The Study On Recurren Neural Network Blind Equalization Algorithm

Equalization technology is an effective method which can remove inter-symbol interference (ISI) in digital communication. Blind equalization is the latest development. It has abroad use because it can adaptively equalize without training sequence. This paper analyzes blind equalization based on neural network, propose two improved algorithms by putting Recursive Neural Network into neural network blind equalization algorithm. The convergence performance of the improved algorithms is illustrated by computer simulations.The main works of this paper can be summarized as follows:(1) In this paper, the fundamental theory and development of neural network and blind equalization is summarized. This paper has summarized the structures of the recurrent neural network, the similarities and differences on these networks are also analyzed.(2) This paper proposes to put two kinds of recurrent neuralnetwork-------diagonal recurrent neural network and quasi-diagonalrecurrent neural network into blind equalization algorithm, using their advantage of the simple structure and the less computational requirement, While with feedback behavior, the recurrent neural network can catch up with the dynamic response of the system, combined the conventional constant modulus algorithm (CMA), the new cost function is proposed, then the steepest descent method is used in the two algorithm. Simulation results show that this algorithm could converge quickly and had the less bit error ratio.(3) The simulation result shows that the convergence performance the new algorithms is better than that of the blind equalization algorithm based on feed-forward neural network.