模糊CMAC的硬件结构分析及其FPGA实现

Hardware Structure Analysis of Fuzzy CMAC and Its FPGA Implementation

提出了模糊CMAC的一种基于FPGA的硬件实现方法。与其它FPGA实现的神经网络相比,包含了可以用于在线学习的权学习算法。分析了模糊CMAC的模型结构及其相应的硬件模块;用VHDL实现基于上述模块的模糊CMAC;对该模糊CMAC进行硬件综合与测试。测试结果表明:该模糊CMAC的FPGA实现方法是可行的,硬件化后的网络具有速度快、精度高、占用器件资源少的特点,是在SOPC中实现模糊CMAC模块的一种有效方法。

This paper proposes a FPGA implementation structure of a fuzzy CMAC.. Compared with other neural networks implemented by FPGA, it contains the learning algorithm which can be employed to realize the on-line learning. The model and the relevant hardware modules of fuzzy CMAC are analyzed. It implements the fuzzy CMAC based on the above hardware modules with VHDL. The design is synthesized and tested. The test result shows that the method of the hardware implementation of the fuzzy CMAC is feasible. The implemented network comprises the characteristics of high speed, good precision and little chip resource. It is an effective method in implementing the module of fuzzy CMAC in SOPC.