its high quality. Acoustic Noise Cancellation (ANC) has gained much attention as a technique to remove noise in speech signal and enhance its quality. When the speech signal and noise both change continuously, then arises the need for adaptive filtering. The heart of the adaptive filter is the adaptive algorithm, which converges rapidly to the changes in the input signal. Least Mean Square (LMS) algorithm is easy to understand and implement, stable and robust but has a disadvantage of slow rate of convergence and gradient noise amplification. Hence Normalized Least Mean Square (NLMS) algorithm is used which provides variable step size to increase convergence speed. Recursive Least Squares (RLS) algorithm is used when high convergence speed is expected. These algorithms are simulated using MATLAB and Simulink and then realized on TMS320C6713 board. Performance of these algorithms is compared using parameters like Signal-to-Noise Ratio (SNR), Mean Square Error (MSE), Misadjustment and Convergence time.