Prediction modelling of rutting depth index for asphalt pavement using de-noising method

ABSTRACT

Intelligent maintenance of pavement requires correct laws of distresses growing. In order to reduce the influence of frequent maintenance on revealing the law of rutting in asphalt pavement, a predictive analysis for rutting depth index (RDI) of asphalt pavement has been conducted via establishing the wavelet-time series prediction model (W-ARMA). The wavelet analysis is firstly used to de-noise the measured RDI data, which can filter out artificial-repairing-caused noise signals that do not vary continuously over time. Then, the rutting evolvement law can be revealed according to the autoregressive moving average (ARMA) model in time series analysis. The results showed that (i) The method of wavelet de-noising to preprocess measured RDI value can reduce the impact of noise signals on rutting predictions; (ii) The improved time series prediction methods are available when applied to rutting analysis based on small size samples (iii) W-ARMA model can directly apply the measured data from real projects, thus simplifying the predictive processing; (iv) The classification of RDI noise signals is given to divide noise factors into the time- and space-based noise; (v) The effective information must be kept during de-noising processing due to the existence of the excessively de-noised or ignored.

KEYWORDS:

• Wavelet analysis
• de-noising
• time series analysis
• rutting depth index
• prediction model

Acknowledgements

Sincere appreciations from the authors go to Prof. J. Murali Krishnan of IIT Madras and several anonymous reviewers for their valuable comments and suggestions. The authors also thank Dr Yunsheng Zhu of the Wuhan University of Technology and Dr Teng Yin of Southeast University, and Mr Tao Wang, Mr Shizhen Yang, Miss Jinghua Huang and Miss Zhijia Tu for their patient revising.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This research is financially supported by National Natural Science Foundation of China (Grant No. 51308429) and the STRIP-WUT (Grant No. 20171049702004).