卡尔曼时间序列模型预测路面平整度指数

Prediction on Roughness Index of Pavement by Kalman Time Series Model

路面平整度的发展趋势受交通量、温度及使用时间等许多因素的影响,很难建立综合全面的预测模型。而时间序列法利用历年的IRI值可解决这个问题。针对传统的时间序列法计算的不足,提出了将卡尔曼滤波应用于时间序列IRI预测模型的方法,能充分利用观测值对状态估值进行实时修正,可有效提高预测模型的预测精度,且无需储存大量的历史观测数据。最后,通过实例证明了该模型有效可行。

Pavement roughness is developed with traffic volume, temperature and service time, as a result, the comprehensive prediction model is difficult to be established. However, it could be solved by time series model using IRI value over the years. In relation to the deficiencies in calculation by traditional time series model, the way in which Kalman filtering is applied to time series prediction model using IRI is suggested, which can carry out real-time correction on state estimates according to observed value and effectively increases the accuracy of the prediction model without a large quantity of observation data stored. Finally, the availability and feasibility of the model is proved through examples.