网络化反馈系统中的动态量化器设计和最优估计

项目名称： 网络化反馈系统中的动态量化器设计和最优估计

项目编号： No.61273109

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 苏为洲

作者单位： 华南理工大学

项目金额： 83万元

中文摘要： 近年来网络控制在理论研究上取得了许多重要进展，但是针对量化等信道特征的反馈系统最优设计与性能分析还有必要进行更深入地研究。其中的关键问题是控制理论中现有的模型与信道的真实特性之间存在着明显的差异，如加性或乘性噪声与量化误差之间的差别。这些差别在中低量化精度下尤其明显。另一方面，在通信和信号处理领域存在着许多值得借鉴的成果，如率-失真理论。本项目的核心思想是试图把信息论中具有反馈机制的编码和预测量化这两种方法引入到反馈控制的研究中。与信息和通信理论不同的是：我们将针对反馈系统特有的实时性和因果性，来研究反馈系统中的动态量化（或编码）问题，进而研究量化误差或编码器的率-失真函数对反馈系统稳定性的影响，以及相应的最优估计与控制问题。从技术上说，我们将运用动态规划和率-失真理论研究有限量化等级（或比特率）条件下的率-失真函数和量化误差最小化问题。进而，研究反馈系统的性能优化问题。

中文关键词： 网络化反馈系统；量化器；率-失真函数；均方稳定；均方最优

英文摘要： During recent years, the research on networked control has been great successful in many research issues. However, optimal design and performance analysis for networked feedback systems in terms of communication channel features such as, quantization, limited data rate, are needed further deeper studies. The key issue is that there exist certain gaps between the channel models used in control area and the features of communication channels, for example the gap between additive or multiplicative noise models and quantization errors. In particular, this gap could be quite significant in the case when the quantization density is lower. On the other hand, there are some very useful tools, such as rate distortion theory, in information and communication theories. In this project, we attempt to study the performance of networked feedback systems by the tools developped in feedback coding and predictive coding areas. Here, the difficulty in the source coding for feedback control systems is that the causality in the signal sequence must be kept, which may not be a serious issue in communication area. By the dynamic progamming and rate distortion theory, we will study the optimal design and the rate distortion function of dynamic coding in networked feedack systems. Furthermore, the performance limits are studied for th

英文关键词： networked feedback systems；quantizer；rate-distortion function；mean-square stabilization；mea-square optimization