Analysis and Design of Switched Control Systems

通过研究取得的主要成果概括如下：1.利用参考调节器和调节器理论在线形成参考输入，对给定的伺服系统提出最优调节控制。将其与参考调速器配合使用，可以大大提高控制性能.一种使用参考调节器的有效方法，对于每个常数参考输入r，确定最大容许集O（r）。要使用引用调控器执行引用输入整形，对于每个t，我们需要确定系统的状态x（t）包含在其中，O（r）。我们提出了一个有效的算法来做这件事.连续时间系统的最大容许集计算提出了一种不确定连续非线性系统最大容许集子集的计算方法。这是对不含不确定性的线性连续时间系统的结果的推广.非线性数字伺服系统的设计离散时间系统是由连续系统离散化得到的，其不确定性类别不易刻画.采用滑模控制的在线参考输入成形应用在线参考输入管理方法，提出了一种带约束系统的滑模控制方案.模型预测型控制：我们提出了一个模型预测型控制方案，这是非常类似的模型预测控制方案。我们的模型预测型控制的优点是它给原来的优化问题的准全局最优解。

The main results obtained through the research are summarized as follows:1. On-line reference input shaping using reference governor and regulator theory.We proposed to use the optimal regulation control for a given servo systems. By using this with reference governor, we can improve the control performance very much.2. An efficient method to use reference governor.For each constant reference input r, the maximal admissible set O (r) is determined. To execute reference input shaping using reference governor, for each t, we need to determine the state x(t) of the systems is contained in which O(r). We proposed an efficient algorithm to do it.3. Maximal admissible set computation for continuous time systems.We proposed a computation method of subset of the maximal admissible set for continuous nonlinear systems with uncertainty. This is a generalization of the previous result for linear continuous time system without uncertainty.4. Design of nonlinear digital servo systems.It is not easy to characterize the uncertainty class for discrete time systems which are obtained by some discretization of continuous systems.5. On-line reference input shaping using sliding mode control.By applying on-line reference input management method, we proposed a sliding mode control scheme for systems with constraints.6. Model predictive type control.We proposed a model predictive type control scheme which is quite similar to the model predictive control scheme. The advantage of our model predictive type control is it gives quasi globally optimal solution of the original optimization problem.