A Robust Adaptive Filtering Algorithm with Double Talk Detection in-Car Communication System

This study focuses on solving the Acoustic Echo Cancellation (AEC) using adaptive filtering algorithms. The paper introduces an approach to incorporate a Double Talk Detection (DTD) mechanism based on Normalized Cross Correlation (NCC) into the Fast Normalized Least Mean Square (FNLMS) algorithm. This integrated algorithm exhibits enhanced convergence speed. The results from simulations validate the effectiveness of this approach in mitigating acoustic echo, particularly in scenarios involving real impulse responses, such as a car cabin impulse response.