基于FPGA的2-D模糊CMAC网络的硬件实现  

Hardware implementation of 2-D fuzzy CMAC based on FPGA

作　　者：白瑞林[1] 沈宪明[1] 章智慧[1] 

机构地区：[1]江南大学控制科学与工程研究中心,江苏无锡214122

出　　处：《计算机工程与设计》2007年第2期430-432,435,共4页Computer Engineering and Design

摘　　要：提出了二维模糊CMAC网络的一种基于FPGA的硬件实现方法。首先,分析了模糊CMAC网络的结构与算法,并以Matlab仿真为依据,得到模糊CMAC网络的FPGA实现所需的参数;在此基础上,对模糊CMAC网络进行硬件模块划分,基于VHDL实现了各硬件模块的功能描述,并对模块结构和权存储方式进行了优化;最后,在特定的FPGA器件上实现了模糊CMAC网络。测试结果表明:该模糊CMAC网络硬件实现具有速度快、精度高的特点,且占用较少的硬件资源,是SOPC中实现模糊CMAC网络模块的一种有效方法。A method of hardware implementation of 2-D fuzzy CMAC based on FPGA is proposed. First, the structure and algorithm of fuzzy CMAC are analyzed, and then the parameters for FPGA implementation according to MATAB simulation are obtained. After that, it partitions hardware blocks for fuzzy CMAC and implements the function of each hardware block with VHDL. Besides, the structure of the blocks and the manner of storing weights are optimized. At last, the fuzzy CMAC based on the given FPGA is imple- mented. The test result shows that the hardware implementation of the fuzzy CMAC has the characteristics of high speed, good precision, and less chip resource. It is an effective method on implementing the modular of fuzzy CMAC in SOPC.

关 键 词：模糊小脑神经网络 现场可编程门阵列 硬件描述语言 函数辨识 硬件实现 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]