Design and Implementation of Hybrid Adaptive Noise Canceller for Audio Signal Processing

An adaptive filter is a self-adjusting digital signal processor that minimizes error signals and adapts to dynamic environments. It uses algorithms like Least Mean Squares (LMS) and Recursive Least Squares (RLS) for tasks like noise cancellation and system identification. Traditional adaptive algorithms are effective with Gaussian noise but struggle with sudden disruptions. This research focuses on developing robust adaptive filters to manage various noise types, utilizing MATLAB models that demonstrate superior performance in real-time applications. For creating noisy signals four types of noises considered AWGN (Additive White Gaussian Noise), Johnson, random and uniform are considered. A hybrid LMS-RLS filter combines the stability of LMS with the quick response of RLS, performing better in non-stationary conditions with improved precision and reduced computational demand.