Pavement performance prediction is necessary for road maintenance and repair (M&R) management and plans. The accuracy of performance prediction affects the allocation of maintenance funds. The international roughness index (IRI) is essential for evaluating pavement performance. In this study, using the road pavement data of LTPP (Long-Term Pavement Performance), we screened the feature parameters used for IRI prediction using the mean decrease impurity (MDI) based on random forest (RF). The effectiveness of this feature selection method was proven suitable. The prediction accuracies of four promising prediction models were compared, including Gradient Boosting Decision Tree (GBDT), eXtreme Gradient Boosting (XGBoost), support vector machine (SVM), and multiple linear regression (MLR). The two integrated learning algorithms, GBDT and XGBoost, performed well in prediction. GBDT performs best with the lowest root mean square error (RMSE) of 0.096 and the lowest mean absolute error (MAE) of 6.2% and the coefficient of determination (R2) reaching 0.974. However, the prediction accuracy varies in numerical intervals, with some deviations. The stacking fusion model with a powerful generalization capability is proposed to build a new prediction model using GBDT and XGBoost as the base learners and bagging as the meta-learners. The R2, RMSE, and MAE of the stacking fusion model are 0.996, 0.040, and 1.3%, which further improves the prediction accuracy and verifies the superiority of this fusion model in pavement performance prediction. Besides, the prediction accuracy is generally consistent across different numerical intervals.

As it was found steep slope and sharp curve sections account for 14% of the accident-prone sections according to the accidents data in Chongqing. Regular design indices with certain thresholds are found not enough for this kind of mountain highways. The goal of this paper is to find an optimization design method for combination of steep slope and sharp curve sections based on the analysis of vehicle driving stability. The overall safety model of steep slope and sharp curve combined section is established, and the relevant coordination relationship between design indices of the front and rear alignment unit is established by using the operating speed estimation model. Taking the maximum slope length corresponding to the specified design speed as the design condition of the front segment, the threshold values and variation rules of the design indices of the rear segment under different design speeds are calculated ensuring driving safety. The safety model is simulated by CarSim software, the trajectory offset and lateral acceleration are used to indirectly reflect the degree of lateral instability, and the results are compared to verify the effectiveness of the simplified safety model established in this paper.

Pavement texture characteristics can reflect early performance decay, skid resistance, and other information. However, most statistical texture indicators cannot express this difference. This study adopts 3D image camera equipment to collect texture data from laboratory asphalt mixture specimens and actual pavement. A pre-processing method was carried out, including data standardisation, slope correction, missing value and outlier processing, and envelope processing. Then the texture data were calculated based on texture separation, texture power spectrum, grey level co-occurrence matrix, and fractal theory to acquire six leading texture indicators and eight extended indicators. The Pearson correlation coefficient was used to analyse the correlation of different texture indicators. The distinction vector based on the information entropy is calculated to analyse the distinction of the indicators. High correlations between ENE (energy) and ENT (entropy), ENT and D (Minkowski dimension) were found. The CON (contrast) has low correlations with HT (macro-texture power spectrum area), ENT and D. However, the differentiation of ENE and HT is more prominent, and the differentiation of the CON is smaller. ENE, ENT, CON and D indicators based on macro-texture and the corresponding original texture have strong linear correlations. However, the microtexture indicators are not linearly correlated with the corresponding original texture indicators. D, WT (micro-texture power spectrum area) and ENT exhibit high degrees of numerical concentration for the same road sections and may be more statistically helpful in distinguishing the characteristics of the pavement performance decay of the road sections.

No abstract

Pavement condition data are collected by agencies to support pavement management system (PMS) for decision-making purpose as well as to construct performance model. The cost of pavement data collection increases with the increase of survey frequencies. However, a lower monitoring frequency could lead to unreliable maintenance decisions. It is necessary to understand the influence of monitoring frequencies on maintenance decision by considering the reliability of performance prediction models. Because of different maintenance conditions of urban roads and highways, their performance show different trends. In this paper, the influence of pavement monitoring frequency on the pavement performance models was investigated. The results indicate that low collection frequencies may result in delay in maintenance action by overestimating pavement performance. The collection frequency for Pavement Condition Index (PCI) can be reduced without compromising the accuracy of performance model, more work should be done to ensure the PCI data quality, thus to guarantee the rationality of maintenance decisions. Effect of frequency reduction on pavement performance (IRI) models of urban roads seems greater than on pavement performance (IRI) models of highways, which may lead to heavier monitoring work for urban roads management. This paper provided an example which demonstrated how a comparative analysis can be performed to determine whether the current data collection plan can provide sufficient data for time series analysis.