An Improved Neuron Controller and its Application

The learning process is the precondition and base for the neural network control (NNC). In order to process the data easily, the attenuation factor usually be used in the learning process of synaptic weight w_i. However, the attenuation factor often influences the convergence of learning algorithm and the learning quality. This paper presents a method with accelerating factor in the learning rule of NNC and applies it into the neuron controller for parameters adaptive learning. The simulations show that the learning and controlling performance can be improved obviously after applying the improved method, especially for the learning control on the different target values