Real time noise-cancellation using ICA, PSO and PE

In order to provide noiseless transmission of speech in wireless communication systems a real-time implementable noise cancellation algorithm is developed. Speech and noise sources are not known but only their mixtures are observed. That system is modeled with instantaneous mixture model. Combination of independent component analysis (ICA) and particle swarm optimization (PSO) algorithms is used to separate speech and noise. However, ICA has an ambiguity such that it is not possible to know which one of the separated signals is speech or noise. As a result, the transmitted signal can be noise, instead of speech. To overcome this ambiguity problem, a pitch extraction (PE) algorithm is developed and combined with ICA-PSO. ICAPSO-PE algorithm is implemented in MATLAB. Contributions of this work are modifying objective functions of ICA algorithm to make them more robust, combining ICA with PSO to make it work fast and robust, and overcoming the ambiguity problem using PE algorithm.