论文部分内容阅读
本文讨论在极点配置的约束下，使［Ｐ］和［Ｖ］·［Ｖ－１］（条件数）极小化的问题，其中Ｐ是（Ａ＋ＢＦ）＇Ｐ＋Ｐ（Ａ＋ＢＦ）＝－２Ｉｎ的工定解，Ｖ是Ａ＋ＢＦ的特征向量矩阵．两种指标都反映了系统鲁棒稳定的程度．通过定义一矩阵函数并引入新的自由变量Ｕ，可放松极点配置的约束，并能系统的推导［Ｐ］／Ｕ及（［Ｖ］·［Ｖ－１］）／Ｕ，从而将鲁棒设计转化为无约束的梯度法寻优，实例说明，本文设计方法的效果很好．
This paper discusses the problem of minimizing [P] and [V] · [V-1] (condition number) under the pole assignment constraint, where P is the sum of (A + BF) ’P + P (A + BF) = -2In The fixed solution, V is the A + BF eigenvector matrix. Both indicators reflect the degree of robustness and stability of the system. By defining a matrix function and introducing a new free variable U, the pole placement constraints can be relaxed and systemically derived [P] / U and ([V] · [V-1]) /U, The rod design is transformed into an unconstrained gradient method, and the example shows that the design method in this paper works well.