不完全测量信息系统的辨识研究

发布时间：2018-11-28 11:47

【摘要】：在控制科学发展初期,人们对控制科学研究还不够深入,在研究系统性能方面,研究者往往忽略相关因素引起的信息缺失,视系统为完全信息系统。而在大数据背景的当今时代,人们对信息系统辨识精度要求越来越高,忽略信息缺失辨识系统参数已不能满足当今社会需求。随着完全信息系统参数辨识缺点逐步突显,广大研究者开始研究不完全信息系统参数辨识算法,部分研究者找到了一种插补辨识算法来降低对系统参数辨识精度的影响,以实现参数的精确辨识并确保系统收敛。因此,本文旨在探索一种样条插补粒子群算法来实现不完全信息系统的参数较精确辨识,通过随机过程知识构建随机变量来分析算法的收敛性,并在实际直流调速系统中检验该算法可行性和有效性。具体步骤如下：首先,针对不完全信息系统,研究了一种基于PSO (Particle Swarm Optimization,粒子群优化)与均值插补法相结合策略实现系统的参数辨识。利用均值插补法对随机丢失信息进行插补,并采用PSO方法对辨识模型参数进行在线迭代优化,以得到合适的误差准则下的最优适应度函数,以实现每个粒子所对应的参数速度和大小迭代,从而实现不完全信息系统参数的实时、精确辨识；在此基础上,进一步分析了该算法的收敛性,通过仿真验证发现本文所提方法在一定条件下可实现不完全信息系统的参数辨识精度的有效提高。其次,将样条插值法引入到不完全信息系统的参数辨识中,以实现系统参数的精确辨识。采用时间戳的信号编码方式,以统计随机缺失的信息量,利用三次样条插补法,对缺失数据进行实时插补,实现完全辨识数据获取；在此基础上,为提高系统辨识精度,进一步采用PSO算法对系统参数进行辨识,并理论推导了该方法的收敛性。最后,将所研究的辨识方法应用到转速反馈控制直流电机调速系统的参数辨识中,在不同缺失率下,实测并获取系统的输入输出数据,利用样条插补辨识法对系统参数进行了辨识,通过验证发现本文所提方法能有效实现直流电机调速系统参数辨识,为该方法进一步的应用和推广奠定了必要的实验基础。
[Abstract]:In the early stage of the development of control science, the research on control science is not deep enough. In the research of system performance, researchers often ignore the lack of information caused by related factors and regard the system as a complete information system. In the background of big data, the accuracy of information system identification is becoming more and more high, and ignoring the missing parameters of information identification system can not meet the needs of the society. As the shortcomings of parameter identification of complete information system become more and more obvious, many researchers begin to study the parameter identification algorithm of incomplete information system. Some researchers have found an interpolation identification algorithm to reduce the influence on system parameter identification accuracy. In order to achieve accurate identification of parameters and ensure the convergence of the system. Therefore, the purpose of this paper is to explore a spline interpolation particle swarm optimization algorithm to realize accurate parameter identification of incomplete information systems, and to analyze the convergence of the algorithm by constructing random variables through the knowledge of stochastic processes. The feasibility and effectiveness of the algorithm are tested in the actual DC speed regulation system. The specific steps are as follows: firstly, a method based on PSO (Particle Swarm Optimization, particle swarm optimization (PSO) and mean interpolation is proposed to identify the parameters of the incomplete information system. The random lost information is interpolated by means of the mean interpolation method, and the parameters of the identification model are iteratively optimized by PSO method in order to obtain the optimal fitness function under the appropriate error criterion. In order to realize the iteration of the velocity and size of the parameters corresponding to each particle, the parameters of incomplete information system can be identified in real time and accurately. On this basis, the convergence of the algorithm is further analyzed, and the simulation results show that the proposed method can effectively improve the precision of parameter identification of incomplete information system under certain conditions. Secondly, the spline interpolation method is introduced into the parameter identification of incomplete information system to realize the accurate identification of the system parameters. Time stamp coding method is used to calculate the amount of random missing information and cubic spline interpolation method is used to interpolate the missing data in real time to obtain the complete identification data. On this basis, in order to improve the accuracy of system identification, the PSO algorithm is used to identify the system parameters, and the convergence of the method is derived theoretically. Finally, the proposed identification method is applied to the parameter identification of the speed feedback control DC motor speed control system. The input and output data of the system are measured and obtained under different missing rates. The parameters of the system are identified by spline interpolation identification method. It is found that the proposed method can effectively identify the parameters of DC motor speed regulation system, which lays a necessary experimental foundation for the further application and popularization of the method.
【学位授予单位】：安徽工程大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP18;N945.14

【相似文献】