Blind Source Separation with Conjugate Gradient Algorithm and Kurtosis Maximization Criterion

Blind source separation (BSS) is a technique for estimating individual source components from their mixtures at multiple sensors. It is called blind because any additional other information will not be used besides the mixtures. Recently, blind source separation has received attention because of its potential applications in signal processing such as in speech recognition systems, telecommunications and medical signal processing. Blind source separation of super and sub-Gaussian Signal is proposed utilizing conjugate gradient algorithm and kurtosis maximization criteria. In our previous paper, ABC algorithm was utilized to blind source separation and here, we improve the technique with changes in fitness function and scout bee phase. Fitness function is improved with the use of kurtosis maximization criterion and scout bee phase is improved with use of conjugate gradient algorithm. The evaluation metrics used for performance evaluation are fitness function values and distance values. Comparative analysis is also carried out by comparing our proposed technique to other prominent techniques. The technique achieved average distance of 38.39, average fitness value of 6.94, average Gaussian distance of 58.60 and average Gaussian fitness as 5.02. The technique attained lowest average distance value among all techniques and good values for all other evaluation metrics which shows the effectiveness of the proposed technique.