利用ARMA模型的直流电机模型的辨识

DC motor identification based on ARMA model

对四旋翼飞行器机体辨识建模时,首先要知道螺旋桨电机的模型.电机的扭矩和角速度取决于电机的输入电压,它们之间的关系与电机线圈的电阻、电感、电机转动惯量、摩擦力、载荷及其他电机的常数有关系.传统电机模型通常利用机理建模,但准确性不高,采用辨识建模能提高模型准确性,给电机一个已知输入,测量其输出的角速度值,然后利用最速下降算法来使电机输出和模型输出值的误差最小,通过ARMA算法来对模型进行辨识,得到系统输入输出的脉冲传递函数,并对得到的系统进行阶跃输入,从而通过比较验证模型的有效性.

When identifying the model of the quadrotor, the first is to know the model of the DC motors. Both the torque and the velocity depend on the voltages, and the relationship is governed by the motor＇ s characteristics such as inductance, resistance, the moment of inertia and so on. The traditional modeling relies on mechanism, but it lacks accuracy. Identification modeling is a better replacement. Here, a response is given to the DC motor, and then the output is measured. An Auto Regressive Moving Average （ARMA） model with a steepest descent algo- rithm is used to minimize the error. In this way, the transfer function has realized. After comparing the step input with the original system, its efficiency can be measured.