Abstract: This paper presents a novel framework for trajectory tracking of robotic manipulators based on the optimal fuzzy clusting neural network system. The model is different from presented neural network。First, typical datum were bolted from source datum by clusting algorithm, and then control rules is optimal and the number of rules is compacted and parameters of membership function are updated. The model not only reduce the time of making rules, avoid exploding the number of rule but also make correct result in simulation work.

Key words: fuzzy neural network, fuzzy clusting, data bolting, fuzzy rule

摘要： 本文提出了新颖的最优模糊聚类神经网络模型对机械手运动轨迹进行控制。该模型与已有的神经网络模型不同之处在于数据首先利用聚类算法对原始数据进行提取优化，然后又进一步优化控制规则以及隶属函数的参数，最终达到模糊聚类神经网络模型的最优化。该模型不但可以缩短规则生成的时间，有效的防止了规则数爆炸，而且在机械手运动控制的应用中效果好。

关键词: 模糊神经网络, 模糊聚类, 数据筛选, 模糊规则