Research of the Neural Network by Back Propagation Algorithm

摘要

In the control of the position of the robots systems one of the more important is to assure the minimum errors between the output and the target. All advanced researches in the word propose to use the neural network (NN) and the learning algorithm like Widrow and Hoff, or Levenberg-Marquard by using the least mean square (LMS) of errors and Delta rule, or back propagation training algorithm. Present paper is showing the mathematical model and numerical simulation of some important neurons types used in many applications that require extreme precision and neural network. All assisted researches were made with the owner LabVIEW virtual instrumentation. The research results and virtual LabVIEW instrumentation can be used in many other mechatronics applications.