摘要
针对模型不确定系统的连续时间广义预测控制,引入无穷时域的性能指标使闭环系统达到稳定.用递推最小二乘法估计系统的未知参数,在每个采样点基于估计的状态模型来计算控制输入.通过施加新的终端等式约束,将无穷时域性能指标的优化问题转化成可解的二次规划问题.利用后退时域性能指标的单调性证明,给出使连续时间广义预测控制达到闭环稳定的条件.仿真算例验证了算法的有效性.
The cost function with infinite horizon is introduced in order to guarantee the closed-loop stability of the generalized predictive control with model uncertainty.The system parameters are estimated by using least-squares estimator,and the control input is calculated based on the estimated system model at each sampling instant.The optimal problem which has a cost function with infinite horizon can be transformed into a solvable quadric programming one,if the new end-point equality constraints is appended.Th...
出处
《鲁东大学学报(自然科学版)》
2008年第1期22-26,共5页
Journal of Ludong University:Natural Science Edition
基金
山东省自然科学基金(Y2005G01)
关键词
连续时间系统
广义预测控制
模型不确定性
性能指标
渐近稳定性
continuous-time system
generalized predictive control
model uncertainty
cost function
asymptotic stability