Dynamic compensation of Markov jump linear systems without mode observation
Dynamic compensation of Markov jump linear systems without mode observation
复制标题
无模态观测的马尔可夫跳跃线性系统的动态补偿
DOI：
10.1109/ecc.2016.7810595
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
U. Hanebeck
中科院分区：
文献类型：
--
作者：
Maxim Dolgov;Christof Chlebek;U. Hanebeck
本文研究了无模式观测的马尔可夫跳跃线性系统的动态输出反馈控制问题。由于该问题的最优非线性控制律是难解的，因此我们假设该问题是一个线性控制器。在此假设下，控制律的计算可以表示为一个包含双线性矩阵不等式的最优化问题。或者，通过引入额外的线性约束并要求某些系统参数为常量，可以将问题转化为线性矩阵不等式。然而，后一种方法是非常有限的，它引入了额外的保守性，这可能会导致较差的性能。因此，我们提出了一种不构成任何非标准约束的迭代算法，并给出了一个数值例子。
In this paper, we address control of Markov Jump Linear Systems without mode observation via dynamic output feedback. Because the optimal nonlinear control law for this problem is intractable, we assume a linear controller. Under this assumption, the control law computation can be expressed in terms of an optimization problem that involves Bilinear Matrix Inequalities. Alternatively, it is possible to cast the problem as a Linear Matrix Inequality by introducing additional linearity constraints and requiring that some system parameters are constant. However, this latter approach is very restricting and it introduces additional conservatism that can yield poor performance. Thus, we propose an alternative iterative algorithm that does not pose any non-standard restrictions and demonstrate it in a numerical example.