信道随机时延等网络不确定性下的均方线性最优控制

随着网络技术的迅速发展，网络化控制系统正在逐步地运用于各个领域。有关这类系统的一个关键问题是如何实时地应对网络在信息传输中产生的不确定性。网络化控制系统中的主要不确定性包括：随机传输时延、丢包、量化与信道噪声、信噪比和带宽约束等。现有的研究结果主要集中于系统稳定性分析和设计。有关性能分析和优化设计方面的研究还有待继续从方法和机理两方面深入。本课题将以信号的功率谱与谱分析为基本工具，针对具有随机时延等不确定性的网络化反馈系统，研究其均方意义下的性能分析与最优设计方法。首先，我们将众多网络随机时延模型中，筛选出适合实际情况的模型；并利用该模型分析系统中信号的功率谱密度函数。然后，运用随机最优控制理论分析上述功率谱密度函数的特点，研究随机时延等信道不确定性对系统性能等约束，并发展出新的最优设计算法。最后，我们将考虑如何把所得结果用于具有多种网络不确定性情形。

With the dramatic development in network technology, networked control systems are applied in many applications. The key issue in these systems is: How to deal with uncertainties which exist in data transmission. These uncertainties include random delays, packet dropout, quantization errors, channel noises, constraints on signal-to-noise ratio and bandwidth, etc. The existing results focus on stabilization problem for these systems. Optimal design and analysis are need further studies. The project attempt to study the optimal control problem for networked systems with random channel delays, and limitation on optimal performance caused by delays. A mean-square small gain theorem like problem will be studied based on the transfer function of a plant and spectral density of communication channels. Then, optimal design method is developed. The design constraints from random delays in communication channels are analyzed. Moreover, these works will be extended to the cases when there exist other type uncertainties, except random, in communication channels