Pavement Surface Condition Index Prediction Based on Random Forest Algorithm
余婷1, 裴莉莉1, 李伟1, 孙朝云1, 呼延菊2
1. School of Information Engineering, Chang'an University, Xi'an Shaanxi, 710064, China; 2. School of Transportation, Southeast University, Nanjing Jiangsu, 211189, China
Pavement Surface Condition Index Prediction Based on Random Forest Algorithm
YU Ting1, PEI Li-li1, LI Wei1, SUN Zhao-yun1, HUYAN Ju2
1. School of Information Engineering, Chang'an University, Xi'an Shaanxi, 710064, China; 2. School of Transportation, Southeast University, Nanjing Jiangsu, 211189, China
摘要With the rise of three-dimensional road detection technology, more and more roads begin to use three-dimensional road detection vehicles to detect road conditions. Pavement surface condition index (PCI), an important indicator of pavement performance, was used as the research target, and the PCI predicted by using the actual pavement data collected by ARAN9000 three-dimensional multifunctional road detection vehicle. Firstly, the data mining technology is used to consider the factors such as pavement distress, environment and pavement structure. and to process and analyze the data related to the pavement surface condition index of a certain highway in Ontario, Canada, such as data cleaning and feature screening. Then, a machine learning prediction model of pavement surface condition index was constructed, and the complex correlation coefficients (R2) of the multiple linear regression model, the neural network model and the random forest model were 0.562, 0.711 and 0.895. Compared with the neural network model, the accuracy of the random forest model in predicting pavement surface condition index was improved by 0.184, the error was reduced by 1.599, and the training speed was improved by 33s. Finally, the random forest model with high precision is selected for optimization. Due to the large number of input variables, it is impossible to determine which line of data is the outlier through simple statistical analysis, so after establishing and predicting the model, the outlier is determined and deleted through the fitting effect between the predicted value and the real value. Then the modified data were used to retrain the model to achieve the optimal training of the current model. The results show that the prediction efficiency and accuracy of the improved random forest model are higher, and the R2 reaches 0.898. The proposed pavement condition index prediction model is accurate and effective, and the prediction results can assist the maintenance department to make more scientific and reasonable maintenance decisions.
Abstract：With the rise of three-dimensional road detection technology, more and more roads begin to use three-dimensional road detection vehicles to detect road conditions. Pavement surface condition index (PCI), an important indicator of pavement performance, was used as the research target, and the PCI predicted by using the actual pavement data collected by ARAN9000 three-dimensional multifunctional road detection vehicle. Firstly, the data mining technology is used to consider the factors such as pavement distress, environment and pavement structure. and to process and analyze the data related to the pavement surface condition index of a certain highway in Ontario, Canada, such as data cleaning and feature screening. Then, a machine learning prediction model of pavement surface condition index was constructed, and the complex correlation coefficients (R2) of the multiple linear regression model, the neural network model and the random forest model were 0.562, 0.711 and 0.895. Compared with the neural network model, the accuracy of the random forest model in predicting pavement surface condition index was improved by 0.184, the error was reduced by 1.599, and the training speed was improved by 33s. Finally, the random forest model with high precision is selected for optimization. Due to the large number of input variables, it is impossible to determine which line of data is the outlier through simple statistical analysis, so after establishing and predicting the model, the outlier is determined and deleted through the fitting effect between the predicted value and the real value. Then the modified data were used to retrain the model to achieve the optimal training of the current model. The results show that the prediction efficiency and accuracy of the improved random forest model are higher, and the R2 reaches 0.898. The proposed pavement condition index prediction model is accurate and effective, and the prediction results can assist the maintenance department to make more scientific and reasonable maintenance decisions.
基金资助:Supported by the Graduate Research Innovation Project of Chang'an University (No. 300103714042)
通讯作者:
YU Ting
E-mail: 2019124044@chd.edu.cn
引用本文:
余婷, 裴莉莉, 李伟, 孙朝云, 呼延菊. Pavement Surface Condition Index Prediction Based on Random Forest Algorithm[J]. Journal of Highway and Transportation Research and Development, 2021, 15(4): 1-11.
YU Ting, PEI Li-li, LI Wei, SUN Zhao-yun, HUYAN Ju. Pavement Surface Condition Index Prediction Based on Random Forest Algorithm. Journal of Highway and Transportation Research and Development, 2021, 15(4): 1-11.
[1] YANG Ru-hai. Specific Application of Preventive Maintenance Technology in Asphalt Pavement Maintenance[J]. Technology Innovation and Application, 2020(8):160-161. (in Chinese) [2] JTG H20-2018, Highway Performance Assessment Standards[S]. (in Chinese) [3] LIU Xian-ming, XIN Gong-feng. Review of Domestic Road Surface Disease Detection Based on Deep Learning[J]. Electronics World, 2021(8):96-98. (in Chinese) [4] SUN Zhao-yun, PEI li-li, LI Wei, et al. Pavement Sealed Crack Detection Method Based on Improved Faster R-CNN[J]. Journal of South China University of Technology:Natural Science Edition, 2020, 48(2):84-93. (in Chinese) [5] KANG Qu. Six Dimensions of High-quality Development of Transportation[J]. Transport Construction and Management, 2018(4):84-87. (in Chinese) [6] CHEN Zhi-jie. Construction Quality Control and Management of High Performance Asphalt Concrete Pavement[D]. Chongqing:Chongqing Jiaotong University, 2016. (in Chinese) [7] LI Bo, HAN Sen, XU Ou-ming, et al. Evaluation of Asphalt Pavement Performance Based on Principal Component Analysis Method[J]. Journal of Chang'an University:Natural Science Edition, 2009, 29(3):15-18. (in Chinese) [8] ZHANG Liang. Prediction of Pavement Condition Index Based on Grey Markov Model[J]. Science Technology and Engineering, 2011, 11(22):5462-5465, 5469. (in Chinese) [9] YAN Ke-zhen, WU Jian-liang. A Method of Asphalt Pavement Performance Evaluation Based on POS-LSSVM[J]. Journal of Highway and Transportation Research and Development, 2009, 26(12):32-34, 100. (in Chinese)[ZK)] [10] FAN Xu-ying, GAO Feng-chun, WANG Hai-long, et al. A Model for Asphalt Pavement Preventive Maintenance Evaluation Based on Improved EW-AHP[J]. Journal of Highway and Transportation Research and Development, 2017, 34(9):8-13. (in Chinese) [11] SOLLAZZO G, FWA T F, BOSURGI G. An ANN Model to Correlate Roughness and Structural Performance in Asphalt Pavements[J]. Construction and Building Materials, 2017, 134:684-693. [12] ABDELAZIZ N, ABD EL-HAKIM R T, EL-BADAWY S M, et al. International Roughness Index Prediction Model for Flexible Pavements[J]. International Journal of Pavement Engineering, 2018, 21(1):88-99. [13] ZHANG C, WANG H, YANG X, et al. A Combinational Prediction Model for Transverse Crack of Asphalt Pavement[J]. KSCE Journal of Civil Engineering, 2018, 22(6):2109-2117. [14] LI W, HUYAN J, XIAO L Y, et al. International Roughness Index Prediction Based on Multigranularity Fuzzy Time Series and Particle Swarm Optimization[J]. Expert Systems with Applications:X, 2019, 2:100006. [15] LI Hai-lian, LIN Meng-kai, WANG Qi-cai. Prediction of Performance of Expressway Asphalt Pavement Based on IFA-SVM[J]. Journal of Highway and Transportation Research and Development, 2019, 36(12):8-14, 78. (in Chinese) [16] SHI Fang. Prediction of Deleterious Synonymous Mutations Based on Random Forest[D]. Hefei:Anhui University, 2018. (in Chinese) [17] ZHOU Zhi-hua. Machine Learning[M]. Beijing:Tsinghua University Press, 2015. (in Chinese) [18] WANG Yao-nan, SUN Wei. Intelligent Control Theory and Application[M]. Beijing:China Machine Press, 2007. (in Chinese) [19] HAN Yan-hua. Prediction of Pavement Performance Based on the Back Propagation Neural Network[J]. Soil Engineering and Foundation, 2013, 27(3):83-85. (in Chinese) [20] ZHAI B, CHEN J. Development of a Stacked Ensemble Model for Forecasting and Analyzing Daily Average PM2.5 Concentrations in Beijing, China[J]. Science of the Total Environment 2018, 635:644-658. [21] TABATABAEE N, ZIYADI M, SHAFAHI Y. Two-stage Support Vector Classifier and Recurrent Neural Network Predictor for Pavement Performance Modeling[J]. Journal of Infrastructure Systems, 2013, 19(3):266-274.
2021, Vol. 15 
