Font Size: a A A

Club System Based On Neural Network Control

Posted on:2012-09-29Degree:MasterType:Thesis
Country:ChinaCandidate:F GaoFull Text:PDF
GTID:2208330335479969Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Research of nonlinear control theory and application has been a focus and difficult issues over the years. Ball & beam system is a typical nonlinear controlled object. It is used to test the effect of control strategies and it is also an ideal means of testing in control theory.For the control issue of the ball & beam system, this paper studies the following:1. Nonlinear system's research status and development, and the main control method of nonlinear systems. Neural network control and the ball & beam control system's research situation and significance.2. The composition of the ball & beam system and working principle of the system were studied. We established a mathematical model of the ball & beam system.3. Introduced the PID control method of the ball & beam syste-m using the traditional control scheme.4. Neuron control method is used to control the ball & beam system. We propose a single neuron control method. We used matlab to achieve a single neuron PID control method. The results of comparative study between the single neuron PID control and conventional PID control methods showed that single neuron PID control performance better than the traditional PID control method.5. Introduced the RBF neural network's nonlinear approximation ability, structure, working process and learning method and so on. We made specific design of RBF neural network controller on the basis of the previous. Achieved control for the ball & beam system. We through the analysis of specific results can be obtained a conclusion that this control method is better than the single neuron PID control method in the actual control.6. Since one of the core issue in this paper is avoiding the mathematical model of ball & beam system which is typical nonlinear system. This leads to new problems when new control scheme is designed. By introducing the RBF neural network online identification,RBF neural network controller is designed based online identification of RBF neural network, and experimental results show the effectiveness of the program.
Keywords/Search Tags:Ball & beam, Intelligent control, Single neuron PID control, RBF neural network control, Online identification
PDF Full Text Request
Related items