摘要
针对饱和约束系统提出了一种鲁棒模型预测控制算法,分别考虑了多面体不确定性和结构反馈不确定性.考虑无穷时域的最坏二次性能指标,通过采用带有饱和特性的反馈控制结构,将控制律的求解转化为一个在线的线性矩阵不等式优化问题.初始时刻优化问题的可行性保证了闭环控制系统的鲁棒稳定性.最后的仿真结果说明了算法的优越性.
A robust model predictive control algorithm for saturated systems is proposed, in which polytopic uncertainty and structure feedback uncertainty are respectively considered. An infinite horizon worst-case quadratic function is considered as performance objective, By using a saturated feedback control structure, the control law is obtained by solving a linear matrix inequality (LMI) optimization problem on-line. The feasibility of optimization problem at the original time implies robust stability of the closed-loop system.
出处
《控制与决策》
EI
CSCD
北大核心
2006年第6期641-645,共5页
Control and Decision
关键词
鲁棒模型预测控制
模型不确定性
线性矩阵不等式
饱和
Robust model predictive control
Model uncertainty
Linear matrix inequality
Saturation
作者简介
李志军(1978-),男,河南平顶山人,讲师,博士,从事模型预测控制、鲁棒控制的研究;Correspondent: LI Zhi-jun, E-mail: lizj78@163.com
刘吉臻(1951-),男,山西岚县人,教授,博士生导师,从事电厂大机组智能优化控制的研究.