The research of blind equalization algorithm based on clonal genetic algorithm and neural network

The performance of modern communication system can be reduced by non-ideal character of channel. The main factor is the inter-symbol interference (ISI) caused by aberration of transmission channel. The equalization technique is an efficient method to overcome ISI and improve the characteristics of the system. And the blind equalization technique is the method that just relies on the prior-information of received channel output sequence to adjust the equalizer weights for rebuilding the sending sequence without a known training sequence available. Genetic algorithm optimizing neural network (GA-BP) is one of the blind equalization methods. A preferable local solution space is offered to the neural network by using GA to optimize weights of the neural network for the fitness function of GA includes the information of the cost function of blind equalization. Then, a precise searching is finished in this space with BP neural network algorithm. But simple genetic algorithm (SGA) produces a new value only based on the mutation operator. It often obtains a solution without high precision. Moreover, deficiencies of SGA such as the unusual slow convergence, bad stability and easily oriented prematurity have become the biggest obstacle for its further application. To solve these problems, a clonal genetic algorithm (CGA) is proposed, to increase the precision of the solution. In this paper the new CGA-BP algorithm is used to realize blind equalization. The computer simulations show that the CGA-BP algorithm obtains good convergence characteristics and satisfied equalization results.