基于TD-RLS算法的四旋翼飞行器系统参数辨识

在对飞行器气动参数进行辨识的过程中,常需要对获得的数据求微分或二次微分,而利用差分法求微分会放大噪声的影响,引入滤波器抑制噪声又会产生相位延迟.针对这一问题,提出了一种跟踪微分器-递推最小二乘(TD-RLS)辨识算法.首先,建立了悬停条件下四旋翼飞行器的系统模型;然后,基于实验室四旋翼平台飞行试验实测数据,将TD-RLS算法应用于飞行器参数辨识.最终的辨识结果表明,在四旋翼飞行器悬停或者小角度飞行条件下,该方法可以实时获得比较精确的系统模型.

Parameter identification of quad-rotor aircraft system based on TD-RLS algorithm

In the parameter identification of aircraft,the obtained data are often used to get differential or quadratic differential.Considering the noise amplification caused by finite difference method for differential and phase delay owing to filter,we proposed a tracking-differentiator and recursive least square (TDRLS) identification algorithm.Firsdy,the system model of a quad-rctor aircraft in hover was built.Then,the algorithm was used for parameter identification of the quad-rotor system based on the measured data of quad-rotor aircraft in our laboratory.The results showed that this method can get accurate system model in hovering or flying with small angle.