一种精简化的球杆系统控制器建模方法

摘要

To simulate human behaviors or actions and solve the problem of dimension disaster of behavior functional simulation model in complicated control, with the cue system as the researching object, the simplified control model for cue system is designed. For this model, the methods of kernel principal component analysis (KPCA) and false nearest neighbor point are applied. In addition, variable selection is implemented through phase-space reconstruction; the least square method is used to realize linear regression to fit the regression mathematical model from simplified input space to output space. The test shows that the control algorithm of this controller is effective, and the controller possesses higher control accuracy and stability.%为了模拟人的控制行为,解决复杂控制中行为功能模拟模型的维度灾难问题,以球杆系统为研究对象,设计了精简化的球杆系统控制器模型.该模型采用核主元分析法和虚假最近邻点法,通过重构相空间实现变量选择；同时,利用最小二乘法进行线性回归,拟合从精简的输入空间到输出空间的回归数学模型.试验证明,该控制器的控制算法有效,具有较高的控制精度和稳定性.