路况预测的动态建模与自适应算法

Dynamic Modelling and Adaptive Algorithm for Pavement Performance Prediction

本文在回归分析模型的预测方法基础上，运用现代时间序列分析方法建立路况预测的动态模型和自适应算法。该方法采用状态空间模型描写路网参数系统，应用自适应ｋａｌｍａｎ滤波方法进行动态系统辩识，应用虚拟噪声补偿技术减小模型误差和观测噪声，并采用自适应递归算法进行路况状态估计和使用性能预测。本方法能自动适应环境因素的影响和过程的时变性，达到良好的预测效果。

Based on the regression analysis model of pavement performance,anew method of pavement performance prediction in terms of dynamic modelling andadaptive algorithm is proposed in the paper by using modern time-serie analysis.Theperfoimance prediction model is reproesented instate space and identified by adaptiveKakman filtering. The technique for compensating filter performanceby a ficitious noise is adopted to reduce the model error and observation noise.Arecursive adaptive algorithm is used to estimate the pavement condition state and topredict the performance variability.The new method with good adaptability forenvironment influence and process variation,will lead to a good prediction effciency.