Research And Application On The Artificial Neural Network To Control The Shape And Gauge Synthesize

The issue bases on the artificial neural network and the emphasis is researching the optimization algorithms on the artificial neural network. Its goals are to put controlling the shape and gauge synthesis and recognizing the flatness pattern into realities. First it is to introduce the basic of the neural network system. Then it introduces emphasis the RBF neural network, the main object we research in the article. It analyses the structure and work mechanism and excellence and shortcoming, it analyses emphasis the algorithms we always use to choose the centers of the data. We design and improve five kinds of algorithms to choose the center of the RBF neural network and simulation them, the results show the algorithms we designed is right and effective. Base on the simulation results we analysis these algorithms and establish their usage field. The plank is used widely in the production and living, its two important indexes are the shape and the gauge. We analysis how to carry out controlling the shape and gauge synthesize based on the diagrammatic view of the sprung equation and the plasticity equation in the article. And establish the controlling mathematical model based on the analysis. Second we use the Legendre orthodoxy polynomials to describe the flatness pattern we meet always. And use the two poles hard restrictor and point matrix to do with the flatness pattern, so two flatness pattern mathematical models are established. The last the shape and gauge are controlled synthesize by the RBF neural network, a two-input and two-output system is built. The discrete Hopfield neural network and the RBF neural network separately recognize two flatness pattern mathematical models. So the system includes the flatness pattern recognized and the shape and gauge controlled synthesize is built.