An on-line self-organizing fuzzy neural neworks for the control of the robotic manipulators
Xiao Xin-lian · Journal of Qiqihar University · 2007
In this paper,based on the dynamic adaptation method a new algorithm for creating a self-organizing fuzzy neural networks(SOFNN),which is implemented to estimate to estimate the uncertainty of the robotic manipulators.The SOFNN can partition the input-output space dynamically,self-learn the Gaussian-type membership functions.Both the number of partitions and their corresponding fuzzy rule configuration are simultaneously and concurrently adaptively.The unique feature of this SOFNN is that it has dynamic self-organizing structure,fast learning speed,good generalization,which is demonstrated on simulation of a two-link robotic manipulators.