CAJ | 학술논문

二阶统计量由于不包含相位信息，具有不等价性，不能用来辨识非最小相位系统，且对噪声敏感，所以只能处理白噪声，对于有色噪声不能够进行理想的数据观测。为克服二阶统计量的缺点，利用参数化的方法可以进行高分辨率的多谱估计，采用高阶累积量对FIR系统的闭式解、线性代数方程以及优化解方法对系统辨识的性能进行比较，为进一步研究FIR系统辨识提供参考。
이계통계량유우불포함상위신식，구유불등개성，불능용래변식비최소상위계통，차대조성민감，소이지능처리백조성，대우유색조성불능구진행이상적수거관측。위극복이계통계량적결점，이용삼수화적방법가이진행고분변솔적다보고계，채용고계루적량대FIR계통적폐식해、선성대수방정이급우화해방법대계통변식적성능진행비교，위진일보연구FIR계통변식제공삼고。
Because of the lack of phase information , the second-order statistical magnitude ( the time domain is related function , and the frequency domain is power spectral ) , being of inequality , is unable to identify the non-minimum phase system .Besides , since the second-order statistical magnitude is sensitive to noise , it is merely used to cope with the white noises;while for the colorful noise , it is unavailable to do the ideal statistic observa-tion.In order to overcome the disadvantages of the second-order statistical magnitude , this paper makes use of parameterized method which is able to take the multispectrum signal estimation in high resolution rate , and it a-dopts the method of high-order statistical magnitude to make a comprehensive comparison on the function of iden-tifying the system among three methods in FIR system , namely the method of equations in closed form , the linear algebraic equations as well as the method of optimization equations , aiming to provide certain references for fur-ther studies on FIR system .