机场道面使用性能的动态自回归预测模型

针对我国机场道面性能观测时间短，观测数据少，使用现有模型预测精度低，不能根据观测值动态更新预测模型等现状，提出了将卡尔曼滤波应用于时间序列预测的方法，建立了动态自回归预测模型，进行机场道面使用性能的预估.选取我国华东某机场的实测道面状况指数为基础数据，进行时间序列建模，应用卡尔曼滤波算法实现时间序列模型参数的实时更新，分析模型的预测效果.时间序列数据较少时，难以建立高精度的自回归模型，通过卡尔曼滤波处理建立的动态自回归预测模型精度明显提高.
A dynamic auto regression model based on the time series analysis with Kalman filter was proposed for pavement condition prediction. Existing prediction models could not be applied to Chinese airports due to the incomplete monitoring data and the complexity to be updated. The time series model was first established based on the pavement condition index (PCI) data of the airport in eastern China. Then Kalman filter algorithm was utilized to update the models. By the comparison with the actual monitoring data, the prediction models are proven to be reliable in Chinese airports. The predictions of the dynamic auto regression model are more accurate than the auto regression model despite the incomplete monitoring data