Font Size: a A A

Fuzzy Control Of Servo System Test Device

Posted on:2010-12-21Degree:MasterType:Thesis
Country:ChinaCandidate:L J SongFull Text:PDF
GTID:2208360275998653Subject:Mechanical Manufacturing and Automation
Abstract/Summary:PDF Full Text Request
With the development of science and technology and the changes in the model of war, it requires weapons to be good maneuverability, rapid collimation and hiting target accurately, and servo system which has a fast response and high precision in static and dynamic states is the key to ensure the combat effectiveness of weapons. In order to verify the feasibility of technical schemes, and shorten the cycle of weapons' development and production, we must test and assess the performance of servo system, but nowadays, the device which is used to test the performance is scarce. Based on the loading and testing device for servo system, this paper introduces partial key technologies in the course of production of this device as follows:First, the working principle, overall structure and hardware of the loading and testing device for servo system are introduced. Some important hardware is introduced in detail, the relevant hardware circuitries are also designed.Because of its reliability, flexibility and real-time performance, CAN bus is widely used in industry. This paper chooses CAN bus to realize communication between the upper computer (IPC) and the lower computer (MCU), and the hardware and software designed in CAN bus communication system are introduced in detail.The upper computer is designed, including the design of the testing interface, the composition of the main program and the realization of the functional modules.On the basis of analyzing the working principle and the characteristics of Magnetic Particle Brake (MPB) which is used to simulate the load torque, this paper adopt improved BP neural network to get the model of MPB, because neural network can approximate any nonlinear function and has strong learning ability.Fuzzy-neural network has been widely applied in non-linear system's control. On the basis of getting the model of MPB, this paper designs a self-learning fuzzy neural network controller to control the MPB. The simulation results validate the feasibility of the controller.
Keywords/Search Tags:CAN bus, Magnetic Particle Brake (MPB), neural network, fuzzy- neural network, system identification, control
PDF Full Text Request
Related items