基于改进的CMAC神经网络与PID并行控制的研究  

Research of CMAC Neural Network and PID Controller Based on Improvement

作　　者：张伟[1] 朱大奇[1] 孔敏[1] 李武朝[1] 

机构地区：[1]江南大学通信与控制工程学院,江苏无锡214000

出　　处：《计算机测量与控制》2005年第12期1359-1360,1375,共3页Computer Measurement &Control

摘　　要：提出一种改进的CMAC神经网络控制算法,利用满打满算单元的先前学习次数作为可信度;将改进的CMAC与PID实现复合控制,由CMAC控制器实现前馈控制,PID控制实现反馈控制;仿真表明,改进算法的响应速度和精度有一定的改善。An improved CMAC neural network controlling algorithm was designed, and the former learned times of the addressed hypecubes was used as the credibility. In this algorithm CMAC neural network was compound with PID algorithm, the feed forward control was realized by CMAC controller while the feed back control was realized by PID controller. Simulation demonstrates that the learning rate and accuracy have been improved.

关 键 词：CMAC PID 神经网络 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]