Orthogonal Wavelet Blind Equalization Algorithm Based On Vector Support Machine

Blind equalization technique is one of key technology in digital communication system, it doesn’t need to send a training sequence, and it will recover the sending signal only according to the statistical characteristics of the received signal sequence. Compared with traditional adaptive equalization, blind equalization technique saves channel bandwidth and raises the channel utilization,thus it has been widely used in the communication field.However, blind equalization has a main contradiction between convergence speed and stead-state error, and it also may be failed in local minimum. So this paper introduces the support vector machine(SVM) into blind equalization and research blind equalization algorithm based on support vector machine and wavelet transform. The main work of this paper contents:(1)The basic knowledge of SVM is introduced, and SVM is applied to blind equalization. Wavelet blind equalization algorithm based on improved support vector machine is proposed. Through the underwater acoustic channel simulation, the results show that the new algorithm has faster convergence speed and smaller stead-state error, it has good performance.(2) Support vector machine wavelet blind equalization algorithm based on improved genetic algorithm is proposed. This algorithm can choose better parameters of Support Vector Machine using Genetic Algorithm’s global optimization ability. And uses support vector machine to initialize weight vector via a short initial data segment. When the weight vector of the equalizer is initialized by SVM, the proposed algorithm carries out the orthogonal wavelet constant modulus blind equalization algorithm. The simulation result with underwater acoustic channel shows that the proposed algorithm converges fast and its’error performance is stable.(3)U-Support Vector Machine is stduied, and U-Support Vector Machine wavelet blind equalization algorithm is proposed by using set U. The simulation result shows that the proposed algorithm having good performance.