摘要
通过分析模糊控制和基于广义基函数的ＣＭＡＣ神经网络，提出一种模糊ＣＭＡＣ（ＦＣＭＡＣ）神经网络。通过ＦＣＭＡＣ权系数的在线学习，实现修正模糊逻辑。给出一种基于ＦＣＭＡＣ的自学习控制器的结构及合适的学习算法，这种网络每次学习少量参数，算法简单。仿真结果表明所提出的控制器优于传统的ＰＩＤ控制器。
A fuzzy CMAC (cerebellar model articulation controller) neural network (FCMAC) was presented based on the theoretic analysis of fuzzy control and CMAC with general basis functions. Fuzzy logical rules were improved through on-line learning of FCMAC weights. A FCMAC based controller structure and a simple learning algorithm were also proposed. In the learning algorithm only small parts of parameters of the FCMAC were adjusted at each learning iteration. Simulation results demonstrated that the proposed controller had better performance than conventional PID controller
出处
《控制与决策》
EI
CSCD
北大核心
1999年第1期77-80,共4页
Control and Decision
