基于标称性能指标的离线鲁棒预测控制器综合

离线鲁棒预测控制综合算法离线确定一个控制律序列，对应一组吸引域，在线根据当前状态的位置选择相应的控制律,该类控制器在线计算量非常小，而可行性和最优性与其它综合算法相比或多或少要差一些,为此，采用标称性能指标而不是“最坏情况”性能指标来改进离线综合算法的可行性和最优性,改进的控制器保持了原有控制器的稳定性以及控制律关于系统状态的连续性.仿真结果说明了采用标称性能指标的优越性。

Synthesizing Off-line Robust Model Predictive Controller Based on Nominal Performance Cost

The off-line synthesis algorithm of robust model predictive control off-line determines a control-law-sequence with corresponding domains of attraction, and the control law is chosen on-line according to the location of the current state. This controller has very low on-line computational burden, but its feasibility and optimality aspects are, more or less, worse than other synthesizing approaches. For this reason, a nominal, instead of "worst-case", performance cost is adopted in this paper to improve the feasibility and optimality. Simulation result illustrates the advantages of nominal performance cost.