A novel technique for blind source separation using bees colony algorithm and efficient cost functions

Abstract

Blind source separation (BSS) technique plays an important role in many areas of signal processing. A BSS technique separates the mixed signals blindly without information about the mixing system. This paper proposes a novel BSS technique using the bees colony algorithm (BCA) in order to achieve the de-mixing system. Cost function is one the important modules for operation of the BCA. So, we have investigated different types of the cost function. These cost functions are based on the balanced combination of two important paradigms, i.e., higher order statistics and information theory. Experimental results show the proposed technique has high separation accuracy, robustness against the local minima, high degree of flexibility and high speed of convergence in noisy and noiseless environments.