摘要
系统辨识是研究建立系统数学模型的理论与方法。从实测的系统输入输出数据或其他数据,用数值的手段重构系统数学模型的办法称为系统辨识。在实际应用中,可以采用许多方法从给定的系统响应数据,如时域响应中的输入和输出数据或频域响应的频率、幅值与相位数据等拟合出系统的传递函数模型,但由于这样的拟合有时解不唯一或效果较差,故一般不对连续系统数学模型进行直接辨识,而更多地对离散系统模型进行辨识。MATLAB的系统辨识工具箱中提出了各种各样的系统辨识函数。在介绍了系统辨识的基本理论和方法的基础上,利用Matlab仿真工具箱对给定实例进行仿真分析。通过仿真结果比较可见,两种方法都能够与实验对象有较好的拟合,而近似最优4阶辅助变量法所得模型拟合精度高于最小二乘法所得模型。
System identification is the theory and method of establishing a mathematical model of system. From the input and output data of measured system, or other data, reconstructed the mathematical model of the system by using numerical means called system identification. In practical applications, there are many methods can he used to fitting a transfer function model from the given response data of system, such as input and output data in the time domain response or frequency, amplitude and phase of the frequency domain response. However, this fitting solution is sometimes not unktue or poor effect. It is generally not to direct identification of continuous system mathematical model but identification of discrete system mathematical model more. There is variety of system identification function in MATLAB System Identification Toolbox. The basic theories and methods of system identification is introduced in this paper, so as an example is simulated and analyzed by the simulation toolbox of MATLAB. Compared the results of simulation, both methods are fit with the experiment object well. For accuracy of approximate the optimal 4-order auxiliary variables method is higher than the least-squares method.
出处
《沈阳师范大学学报（自然科学版）》
CAS
2013年第4期527-530,共4页
Journal of Shenyang Normal University:Natural Science Edition
基金
国家自然科学基金资助项目(61174175)
关键词
系统辨识
仿真分析
拟合精度
system identification
simulated and analysis
accuracy of approximate
