一种改进的最大似然法在直流电机参数辨识中的应用

Applications of an improved method of maximum likelihood in DC motor parameter identification

以D17直流电机实测数据为样本,研究了最大似然法进行直流电机参数辨识问题。该文对最大似然辨识迭代算法进行了改进,提出了滑块递推方法,建立了直流电机最大似然辨识模型,对模型参数进行了辨识和分析,最后对辨识算法进行了验证。结果表明,改进的最大似然法可以快速、准确地辨识出直流电机所需的参数,并具有较高的精度,克服了最小二乘法辨识微分方程参数时存在的缺陷,也解决了最大似然法因观测数据静态分组而导致的辨识精度下降问题,为控制系统/部件建模提供了可行的方法。

The sample using measured data of D17 DC motor, study DC motor parameter identifica-tion problem using the maximum likelihood method. In this paper, iterative maximum likelihood algo-rithm is improved, slider recursive method is proposed, create a DC motor maximum likelihood identifica-tion model, identify and analyze the model parameters, and analyze the accuracy of the recognition re-sult. The results showed that improved push the maximum likelihood method can quickly and accurately identify the parameters required for DC motor, and has a high accuracy, overcomes the defects of the pa-rameters of differential least square method, also solves the maximum likelihood method the problem of i-dentification precision drop because of observational data packet, provides a feasible method for the con-trol system / component modeling.