Schur分解降维算法研究.doc

实用标准文案 PAGE II - - 文档 基于Schur分解的控制系统降维算法研究 摘要 模型降阶是在一定条件下将一个阶次较高的系统转化为一个近似的较低系统的过程，以便降低大型复杂系统的理论分析的难度及减少数据运算量。一般的模型降阶方法应满足的基本条件为降阶系统与原系统的近似误差足够小降阶系统能保持原系统的某些性能例如稳定性、动态指标和稳态指标等，同时保证降阶计算过程稳定而高效等。对于现代大型或复杂系统，由于其高维数或复杂性，若直接进行分析则显得较为相对困难或数值模计算拟耗时过长甚至对某些问题无法模拟。在这样的背景下,对模型的规模或阶数进行有效的降阶处理就显得尤为重要了。目前,广泛应用于导航系统、力学系统、控制系统、定位测量系统等。 系统模型降维作为一个理论课题，自60年代末至今，国内外广大控制理论研究的及工程应人员提出了大量模型降阶的方法。一般而言，高阶模型的降阶方法可归纳为时域降阶法和频域降阶法，具体有：频率拟合Pade逼近法、连分式降阶法、集结法、摄动法、Pade逼近法和Rotch逼近法及它们的改进方法等。本篇论文将挑出其中的一些典型方法做出一个概括性的介绍。 本篇论文主要分析了基于Schur分解的控制方法进行研究，并通过数学实验仿真验证此方法验证可以达到高阶模型降阶的要求。 关键词　模型降阶；时域降阶；频域降阶；Schur分解 The decomposition of the control system based on Schur dimension reduction algorithms Abstract The reduced order model is in certain conditions will be a order times higher system into a similar low the process of the system, so as to reduce the large complicated system theory the difficulties of analysis and reduce the quantity of data. The general model the reduced order the basic conditions should satisfy the method for the reduced order system and the original system approximation error reduced order system is small enough to keep some of the original system properties such as stability, the dynamic performance and steady state index, while ensuring the reduced order calculation process stable and high efficiency, etc. For modern large or complex system, due to its high dimension or complexity, if direct analysis appears more relatively difficult or numerical model calculation is too time consuming and even to some problem cannot simulation. In this context, the scale of the model or order number of effectively reduced order processing is particularly important. At present, widely used in navigation system, mechanical system, control system, positioning measurement system, etc. System model dimension reduction as a theory problem, since the s so far, the domestic and foreign general control theory research and engineering personnel should be put forward a method of the reduced order model. Gen