An improved learning rule for a neuron controller in a quick locating system
Song Yibin, Wang Peijin · 2002
As an important branch of intelligent control, neural network control (NNC) is used in many kinds of high performance systems more and more widely. However, there are many parameter decisions, which is perplexing in system design. It is very important for a neural network controller, such as a single neuron controller, to make a suitable decision of learning rate, weighing factors, attenuation factor and sampling period. The paper presents a design of a neuron controller (NC) using the method of learning rate self-setting and applies the improved NC to a quick locating system. The comparisons of the system performance before and after using the method of learning rate self-setting are made. The simulation results show the improved NC is effective and feasible and the system performance is satisfactory.