基于抗差卡尔曼滤波的 GM（1，1）模型在变形预计中的应用

摘要

In this paper , the problem which traditional gray model is vulnerable to be influenced by random disturbance of observation data during the model is built , was solved by building the GM (1,1) model based on robust estimate of Kalman filter , taking advan-tage of robust estimate of Kalman filter can effectively estimate observation value which contains random noise .According to the related case, proved that the GM(1,1) model based on robust estimate of Kalman filter could enhance the forecast precision of deformation monitoring and reflect the deformation tendency better .%针对传统灰色模型建模过程中易受观测数据随机噪声干扰的影响，利用抗差卡尔曼滤波理论能够有效地估计含有噪声的观测值的优点，构建了基于抗差卡尔曼滤波的GM（1，1）模型。结合实例，验证了该模型在一定程度上可以提高变形监测预测精度，更好地反映观测对象的变形趋势。