Adaptive echo cancellation using a genetic algorithm

It is well known that processing waveforms containing echoes is a nontrivial problem. This problem arises in such situations as telephone communications, high fidelity recording, and seismic data processing. In such cases, one is interested in detecting echoes and removing their presence. In this thesis, a method is introduced which uses the Hilbert Transform (HT) and Genetic Algorithm (OA) to obtain echo cancellation. The Hilbert Transform and Genetic Algorithm method (HTGA) is based upon detecting the echo position and eliminating it with a near optimal subtraction. The detection of the echo is done with the Hilbert Transform and the cancellation is implemented via the GA. Simulated data will be presented to demonstrate that this method can suppress an echo or multiple echoes. Finally, an adaptive filter using the least-mean-square (LMS) algorithm will be implemented in place of the GA and a subjective comparison between the two optimization methods will be made.