基于支持向量机的非线性信道盲均衡算法研究

针对非线性信道盲均衡问题,考察了一种基于支持向量机的单输入单输出(SISO)盲均衡算法,该算法利用通信信号的常数模特性,将非线性盲均衡问题转化为非线性支持向量回归问题。在此基础上,本文利用分集技术,将该算法拓展至单输入多输出(SIMO)的情况。对两种算法进行的计算机仿真表明,基于支持向量机的SISO盲均衡算法能够有效地抑制信道中的非线性码间干扰;本文提出新算法由于更好地利用了信道的时空特性,具有剩余平均模误差小的优点。

Study on Nonlinear Channel Blind Equalization Algorithm Based on Support Vector Machines

Aimed at the problem of nonlinear channel blind equalization,a Single-Input Single-Output(SISO) algorithm based on support vector machines is investigated in this paper.The nonlinear blind equalization problem is transformed into a nonlinear support vector regression problem by utilizing the constant modulus characteristic of the communication signal.Based on this work,temporal and spatial diversity technique is employed and the SISO algorithm is extended to Single-Input Multiple-Output(SIMO) situation in this paper.Computer simulations performed on these two algorithms show that the SISO algorithm can remove the nonlinear inter-symbol interference efficiently,while the new algorithm proposed in this paper has lower residual average modulus error due to the utilizing of temporal and spatial characteristics of the channel.