模糊小脑模型神经网络
Fuzzy CMAC Neural Network
-
摘要: 提出输入层具有一定隶属度的模糊小脑模型神经网络(Fuzzy CMAC),它比小脑 模型CMAC(Cerebellar Model Articulation Controller)能更真实地描述客观世界.给出n维 Fuzzy CMAC算法,仿真结果表明Fuzzy CMAC比小脑模型CMAC具有如下优点:学习收敛 速度快得多,可以学习模糊规则.Fuzzy CMAC比CMAC优越,使CMAC成为Fuzzy CMAC 的特例.Abstract: A Fuzzy Cerebellar Model Articulation Controller (FCMAC) is proposed in this paper. A fuzzy membership function μ(k) is introduced into the FCMAC's input layer. The FCMAC can describe thd world more really than the CMAC can. The FCMAC algorithms for n dimensional problem are given. Simulation results show that the FCMAC has a faster convergence speed than CMAC. The FCMAC can also learn fuzzy reasoning rules (for fuzzy control). In a word, the FCMAC is a better neural network than CMAC, and it makes CMAC as the special case of FCMAC.
计量
- 文章访问数: 3692
- HTML全文浏览量: 123
- PDF下载量: 1242
- 被引次数: 0