摘要
介绍了应用基于GA的ANFIS的自适应噪声消除的方法,阐述了基本思想和算法实现过程。神经网络采用五层的ANFIS网络结构,采用自适应GA对模糊规则前件部分的隶属函数参数进行训练,避免了原有BP算法极易陷入局部最优的缺点,可获得全局最优解,用BP算法来调节和优化具有局部性的推理规则结论部分的权值。应用结果表明了该方法的有效性,收敛速度更快、误差更小,滤波率达到了预期要求。
Application of adaptive noise cancellation with ANFIS based on GA is presented, explains its main idea and the implementation procedure of the algorithm. The neural network is based five layer ANFIS network, the study of the parameters of subjection function of the foregoing part of the fuzzy rulers adopts adaptive GA, avoiding the shortcoming of BP algorithm which is casy to fall to part optimum and can get the whole optimum. With BP algorithm regulates and optimizes the joint value of the conclusion part of the fuzzy rulers, which has the partial Characteristic. Its application result indicates this method is validity,with faster simulation rote, less error, and the rate of noise filter reaches expected value.
出处
《计算机技术与发展》
2007年第5期52-54,58,共4页
Computer Technology and Development
关键词
自适应
模糊
神经网络
遗传算法
信号
噪声
adaptive
fuzzy
neural networks genetic algorithm
signal noise
