Research on Modeling Method Based Kernel Principal Component Analysis for Ball and Beam System
To simulate the human control behavior,and to solve the dimension disaster problem of the simulation model which can simulate complicated controlling behavior,the ball and beam system is researched.The kernel principal component analysis and false nearest neighbor method is adopted,and the simplified ball and beam system controller is designed.The embedding dimension of input time series is determined by the false nearest neighbor method,and then the characteristic value is extracted by the kernel principal component analysis method,so the nonlinear auxiliary variable space feature can be extracted,phase space is reconstructed and the variable is selected.Last from simplified input space to the output space regression mathematical model is fit by using the method of least squares linear regression.Test shows that the control algorithm is effective,with high control precision and stability.
ball and beam system kernel principal component analysis phase reconstruction system modeling false nearest neighbor method
Zefang Wang Miao Song
Electrical Information Engineering Department,Zhengda Software Polytechnic of Chongqing,Chongqing 400056
国际会议
重庆
英文
292-296
2012-11-10（万方平台首次上网日期，不代表论文的发表时间）