Super-Resolution Control and Soft-Specification Control;Realization of Systems with High Functionality

考虑了基于量化输入/输出的超分辨率控制和软指定控制问题，建立了基于低分辨率信号的基本控制框架。在量化输入控制方面，将动态量化器扩展到多速率系统，提出了一种性能最优的动态量化器设计方法。并将其推广到一类非线性动力系统。得到了输入仿射系统的解析解。讨论了一种利用抖动信号的新型量化器，并对其性能进行了分析。在量化输出控制方面，提出了一种基于信号投影的容噪系统辨识方法。目前已有几种基于系统模型和量化输出的系统状态重构方法。对于软指定控制，给出了用有界有限不变集描述镇定问题的数学框架。对于具有周期输入的系统，给出了一种产生相关输入以使系统轨迹保持在不变集中的方法。在此基础上，提出了一种利用实用稳定技术实现软跟踪的方法，并通过移动机器人实验验证了该方法的有效性。

Super-resolution control based on quantized input/output and softly specifiedcontrol problems are considered to establish a fundamental control framework based on low-resolution signals. As for quantized input control, dynamic quantizers are extended to multi-rate systems, and a design method of optimal dynamic quantizers is proposed with its optimal performance. Also, it isextended to a class of nonlinear dynamical systems. An analytical solution is obtained for input affine systems. Furthermore, a new type of quantizers which utilize dither signals are discussed, and its performance is analyzed. As for quantized output control, a noise tolerant system identification method is proposed based on signal projection. A few methods have been developed to reconstruct the system state based on the system models and the quantized output. As for softly specified control, a mathematical framework is given to describe the stabilization problem in term of bounded finite invariant sets. For systems with periodic input, a method to produce relevant inputs to keep the system trajectory in an invariant set. Furthermore, a method is proposed to achieve soft tracking via practical stabilization techniques, and its effectiveness is validated through some experiments using mobile robots.