基于递推最小二乘法的飞行器模型参数在线辨识

Online Identification of Aircraft Model Parameters Based on Recursive Least Squares Method

为了实现飞行控制系统的在线更新、飞行能力在线评估以及飞行器故障的在线检测,需对飞行器动力学模型中未知参数的在线辨识方法进行研究。首先,在分析飞行器特性的基础上选取纵向小扰动线性运动方程和纵向非线性运动方程作为辨识的模型;之后采用递推最小二乘法分别对单输入多输出线性运动模型和多输入多输出非线性运动模型里面的线性子系统进行参数辨识。仿真结果表明,递推最小二乘法收敛速度快、计算效率高、辨识精度基本满足工程要求,可以用于在线辨识;但是该方法只适用于简单线性模型,无法直接用于非线性复杂运动模型。

In order to realize online update of the flight control system, online assessment of flight capability and online detection of an aircraft fault, the online identification method of unknown parameters in the aircraft dynamics model was studied. Firstly, based on the analysis of aircraft characteristics, the longitudinal small disturbance linear motion equation and the longitudinal nonlinear motion equation are selected as the identification model. Then, the recursive least squares method is used to identify the parameters of the Single Input and Multiple Output (SISO) linear motion model and the linear subsystem of the multiple input and multiple output (MIMO) nonlinear motion model. The simulation results show that the recursive least squares method has fast convergence speed and high computational efficiency. The identification accuracy meets the engineering requirements and can be used for online identification. However, this method is only suitable for simple linear models and cannot be directly adopted for nonlinear complex motion models.