无人直升机频域辨识模型参数摄动分析

摘要

This paper presents a model perturbation analysis method based on frequency domain identification for aircraft. Hie model parameter perturbation estimation algorithm was analyzed and proved through the combination of Cramer-Rao Theorem and frequency domain identification. In the meanwhile, the model structure was refined by related indexes, which solved the model order-selecting problem. The model can describe system more precisely through perturbation analysis and model structure refine. The yaw channel frequency domain identification model's parameter perturbation was analyzed for a small-scale unmanned rotorcraft. The comparison of theoretical analysis and the statistical result showed that the effect of theoretical method is consistent with that of experimental method, and the time domain verification showed the model perturbation analysis method is feasible and effective.%提出了一种基于飞行器频城辩识模型的模型参数摄动分析方法.该方法通过将克莱姆-拉奥定理与频域辩识方法相结合对辨识模型参数的摄动范围估计算法进行了分析及证明.同时运用摄动范围计算过程中的相关指标对模型结构进行了优化,解决了飞行器模型辨识过程中模型阶次较难确定的问题.通过时辨识模型的摄动分析及模型结构优化使得模型能更准确地对系统进行描述.根据一种小型无人直升机悬停状态下的偏航通道飞行实验数据,利用该方法对其频域辨识模型参数撮动进行了分析并对其模型结构进行了优化,理论分析结果与统计结果的比较表明理论与实验方法效果一致,模型时域预测输出与实际输出的比较验证了摄动分析方法的可行性及有效性.