【摘要】 改进RBF模糊神经网络前件和后件的结构和学习算法,克服了RBF模糊神经网络模糊规则冗余的缺点。利用该系统对含噪声的非线性信号逼近,达到消除噪声的目的。同时,应用该系统对地震信号进行滤波处理仿真,结果表明改进后的RBF模糊神经网络具有学习算法简单,计算量小,实时性好,而且能有效地抑制噪声。更多还原

【Abstract】 The structure and learning algorithm of the antecedent and subsequent network were improved to overcome the shortcoming of redundant fuzzy rule.Based on the approximation of nonlinear noise signal,the system could get to noise cancellation.At one time,it was applied to the filter processing of seismic signal,and simulation results testified the improved RBF fuzzy neural network’s nature of learning algorithm simplicity,a little quantity in its calculation and good for processing in real-time.It is shown that the filter is good for restraining noise with improved RBF fuzzy neural network.更多还原