基于人工免疫网络的正交小波盲均衡算法

Orthogonal Wavelet Blind Equalization Algorithm Based on Artificial Immune Network

传统的常数模盲均衡算法存在收敛速度慢、均方误差大、易陷入局部极小值点等缺点。为此,提出一种基于人工免疫系统的正交小波盲均衡算法。该算法将均衡器系数向量作为抗体,经过抗体克隆、变异和抑制等操作,搜索到适应度值最高的抗体,即均衡器的最优系数,使权向量跳出局部最优点,接近全局最优点,并利用正交小波变换改善常数模盲均衡算法的收敛性,降低均方误差。仿真实验结果表明,该算法收敛速度快、均方误差小,能得到全局最优解。

The traditional constant modulus blind equalization algorithm has some disadvantages such as slow convergence speed,large mean square error,and easily falling into local minimum points.Aiming at these shortcomings,wavelet blind equalization algorithm based on artificial immune system is presented.In the presented algorithm,the equalizer coefficient vector is regarded as antibodies and the optimal solutions can be obtained by replication,mutation and suppression.It can make weight vector escape from local minima and approach the global minima.Its convergence speed can be improved via using orthogonal wavelet transform and its mean-square error can be reduced by artificial immune algorithm.Simulation results show that this presented algorithm has rapid convergence speed,small mean square error and can obtain global optimum solution.