A Research on Control Methods of Ball and Beam System Based on Adaptive Neural Network

The composition and control process of the ball and beam system are introduced in the paper, a mathematical model of ball and beam system based on Lagrange equations is set up with the application of the energy balance principle concerning dynamics, and on the base of analysis on such model, the design of a controller is worked out by utilizing the techniques of adaptive neural network. A reference model and the training algorithm equation having realized control on adaptive neural network have been deduced, and, in combination with the designed parameters of the ball and beam system, a simulation experiment concerning the system is carried out which as a result shows that, compared with the traditional neural control method, the new control method has advantages such as high convergence speed, low influence of initial weight, etc. and has better solved the design and control issues concerning multiple-parameter nonlinear system.