Evolutionary Particle Swarm Algorithm Based on Higher Order Cumulant Fitting for Blind Channel Identification

Evolutionary particle swarm algorithm is proposed for blindly identifying the communication channels. The channel coefficients vector candidates are evaluated by scoring positions of a swarm of particles flying through the multimodal problem space based on their values of the higher order cumulant cost function. The flying is constituted by the interaction of these particles and the alteration arrangement among dimensions of each particle to jump out of the potential local minimum. This procedure leads to the new algorithm with superiority performance but with its complexity comparable to those of genetic algorithm and simulated annealing algorithm.