Research of Blind Signals Separation with Genetic Algorithm and Particle Swarm Optimization Based on Mutual Information

Blind source separation technique separates mixed signals blindly without anyinformation on the mixing system. In this paper, we have used two evolutionaryalgorithms, namely, genetic algorithm and particle swarm optimization for blindsource separation. In these techniques a novel fitness function that is based on themutual information and high order statistics is proposed. In order to evaluate andcompare the performance of these methods, we have focused on separation of noisyand noiseless sources. Simulations results demonstrate that proposed method foremploying fitness function have rapid convergence, simplicity and a more favorablesignal to noise ratio for separation tasks based on particle swarm optimization andcontinuous genetic algorithm than binary genetic algorithm. Also, particle swarmoptimization enjoys shorter computation time than the other two algorithms forsolving these optimization problems for multiple sources.

Keywords:

Blind source separation , mutual information , high order statistics , Continuousand Binary genetic algorithm , Particle swarm optimization