用B样条神经网络设计自适应模糊控制器

Designing Adaptive Fuzzy Controllers Based on B-Spline Neural Networks

本文提出一种可用于设计自适应模糊控制器的模糊化Ｂ样条神经网络，并给出了合适的训练算法．由于这种网络在每次训练时仅需对少量权重进行调整，因此构成的模糊控制器学习速率快，可应用于过程控制中．本文最后以电厂中过热汽温的控制为例，说明本文的设计方法是有效的．

A kind of fuzzified B-spline neural network (BSNN) which can be applied to the design of adaptive fuzzy controllers is proposed in this paper, and a suitable training algorithm is also offered. Because there are only a few weights of 'the network should be adjusted during each training period, the designed fuzzy controllers have fast learning rate and can be employed for process control practice. A simulating example of temperature control in power plant shows that the proposed method is effective.