摘要
提出了一种基于子空间辨识的模型算法控制器设计方法。在输入信号满足持续激励的条件下,先利用子空间辨识方法提取出系统有限单位脉冲响应序列,再利用单位脉冲序列构成的矩阵设计预测控制器。该方法将独立的对象辨识和控制器设计整合在一起,简化了控制器的设计过程,是一种完全的数据驱动方式。工业过程对象的仿真实验详细给出了该算法的设计过程。
A approach to model algorithm control （MAC） design is proposed based on subspace method identification. Zn the case thatthe input sequences satisfy the condition of persistency of excitation, the finite impulse response （FIR） sequences of the system are firstly extravagated hy using SMI. Then, MAC is designed directly from a matrix formed by FIR sequences. The proposed method not only unites the procedure of both system identification and controller design, but also simplifies the controller design process, and is a totally data-driven technology. The proposed approach is demonstrated using a simulated process.
出处
《控制工程》
CSCD
北大核心
2011年第5期668-670,696,共4页
Control Engineering of China
基金
国家高技术研究发展计划项目(2007AA04Z195)
国家水体污染控制与治理科技重大专项课题(2008ZX07010-003)
关键词
子空间辨识
模型算法控制
有限单位脉冲响应
subspace identification method
model algorithm control
finite impulse response
