【摘要】 针对前馈神经网络的特性和盲均衡中CMA算法的特点构造出一个新的目标函数J(n)。该目标函数综合考虑使盲均衡的代价函数和神经网络的误差函数同时减小。仿真结果表明,使用此算法加快了均衡过程的收敛速度、减小了剩余误差及误码率。更多还原

【Abstract】 In this paper,a target function J(n) is introduced for feed-forward neural network,the decreases both in the cost function of blind equalization and in the err function of neural network being considered.Simulation experiments show that this algorithm accelerates the convergence,and reduces the steady-state error and the bit-error-rate.更多还原