Blind equalization algorithm based on least squares support vector regressor
Huilin Yao, Songlan Zhang, Zhenpeng Wang · 2011
Support vector machine was an effective tool for solving limited learning samples. In the paper, one new blind equalization algorithm based on least squares support vector regressor (LSSVR) was developed. In this method, an iterative reweighed quadratic programming procedure for attaining the solutions has been used in this paper and the simulation results show that it has the good generalization characteristics. The solving steps are also given and it can be generalized to nonlinear blind equalization using kernel functions. The algorithm equalizes the received base-band signals with serious inter-symbol interference. By the algorithm, this method realizes error-free transmission in typical telephone channel. Simulation results show that the proposed algorithm has better convergence performances than the constant modulus algorithm. Especially in small sample cases, the algorithm has lower computational complexity and faster convergence speed.