Asphalt pavement roughness prediction model based on weighted stacking ensemble learning

Authors

• Wang Jiaqi, School of Future Transportation , Chang’an University , Xi’an, Shaanxi 710018 , ChinaFollow
• Tian Shun, School of Future Transportation , Chang’an University , Xi’an, Shaanxi 710018 , ChinaFollow
• Ren Youjun, China Communications Construction Second Engineering Bureau Co ., Ltd., Xi’an, Shaanxi 710119 , China
• Li Yingyi, School of Future Transportation , Chang’an University , Xi’an, Shaanxi 710018 , China
• Ren Feihong, School of Architecture , Chang’an University , Xi’an, Shaanxi 710018 , China

Corresponding Author

田顺，男，博士，副教授.E-mail：[email protected]

Abstract

Pavement roughness is a key indicator for evaluating pavement performance, which directly affects ride comfort and maintenance decisions. Based on the Long-Term Pavement Performance Program (LTPP ) database in the United States, data from 202 asphalt roads in California, Florida, and Minnesota were extracted. Meanwhile, multidimensional features such as pavement structure, climate, traffic loads, and pavement distress were combined to predict the variation of the international roughness index (denoted as IIRI when it is taken as the indicator variable ). Six regression models, including XGBoost, LightGBM, CatBoost, RandomForest, ExtraTrees, and NeuralNetTorch, were employed to address the issue of insufficient prediction accuracy of single regression models. The training prediction results of base models were integrated by stacking and weighted fusion methods, with Stack-Weight L 2 and Stack-Weight L 3 models built according to the number of learner layers. The results show that the Stack-Weight L 2 and L 3 models outperform single algorithms in terms of prediction performance, with the base models CatBoost and LightGBMXT (the fusion model of LightGBM and ExtraTrees ) yielding the optimal performance. It is demonstrated that machine learning ensemble algorithms can effectively improve the prediction accuracy of pavement roughness.

Publication Date

10-29-2026

DOI

10.14048/j.issn.1671-2579.2026.05.030

First Page

284

Last Page

292

Submission Date

October 2026

Recommended Citation

Jiaqi, Wang; Shun, Tian; Youjun, Ren; Yingyi, Li; and Feihong, Ren (2026) "Asphalt pavement roughness prediction model based on weighted stacking ensemble learning," Journal of China & Foreign Highway: Vol. 46: Iss. 5, Article 30.
DOI: 10.14048/j.issn.1671-2579.2026.05.030
Available at: https://zwgl1980.csust.edu.cn/journal/vol46/iss5/30

Reference

[1]杜艳花.基于时间序列的沥青路面平整度预测方法 [J].兰州交通大学学报,2017,36(1):41-46.Du Yanhua.Method of prediction of asphalt pavement roughness based on time series method [J].Journal of Lanzhou Jiaotong University,2017,36(1):41-46.

[2]斯日古楞,毅力果奇,阿荣,等.基于智能手机的路面平整度评价方法研究 [J].中外公路,2025,45(6):58-66.Si Riguleng,Yi Liguoqi,A Rong,et al.Research on pavement roughness evaluation method based on smartphones [J].Journal of China & Foreign Highway,2025,45(6):58-66.

[3]黄考取,陈宇,徐文军,等.八车道高速公路交通轴载与沥青路面使用性能指标的车道差异性 [J].公路,2024,69(5):424-432.Huang Kaoqu,Chen Yu,Xu Wenjun,et al.Lane variability of traffic axle load and asphalt pavement performance indexes for eight-lane expressway [J].Highway,2024,69(5):424-432.

[4]Wu Z,Yang X M,Zhang Z J.Evaluation of MEPDG flexible pavement design using pavement management system data:Louisiana experience [J].International Journal of Pavement Engineering,2013,14(7):674-685.

[5]张金喜,郭旺达,宋波,等.基于随机森林的沥青路面性能预测 [J].北京工业大学学报,2021,47(11):1256 -1263.Zhang Jinxi,Guo Wangda,Song Bo,et al.Performance prediction of asphalt pavement based on random forest [J].Journal of Beijing University of Technology,2021,47(11):1256 -1263.

[6]邢小亮,杜守军,闫杰,等.基于 XGBoost-Apriori 的山区高速公路事故严重度预测与关联分析研究 [J/OL ].中外公 路,2025 -10-17.https://link.cnki.net/urlid/ 43.1363.U.20251017.1944.025.Xing Xiaoliang,Du Shoujun,Yan Jie,et al.Study on prediction and correlation analysis of highway accident severity in mountainous areas based on XGBoost-apriori[J/OL ].Journal of China & Foreign Highway,2025 -10-17.https://link.cnki.net/urlid/ 43.1363.U.20251017.1944.025.

[7]Baykal T,Ergezer F,Eriskin E,et al.Using ensemble machine learning to estimate international roughness index of asphalt pavements [J].Iranian Journal of Science and Technology,Transactions of Civil Engineering,2024,48(4):2773 -2784.

[8]Ali A,Esekbi M,Sreh M.Development of international roughness index from pavement distress using artificial neural networks (ANNs ) [J].International Journal of Advanced Engineering and Business Sciences,2023,4(1):179-192.

[9]Ali A A,Hussein A,Heneash U.Performance of soft computing technique in predicting the pavement international roughness index:Case study [J].International Journal of Pavement Research and Technology,2025,18(2):346-364.

[10]赵飞,施发垦,陈利.基于 BP神经网络和 Logistic回归分析的沥青路面平整度模型预测 [J].筑路机械与施工机械化,2019,36(7):110-114.Zhao Fei,Shi Faken,Chen Li.Predictive modelling of asphalt pavement roughness using BP neural network and Logistic regression analysis [J].Road Machinery & Construction Mechanization,2019,36(7):110-114.

[11]马鑫源,刘泉,吴建涛.基于机器学习方法的路面平整度预测及模型解释研究 [J].市政技术,2023,41(8):39-44，80.Ma Xinyuan,Liu Quan,Wu Jiantao.Research on pavement roughness prediction and model interpretation based on machine learning methods [J].Journal of Municipal Technology,2023,41(8):39-44，80.

[12]呙润华,于向前.基于编码器 ‒解码器结构的路面平整度预测 [J].同济大学学报 (自然科学版 ),2023,51(8):1182 -1190.Guo Runhua,Yu Xiangqian.Pavement roughness prediction based on encoder-decoder structure [J].Journal of Tongji University (Natural Science ),2023,51(8):1182 -1190.

[13]Choi J,Adams T M,Bahia H U.Pavement roughness modeling using back-propagation neural networks [J].Computer-Aided Civil and Infrastructure Engineering,2004,19(4):295-303.

[14]Damirchilo F,Hosseini A,Mellat Parast M,et al.Machine learning approach to predict international roughness index using long-term pavement performance data [J].Journal of Transportation Engineering,Part B:Pavements,2021,147(4):04021058.

[15]Marcelino P,De Lurdes Antunes M,Fortunato E.Comprehensive performance indicators for road pavement condition assessment [J].Structure and Infrastructure Engineering,2018,14(11):1433 -1445.

[16]Gong H R,Sun Y R,Shu X,et al.Use of random forests regression for predicting IRI of asphalt pavements [J].Construction and Building Materials,2018,189:890-897.

[17]Luo Z Y,Wang H,Li S L.Prediction of international roughness index based on stacking fusion model [J].Sustainability,2022,14(12):6949.

[18]杨术,艾毅,盛善望,等.三元工业固废混料优化设计及强度试验研究 [J].中外公路,2026,46(3):89-99.Yang Shu,Ai Yi,Sheng Shanwang,et al.Optimization design and strength test study of triple-source industrial solid waste mixtures [J].Journal of China & Foreign Highway, 2026,46(3):89-99.

[19]陈云樱,吴积钦,徐可佳.决策树中基于基尼指数的属性分裂方法 [J].微机发展,2004,14(5):66-68.Chen Yunying,Wu Jiqin,Xu Kejia.Using gini-index for attribute selection in decision trees [J].Microcomputer Development,2004,14(5):66-68.

[20]于玲,吴铁军.集成学习:Boosting 算法综述 [J].模式识别与人工智能,2004,17(1):52-59.Yu Ling,Wu Tiejun.Assemble learning:A survey of boosting algorithms [J].Pattern Recognition and Artificial Intelligence,2004,17(1):52-59.

[21]Chen T Q,Guestrin C.XGBoost:A scalable tree boosting system [C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.San Francisco California USA.ACM,2016:785-794.

[22]Meng Q.LightGBM:A highly efficient gradient boosting decision tree [C]//Proceedings of the 31st Conference on Neural Information Processing Systems.Long Beach,CA:Curran Associates Inc.,2017:3149 -3157.

[23]Prokhorenkova L,Gusev G,Vorobev A,et al.CatBoost:Unbiased boosting with categorical features [J].2018,31:6638 -6648.

[24]Del Ser J,Lana I,Bilbao M N,et al.Road traffic forecasting using stacking ensembles of echo state networks [C]//2019 IEEE Intelligent Transportation Systems Conference (ITSC ).Auckland,New Zealand.IEEE,2019:2591 -2597.