拡張カルマンフィルタを用いた直流サーボモータのパラメータ同定

抄録

Parameters of a DC servomotor, such as torque coefficient Kt, induced voltage coefficient Ke, armature resistance R, and damping coefficient D are identified at several operating points of the motor using the extended Kalman filter as a parameter estimator. A known M series random signal is used as a test signal to fluctuate input voltage of the motor so that the performance of the parameter estimation can be much improved. It is seen from the results of the estimation that a test signal suitable for this estimation is dependent on the kinds of parameters to be identified; a M series signal with shorter maximum pulse width is suitable for estimating R and Kt, while a signal with longer is suitable for D, and any signal is all right for Ke. The results are proven by investigating the observability of each paramater estimation system. It is also seen from the results of this identification that Kt or Ke is identified close to its nominal value offered from the manufacturer, while R or D is identified somewhat far from its nominal value. R shows a slight decrease with an increase in motor current. This suggests that the effect of brush contact resistance is not negligible. It is also suggested that it is difficult to estimate in advance D in actual operating conditions.