摘要
在输入-输出数据中带有噪声时,传统的最小二乘辨识算法给出模型参数的有偏估计。当噪声方差的估计值可获得时,采用偏差补偿算法能够得到模型参数的一致性估计。在辅助变量算法的基础上结合偏差补偿算法进行推广得到偏差补偿辅助变量辨识算法。采用适用于噪声环境的偏差补偿辅助变量辨识算法,可准确地辨识飞机的颤振模态参数,该算法结合传递函数模型,将带噪声系统的辨识问题转化为迭代求解问题,用来解决输入噪声为白噪声,而输出噪声为有色噪声的复杂辨识情况。利用该算法可将噪声的方差值和传递函数中的模型参数迭代地估计出来。最后利用试飞试验数据辨识飞机颤振的系统参数,将算法与经典的辅助变量算法进行比较,验证了该方法的有效性。
The traditional least-squares identification method generally gives biased parameter estimates when the observed input-output data are corrupted with noise. If estimates of the noise variances are available, then the principle of biased compensated method can readily be used to obtain consistent estimates. We extended the biased compensated method and instrumental variable method to get the bias compensated instrumental variable method. The bias compensated instrumental variable method was adopted for aircraft flutter modal parameter identification under noisy environment. Combining with a rational transfer function model, the identification of system with noisy data was transformed into a iterate problem. The input noise was supposed to be white ,while the output noise was assumed to be colored. Using this algorithm,the noise variance parameters and the model parameters can be obtained iteratively. The simulation with real fli^ht test data shows the efficiency of the algorithm.
出处
《电光与控制》
北大核心
2011年第12期70-74,共5页
Electronics Optics & Control
关键词
飞机颤振
参数辨识
辅助变量
偏差补偿
aircraft flutter
parameter identification
instrumental variable
bias compensate
