飞机颤振模态参数辨识试验的快速滤波算法

研究飞机颤振随机模型中实际输入-输出信号序列的最优滤波估计问题,利用矩阵论中的矩阵因式分解和统计信号处理中的条件期望公式,将由新息过程构成的块Toeplitz矩阵进行三角分解,得到一种有效的递推滤波算法。对于滤波输入-输出信号的估计值,推导该算法下的估计误差和方差表达式。最后用仿真算例验证采用滤波后得到的输入-输出信号估计值作为飞机颤振模态参数辨识试验的观测信号可得到较为准确的传递函数,进而使得模态参数的辨识也更精确。

A Fast Filtering Algorithm in the Aircraft Flutter Modal Parameter Identification Experiment

This paper,studied the optimal filtering estimation problem of the actural input and output sequences in the aircraft flutter modal parameter identification experiment. Based on the matrix factor decomposition from matrix theory and conditional expectation formula from statistical signal processing, an efficient recursive filtering algorithm was proposed. It is an innovation based approach that relies on the triangular decomposition of block Toeplitz matrices. We derived the estimation error and covariance expression about the estimation of the filtered input and output sequences. Finally in the simulation, we applied the filered input and output sequences to be the observed signals,which were used in the aircraft flutter modal parameter identification experiment. Then we can obtain the more accurate transfer function and the modal parameters.