The Study of Appling Support Vector Machines to the Missile Control System-Take Roll for Example

• Main Authors: Tan, Yen-Chiao, 譚彥樵
• Other Authors: Deng, Shih-Kang
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/58203466331108373757

碩士 === 國防大學理工學院 === 兵器系統工程碩士班 === 101 === This thesis used missile flying control theory to define six-degree of freedom, three-coordinate system, and rolling motion equations for missile maneuvering and used traditional PID controller for analyzing the results by gaining and out response. At last, this research introduced the Support Vector Machines technique, to construct an adapted control algorithm by learning mode into PID controller. This newly defined controller (SVRC) can determine the extremely short-term changes of control which commands to predict the output of the next time step precisely. This research employed LABVIEW (laboratory virtual instrument engineering workbench), an image programming language, and MATLAB, a support vector machine toolbox. Then, we integrated the SVR technique into autopilot that adopted a dual-feedback flight control system structure of acceleration and angular velocity. The simulation results show that the SVR controller (SVRC) can control missile circumrotate acceleration commands precisely and achieve adaptive control.