多输入-多输出网络量化系统的分析与综合研究

项目名称： 多输入-多输出网络量化系统的分析与综合研究

项目编号： No.61503106

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 自动化技术、计算机技术

项目作者： 徐雍

作者单位： 广东工业大学

项目金额： 22万元

中文摘要： 网络量化系统集中解决网络传输带宽受限问题，是目前控制领域的研究热点。由于多输入-多输出（MIMO）不确定系统的H∞范数很难解析表达，现有大多数研究只考虑单输入-单输出（SISO）系统，导致其应用范围有限。本项目针对MIMO网络量化系统开展以下四个方面的研究。.首先，利用提升技术和鲁棒控制分析方法，研究单通道传输约束下MIMO系统的量化调度反馈控制问题，给出调度传输策略和最粗糙的量化密度。其次，采用概率统计方法，建立基于Markov链的网络负载时变模型，设计对应的多密度量化器。通过博弈论方法，解决MIMO系统的混合H2/H∞量化控制问题。再次，通过Chapman-Kolmogorov方程，并提取对数量化器的结构特征，设计多通道传输下MIMO Markov系统的状态重置次优滤波器。最后，进行数值仿真和实验验证。

中文关键词： 实时反馈编码（量化）；信道与反馈协同设计；调度策略

英文摘要： Networked systems with quantized information focus on overcoming the capacity constraint of communication channels, which are the recent research hotspot in the control field. Most of existing works concern the single-input and single-output (SISO) systems, because the analytic expression of the H∞ norm of the uncertain multiple-input and multiple-output (MIMO) systems is hard to obtain, which leads to their application is limited. This project will study MIMO networked systems with quantized information from the following four aspects:.Firstly, the problem of the quantized-scheduling feedback control for MIMO systems with one channel constraint is studied by adopting the lifting technique and the robust analysis method. The scheduling strategy and the coarsest quantizer density are obtained. Then, a Markov-based load-varying model of wireless network is established by using the probability statistics method, and the corresponding multi-density quantizer is designed. According to the game theory method, the mixed H2/H∞ quantized control problem is solved for MIMO systems. Furthermore, based on the Chapman-Kolmogorov equation, the state reset sub-optimal filter for MIMO Markov systems over multi-communication channels is designed by utilizing the structure feature of the logarithmic quantizer. Finally, the methods and main results of this project will be verified by numerical simulations and experiments.

英文关键词： Quantization;channel and controller co-design;scheduling