Road surface comprehensive service performance evaluation method, device, equipment and storage medium

Abstract

The application discloses a pavement comprehensive service performance evaluation method and device, equipment and a storage medium, and relates to the technical field of pavement service performance evaluation. The method comprises the following steps: obtaining an integrated learning data set, wherein the integrated learning data set comprises pavement performance parameters, environmental parameters and traffic parameters; performing statistical analysis on the performance parameters based on the environmental and traffic parameters and extracting disease characteristics, so that disease identification is more in line with actual road conditions. The preset maintenance standard is combined with the disease characteristics to generate expert experience rules, thereby solving the problem of excessive dependence on data-driven and ignoring engineering experience. Based on the integrated learning data set and the expert experience rules, an integrated learning model is trained to obtain a pavement performance evaluation model. The current pavement performance parameters, environmental parameters and traffic parameters are input into the pavement performance evaluation model to obtain a prediction result. The specification standard calculation result is compared with the prediction result for verification, thereby significantly improving the evaluation accuracy and self-adaptive ability.

Description

Technical Field

[0001] This application relates to the field of pavement service performance evaluation technology, and in particular to pavement comprehensive service performance evaluation methods, devices, equipment and storage media. Background Technology

[0002] With the development of road surface testing technology towards automation and intelligence, road surface performance testing data has shown an explosive growth trend, providing a massive data foundation for road surface performance evaluation.

[0003] Currently, traditional evaluation methods rely too heavily on data patterns and neglect engineering experience when faced with complex traffic environments and large-scale data, resulting in insufficient model adaptability and significant deviations between evaluation results and actual road conditions.

[0004] Therefore, how to effectively improve the accuracy and adaptability of pavement service performance evaluation is an urgent problem to be solved. Summary of the Invention

[0005] The main objective of this application is to provide a method, apparatus, equipment, and storage medium for evaluating the comprehensive service performance of road surfaces, aiming to solve the technical problems of low accuracy and insufficient adaptability in road surface service performance evaluation.

[0006] To achieve the above objectives, this application proposes a method for evaluating the comprehensive service performance of road surfaces, the method comprising:

[0007] Obtain an ensemble learning dataset, which includes road performance parameters, environmental parameters, and traffic parameters;

[0008] Based on the environmental and traffic parameters, statistical analysis is performed on the pavement performance parameters to extract disease characteristics.

[0009] By combining the preset maintenance standards with the characteristics of the disease, expert experience rules are generated.

[0010] Based on the ensemble learning dataset and the expert experience rules, an ensemble learning model is trained to obtain a road performance evaluation model.

[0011] The current pavement performance parameters, environmental parameters, and traffic parameters are input into the pavement performance evaluation model to obtain prediction results. The calculation results of the standard are then compared and verified with the prediction results to obtain evaluation results and verification results.

[0012] In one embodiment, the step of obtaining the ensemble learning dataset, which includes road performance parameters, environmental parameters, and traffic parameters, includes:

[0013] Using road segments as units, multi-dimensional testing of various road surface performance characteristics is conducted, and the test values ​​of various performance characteristics corresponding to each road segment are calculated according to the performance index calculation formula.

[0014] Based on the road surface technical condition index, the test values ​​of the various performance aspects are weighted and calculated to obtain the comprehensive performance evaluation value of each road section.

[0015] The environmental and traffic parameters corresponding to the area where the road is located are obtained based on the climate spatial distribution map and the average daily traffic flow.

[0016] In one embodiment, the step of statistically analyzing the pavement performance parameters based on the environmental parameters and traffic parameters to extract the characteristics of pavement defects includes:

[0017] The pavement performance parameters are statistically analyzed according to the grading standards to obtain the grade distribution of each performance index and identify key deterioration indicators.

[0018] Based on the aforementioned environmental and traffic parameters, the correlation between key deterioration indicators and environmental factors is analyzed.

[0019] Based on the aforementioned correlation patterns, a preset disease type database is matched to extract disease characteristics under the current environmental traffic conditions.

[0020] In one embodiment, the step of combining preset maintenance standards with the disease characteristics to generate expert experience rules includes:

[0021] The maintenance thresholds in the preset maintenance standards are matched according to the disease characteristics, and the maintenance thresholds correspond to the sub-indicators;

[0022] The road section repair demand is calculated based on the aforementioned maintenance threshold.

[0023] Based on the mapping relationship between the degree of repair demand and the characteristics of the disease, expert experience rules are generated.

[0024] In one embodiment, the step of training an ensemble learning model based on the ensemble learning dataset and the expert experience rules to obtain a road performance evaluation model further includes:

[0025] A composite loss function is constructed based on the expert experience rules, and the composite loss function includes an error loss term and a constraint loss term.

[0026] The ensemble learning model is trained based on the ensemble learning dataset, and the composite loss function is defined as the optimization objective to generate a road surface performance evaluation model.

[0027] In one embodiment, the step of constructing a composite loss function based on the expert experience rules, the composite loss function including an error loss term and a constraint loss term, includes:

[0028] The expert experience rules are transformed into constraint functions, which are used to penalize the output of predicted values ​​that deviate from the expert feasible region.

[0029] The constraint function and the regression error function are linearly combined to obtain the composite loss function;

[0030] The composite loss function is expressed as:

[0031]

[0032] Where Obj is the composite loss function, F error Let F be the regression error function. constraint Let α be the constraint function. i Let y be the weight corresponding to the i-th sample. i Let i be the true value of the i-th sample. Let M be the predicted value for the i-th sample, M be the total number of expert experiences, N be the total number of samples, and β be the predicted value for the i-th sample. k,i Let q(i) be the weight of the m-th expert experience corresponding to the i-th sample, and q(i) be the logical judgment value of the i-th sample.

[0033] In one embodiment, the step of inputting current pavement performance parameters, environmental parameters, and traffic parameters into the pavement performance evaluation model to obtain prediction results, and comparing and verifying the prediction results with the standard calculation results to obtain evaluation results and verification results, includes:

[0034] Input the current pavement performance parameters, environmental parameters, and traffic parameters into the pavement performance evaluation model, and output the predicted value of the technical condition index;

[0035] Calculate the standard value of the technical condition index for the same road section based on the standard specifications;

[0036] The predicted value of the technical condition index is compared with the standard value of the technical condition index to obtain the evaluation result;

[0037] The evaluation results were verified and analyzed based on the maintenance thresholds to obtain the verification results.

[0038] Furthermore, to achieve the above objectives, this application also proposes a pavement comprehensive service performance evaluation device, which includes:

[0039] The parameter acquisition module is used to acquire the ensemble learning dataset, which includes road performance parameters, environmental parameters, and traffic parameters.

[0040] The feature extraction module is used to perform statistical analysis on the pavement performance parameters based on the environmental parameters and traffic parameters, and to extract the features of the defects.

[0041] The rule generation module is used to combine preset maintenance standards with the disease characteristics to generate expert experience rules;

[0042] The performance evaluation module is used to train an ensemble learning model based on the ensemble learning dataset and the expert experience rules to obtain a road surface performance evaluation model.

[0043] The comparison and verification module is used to input the current pavement performance parameters, environmental parameters, and traffic parameters into the pavement performance evaluation model to obtain prediction results, and compare and verify the calculation results of the standard with the prediction results to obtain evaluation results and verification results.

[0044] In addition, to achieve the above objectives, this application also proposes a pavement comprehensive service performance evaluation device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the pavement comprehensive service performance evaluation method as described above.

[0045] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the road surface comprehensive service performance evaluation method as described above.

[0046] One or more technical solutions proposed in this application have at least the following technical effects:

[0047] An ensemble learning dataset was acquired, which includes road performance parameters, environmental parameters, and traffic parameters. By acquiring this ensemble learning dataset (including road performance, environmental, and traffic parameters), the problem of poor regional adaptability caused by the single data dimension in traditional evaluation methods was solved, and the data foundation was made more comprehensive.

[0048] By statistically analyzing performance parameters and extracting disease characteristics based on environmental and traffic parameters, the traditional method ignores the influence of climate and traffic factors, making disease identification more consistent with actual road conditions.

[0049] By combining preset maintenance standards with disease characteristics to generate expert experience rules, the problem of over-reliance on data-driven approaches and neglect of engineering experience is solved, and the rules can be expressed in a quantifiable way.

[0050] Based on the ensemble learning dataset and expert experience rules, an ensemble learning model is trained to obtain a road performance evaluation model. The evaluation model is obtained by training the ensemble learning model based on the dataset and expert rules. At the same time, the model accuracy is improved by using expert constraint functions, which solves the problems of low efficiency in processing massive amounts of data and insufficient model adaptability.

[0051] The current pavement performance parameters, environmental parameters, and traffic parameters are input into the pavement performance evaluation model to obtain prediction results. The predicted results are then compared and verified with the standard calculation results to obtain evaluation results and verification results. By inputting parameters, outputting prediction results, and comparing and verifying them with the standard calculation results, the problem of significant deviations between evaluation results and actual road conditions is solved, achieving an effective integration of subjective and objective evaluations, and ultimately significantly improving the accuracy and adaptability of the evaluation. Attached Figure Description

[0052] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0053] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 This is a flowchart illustrating the first embodiment of the pavement comprehensive service performance evaluation method of this application;

[0055] Figure 2 This is a flowchart illustrating the second embodiment of the pavement comprehensive service performance evaluation method of this application;

[0056] Figure 3 This is a flowchart illustrating the third embodiment of the pavement comprehensive service performance evaluation method of this application;

[0057] Figure 4 This is a flowchart illustrating the fourth embodiment of the pavement comprehensive service performance evaluation method of this application;

[0058] Figure 5 This is a schematic diagram of the fitness value change curve during the hyperparameter optimization process in an embodiment of this application;

[0059] Figure 6 RMSE and R during the training and verification process of the embodiments of this application 2 Schematic diagram of the changes;

[0060] Figure 7 This is a feature importance diagram of the model in the embodiments of this application;

[0061] Figure 8 This is a scatter plot of the predicted values ​​from the embodiments of this application and the calculated values ​​from the "Highway Technical Condition Assessment Standard";

[0062] Figure 9 This is a schematic diagram of the module structure of the pavement comprehensive service performance evaluation device according to an embodiment of this application;

[0063] Figure 10 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the comprehensive service performance evaluation method for road surfaces in this application embodiment.

[0064] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0065] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0066] To better understand the technical solution of this application, the following detailed description, in conjunction with the accompanying drawings, through specific embodiments and application scenarios, will illustrate the pavement comprehensive service performance evaluation method, pavement comprehensive service performance evaluation device, pavement comprehensive service performance evaluation equipment, and storage medium provided in the embodiments of this application.

[0067] It should be noted that the comprehensive service performance evaluation method for road surfaces can be applied to terminals, and can be executed by the hardware or software in the terminal.

[0068] The terminal includes, but is not limited to, portable communication devices such as mobile phones or tablets with touch-sensitive surfaces (e.g., touchscreen displays and / or touchpads). It should also be understood that, in some embodiments, the terminal may not be a portable communication device, but rather a desktop computer with touch-sensitive surfaces (e.g., touchscreen displays and / or touchpads).

[0069] The following embodiments describe a terminal including a display and a touch-sensitive surface. However, it should be understood that the terminal may include one or more other physical user interface devices such as a physical keyboard, mouse, and joystick.

[0070] The evaluation of pavement service performance is primarily conducted according to the "Highway Technical Condition Assessment Standard" (JTG 5210-2018). This standard calculates a comprehensive index by assigning weights to various pavement performance indicators, and uses this index to evaluate the overall condition of the pavement. The weights are based on historical data statistics and expert judgment. While this evaluation method has advantages such as simplicity, ease of operation, and high standardization, it has gradually revealed some limitations in practical application: it fails to adjust the weights of various pavement performance indicators for different geographical and climatic environments, making it difficult to cope with diverse environments and leading to a certain degree of deviation between the evaluation results and actual road conditions; furthermore, the method is highly subjective in the actual evaluation process, making it difficult to meet the needs of refined maintenance of high-grade highways. Therefore, how to effectively improve the accuracy and adaptability of pavement service performance evaluation is a problem that urgently needs to be solved.

[0071] Based on this, the embodiments of this application provide a method for evaluating the comprehensive service performance of road surfaces, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the pavement comprehensive service performance evaluation method of this application.

[0072] In this embodiment, the pavement comprehensive service performance evaluation method includes steps S10 to S50:

[0073] Step S10: Obtain the ensemble learning dataset, which includes road performance parameters, environmental parameters, and traffic parameters.

[0074] It should be noted that the ensemble learning dataset can be understood as a structured collection of data collected through pavement inspection equipment and survey methods, including multi-dimensional indicators used to train the ensemble learning model. Specifically, it includes pavement performance parameters, environmental parameters, and traffic parameters. Among these, pavement performance parameters are quantitative indicators reflecting pavement performance, such as the Pavement Condition Index (PCI), Riding Quality Index (RQI), Rutting Depth Index (RDI), Pavement Surface Wearing Index (PWI), Pavement Bump Index (PBI), and Pavement Structure Strength Index (PSSI). Environmental parameters are climate-related data for the pavement's location, including average temperature and precipitation, used to analyze the impact of the environment on the pavement. Traffic parameters are traffic volume data, such as average daily traffic flow, reflecting the effect of traffic load on pavement service performance.

[0075] Step S20: Statistical analysis of pavement performance parameters is performed based on environmental and traffic parameters to extract pavement defect characteristics.

[0076] It should be noted that the characteristics of pavement damage can be understood as typical problem patterns derived from statistical analysis, such as damage types prone to occur under specific climatic or traffic conditions (e.g., rutting or deterioration of skid resistance), used to reflect the actual deterioration trend of the pavement. For example, environmental parameters and traffic parameters can be correlated, and pavement performance parameters can be mathematically processed (e.g., statistical grade distribution, analysis of the advantages and disadvantages of indicators, etc.) to identify data patterns.

[0077] Step S30: Combine the preset maintenance standards with the disease characteristics to generate expert experience rules.

[0078] For example, the preset maintenance standard can be the performance index repair value given in the "Specification for Maintenance Design of Highway Asphalt Pavement", used to determine whether the pavement needs maintenance intervention. It should be noted that expert experience rules can be understood as logical rules formed by combining the characteristics of the defects and the maintenance standards, such as decision conditions based on the Repair Maintenance Rate (RMR), used to guide the model to learn engineering experience.

[0079] Step S40: Based on the ensemble learning dataset and expert experience rules, train the ensemble learning model to obtain the road performance evaluation model.

[0080] For example, ensemble learning models can employ machine learning frameworks built using ensemble algorithms (such as decision tree-based models) to process high-dimensional data and optimize prediction accuracy. Pavement performance evaluation models refer to trained prediction frameworks used to output predicted values ​​for the Pavement Quality Index (PQI).

[0081] Step S50: Input the current pavement performance parameters, environmental parameters, and traffic parameters into the pavement performance evaluation model to obtain the prediction results. Then, compare and verify the calculation results of the standard with the prediction results to obtain the evaluation results and verification results.

[0082] It should be noted that the Pavement Technical Condition Index (PQI) calculated using the formula in the "Highway Technical Condition Assessment Standard" can be used as the standard calculation result. The prediction result can be understood as the PQI prediction value output by the pavement performance evaluation model, reflecting the model's assessment of the current road conditions. The evaluation result can be understood as the pavement service performance level classification (e.g., excellent, good) based on the prediction result, used for actual maintenance decisions. The rationality of the derived model will be evaluated by comparing the prediction result with the standard calculation result (e.g., analyzing the goodness of fit or consistency rate).

[0083] In this embodiment, by acquiring an ensemble learning dataset (including pavement performance, environmental, and traffic parameters), the problem of poor regional adaptability caused by the single data dimension in traditional evaluation methods is solved, achieving a comprehensive data foundation. Statistical analysis of performance parameters based on environmental and traffic parameters and extraction of defect features address the issue of neglecting the influence of climate and traffic factors in traditional methods, making defect identification more closely aligned with actual road conditions. Combining preset maintenance standards with defect features to generate expert experience rules solves the problem of over-reliance on data-driven approaches while ignoring engineering experience, achieving quantifiable expression of the rules. An ensemble learning model is trained based on the dataset and expert rules to obtain the evaluation model. While improving model accuracy through expert constraint functions, the problems of low efficiency in processing massive amounts of data and insufficient model adaptability are solved. By inputting parameters, outputting prediction results, and comparing and verifying them with standard calculation results, the problem of significant deviations between evaluation results and actual road conditions is solved, achieving effective integration of subjective and objective evaluations, ultimately resulting in a significant improvement in evaluation accuracy and adaptability.

[0084] Reference Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment of the pavement comprehensive service performance evaluation method of this application, based on the above. Figure 1 The first embodiment shown presents a second embodiment of the pavement comprehensive service performance evaluation method of this application.

[0085] In the second embodiment, step S10 includes:

[0086] Step S101: Taking road segments as units, conduct multi-dimensional tests on various performance aspects of the road surface, and calculate the test values ​​of various performance aspects corresponding to each road segment according to the performance index calculation formula.

[0087] For example, the road is divided into independent inspection sections to ensure data spatial consistency. Six categories of indicators are collected using equipment such as multi-functional inspection vehicles: road surface damage, smoothness, rutting, skid resistance, bumps, and structural strength. The corresponding performance values ​​for each road section are calculated according to the calculation rules of the six indicators PCI, RQI, RDI, PWI, PBI, and PSSI defined in the "Highway Technical Condition Assessment Standard".

[0088] Step S102: Based on the road surface technical condition index, the test values ​​of various performance aspects are weighted and calculated to obtain the comprehensive performance evaluation value of each road section.

[0089] For example, the values ​​of six indicators, such as PCI, can be linearly combined according to the standard weights to obtain the comprehensive performance evaluation value of each road segment.

[0090] Step S103: Obtain the environmental parameters and traffic parameters corresponding to the area where the road is located based on the climate spatial distribution map and the average daily traffic flow.

[0091] It should be noted that the climate spatial distribution map can be the geographical distribution data of temperature or precipitation, and the traffic parameter can be the average number of vehicles passing through the road segment per unit time.

[0092] In this embodiment, multi-dimensional detection and standardized calculation based on road segment units are adopted, solving the problems of low efficiency and incomplete indicator coverage in traditional manual detection, and realizing automated collection and standardized processing of detection data. Weighted fusion of multi-dimensional indicators based on standardized formulas solves the evaluation bias problem caused by subjective weighting and establishes an initial evaluation benchmark that conforms to industry standards. Integrating climate spatial distribution and traffic flow data solves the problem of evaluation models ignoring environmental and load factors, constructing a multi-dimensional dataset to provide data support for subsequent differential analysis.

[0093] In one implementation, step S20 includes: statistically analyzing pavement performance parameters according to the grade classification standard to obtain the grade distribution status of each performance index and identify key deterioration indicators; combining environmental parameters and traffic parameters to analyze the correlation between key deterioration indicators and environmental factors; and matching a preset disease type library based on the correlation pattern to extract disease characteristics of the current environmental traffic conditions.

[0094] It should be noted that the grading standards can be understood as the performance index grade thresholds defined in the "Highway Technical Condition Assessment Standard" (JTG5210-2018), used to quantitatively evaluate pavement condition. Key deterioration indicators are performance parameters significantly below average identified through statistical analysis, used to reflect the core defects of road sections. Environmental factors can be understood as the dynamic combination of climate parameters (temperature, precipitation) and traffic parameters (traffic volume), jointly affecting the pavement deterioration process. Correlation patterns are the causal relationships between specific environmental factors and key deterioration indicators discovered statistically, used to reveal the causes of road damage. The pre-set disease type library is a set of typical diseases (such as rutting, insufficient skid resistance, etc.) established based on standards and historical data, used to match actual data characteristics. Based on the correlation patterns and disease characteristics extracted from the disease type library, targeted maintenance is carried out on diseased road sections.

[0095] In this implementation, hierarchical statistical analysis is used to identify key degradation indicators, solving the problem of traditional methods neglecting localized performance degradation and achieving precise defect location. Combining the correlation patterns of environmental and traffic parameters addresses the issue of static isolation of climate and load factors. Based on these correlation patterns, features are extracted from a pre-set disease type database, overcoming the problem of strong subjectivity and reliance on experience, and achieving standardization and quantifiability in disease diagnosis.

[0096] Reference Figure 3 , Figure 3 This is a flowchart illustrating the third embodiment of the pavement comprehensive service performance evaluation method of this application, based on the above. Figure 2The second embodiment shown presents a third embodiment of the pavement comprehensive service performance evaluation method of this application.

[0097] In the third embodiment, step S30 includes:

[0098] Step S301: Match the maintenance threshold in the preset maintenance standard according to the disease characteristics. The maintenance threshold corresponds to the sub-indicators.

[0099] It should be noted that the maintenance thresholds in the preset maintenance standards can refer to the performance index thresholds defined in the "Specifications for Maintenance Design of Highway Asphalt Pavement" to determine whether the pavement requires maintenance intervention. Understandably, the maintenance thresholds are numerical limits directly corresponding to sub-indicators (such as PCI and RQI); when the index value falls below the threshold, a repair requirement is triggered.

[0100] Step S302: Calculate the road section repair demand based on the maintenance threshold.

[0101] It should be noted that the road section repair demand can be understood as a quantitative indicator calculated by comparing the values ​​of sub-indicators with the maintenance threshold, which is used to reflect the urgency of the road section needing immediate repair.

[0102] Step S303: Generate expert experience rules based on the mapping relationship between repair needs and disease characteristics.

[0103] For example, the mapping relationship can be a statistical association between repair demand and defect characteristics (such as a positive correlation between the rut index (RDI) and repair demand in high-traffic areas). By combining defect characteristics and repair demand to form quantifiable decision logic (such as "prioritize repair when RDI is below a threshold and traffic volume is high"), the model can be guided to learn engineering experience.

[0104] In this embodiment, dynamic matching of disease characteristics and maintenance thresholds is employed to address the issues of strong subjectivity in maintenance decisions and lack of integration with actual disease patterns in traditional evaluations, thus achieving the objective application of maintenance standards. The calculation of repair demand based on thresholds solves the problem of a lack of quantifiable repair indicators, generating measurable road segment repair priorities. The construction of expert experience rules through mapping relationships addresses the challenge of integrating expert knowledge into data models, achieving quantifiable and integrable rules, ultimately improving the alignment between evaluation results and actual maintenance needs.

[0105] Reference Figure 4 , Figure 4 This is a flowchart illustrating the fourth embodiment of the pavement comprehensive service performance evaluation method of this application, based on the above. Figure 3 The third embodiment shown presents a fourth embodiment of the method for evaluating the comprehensive service performance of pavement in this application.

[0106] In the fourth embodiment, step S40 includes:

[0107] Step S401: Construct a composite loss function based on expert experience rules.

[0108] It should be noted that the composite loss function includes an error loss term and a constraint loss term. The error loss term is the squared error between the predicted value and the true value, and the constraint loss term is the penalty term for the predicted value deviating from the expert feasible region.

[0109] Preferably, step S401 includes: converting expert experience rules into constraint functions, the constraint functions being used to penalize the output of predicted values ​​that deviate from the expert feasible region; and linearly combining the constraint functions with the regression error function to obtain a composite loss function.

[0110] For example, the loss function of the expert constraint function part can be expressed as Equation (1).

[0111]

[0112] In formula (1), F constraint Here, M is the constraint function (i.e., the loss function of the expert constraint function part), M is the total number of expert experiences, N is the total number of samples, and β is the loss function. k,i Let q(i) be the weight of the m-th expert experience corresponding to the i-th sample, and let q(i) be the logical judgment value of the i-th sample. Here, q(i) is represented by formula (2).

[0113]

[0114] In formula (2), k(i) is the predicted value of the i-th sample, and K(i) is the feasible region given by the i-th sample based on expert experience. The composite loss function is expressed as formula (3).

[0115]

[0116] In formula (3), α i Let y be the weight corresponding to the i-th sample. i Let i be the true value of the i-th sample. Let F be the predicted value of the i-th sample. error is the regression error function (i.e., the error loss function of the regression model), and Obj is the composite loss function.

[0117] Step S402: Train the ensemble learning model based on the ensemble learning dataset, define the composite loss function as the optimization objective, and generate a road surface performance evaluation model.

[0118] For example, the MinMaxScaler feature scaling method can be used to normalize the input indicators, linearly mapping the data to the [0,1] range to avoid the impact of differences in indicator units and value ranges on training results. Define the input, output, and label indicators for the pavement performance evaluation model. The input indicators are normalized PCI, RQI, RDI, PWI, PBI, and RMR values; the output indicator is the PQI value; and the label is the PQI value calculated according to the "Highway Technical Condition Evaluation Standard". The dataset can be divided into training, validation, and test sets in a 6:2:2 ratio. Hyperparameter ranges are given experimentally, and hyperparameter optimization is performed on the ensemble learning model, recording the changes in fitness values ​​during optimization. The hyperparameter optimization result is the optimal hyperparameter combination. This combination is used to train the model, recording evaluation indicators such as root mean square error (RMSE) and coefficient of determination (R²) during training and validation. 2 The changes in ) and the corresponding RMSE and R for the training and validation sets. 2 The fitness value changes during the hyperparameter optimization process, as shown below. Figure 5 As shown, the optimization process is divided into three stages: initial, middle and late. In the initial stage, the fitness value drops rapidly, then remains relatively stable, then fluctuates slightly before continuing to decrease, and finally stabilizes. In the middle stage, the fitness value exhibits three states: stable, fluctuating, and decreasing, and the optimization effect continues to improve. In the late stage, the fitness value begins to stabilize after about 45 iterations, and the final fitness value is 0.45, which is lower than 1, indicating that the algorithm has found the global optimum.

[0119] For example, by optimizing the hyperparameters of the service performance evaluation model, the optimal combination of hyperparameters is obtained as follows: learning rate 0.06, maximum number of iterations 80, and maximum tree depth 4. Figure 6 As shown, the RMSE values ​​of both the training and validation sets of the model initially decreased rapidly during the training iterations, then gradually stabilized; while the R-squared value initially increased rapidly, then gradually stabilized. For the optimized service performance evaluation model, the training set R... 2 The score reached 0.897, and the RMSE value was 0.703; the validation set R... 2 The RMS value reached 0.902, with an RMSE of 0.719. The RMSE and RMS values ​​for the training and validation sets are... 2 The values ​​are very close, indicating that the model has good generalization ability and there is no overfitting.

[0120] In this embodiment, by transforming expert experience rules into constraint loss terms, the problem of pure data-driven models ignoring engineering knowledge is solved, and quantifiable fusion of expert experience is achieved. A composite loss function is defined using a linear combination of error loss terms and constraint loss terms, addressing the inability of fixed-weight evaluation methods to dynamically balance data patterns and experience-based judgments, and establishing an adaptive optimization mechanism. Training the model with the composite loss function as the optimization objective solves the problems of low efficiency in processing massive amounts of data and insufficient model generalization ability, improving the consistency between evaluation results and actual road conditions.

[0121] In one implementation, based on the above embodiments and implementation methods, step S50 includes: inputting current pavement performance parameters, environmental parameters, and traffic parameters into the pavement performance evaluation model, and outputting a predicted value of the technical condition index; calculating a standard value of the technical condition index for the same road section based on the standard specifications; comparing the predicted value of the technical condition index with the standard value of the technical condition index to obtain an evaluation result; and performing a verification analysis on the evaluation result based on the maintenance threshold to obtain a verification result.

[0122] It should be noted that the current pavement performance parameters, environmental parameters, and traffic parameters can be understood as the quantified values ​​of six indicators—PCI, RQI, RDI, PWI, PBI, and PSSI—obtained in real time by detection equipment under current environmental and traffic conditions. These parameters are input into the pavement performance evaluation model to obtain the predicted technical condition index. The standard value of the technical condition index for the same road section can be calculated based on the calculation formula in the "Highway Technical Condition Assessment Standard" (JTG5210-2018). The predicted technical condition index value is compared with the standard value to obtain the evaluation result, which can be a pavement service performance level (e.g., excellent, good, medium, poor, or suboptimal) based on the predicted value. By comparing the deviation between the predicted and standard values ​​and verifying whether the sub-indicators are below the maintenance threshold, the rationality of the pavement performance evaluation model is determined.

[0123] For example, predictions are made using a pavement performance evaluation model to obtain the corresponding evaluation indices RMSE and R. 2 The values ​​are used to analyze the fit between the predicted values ​​and the labels on the test set. The final test set R... 2 With a value of 0.864 and an RMSE of 0.812, the performance metrics on the test set are at a high level, indicating that the model has good generalization ability and prediction accuracy. Feature importance plots can be generated to help interpret the model and measure the contribution of each input feature to the model's prediction results, thereby evaluating whether the pavement performance evaluation model has effectively learned the expert constraint function. Figure 7This is a feature importance diagram for the pavement performance evaluation model. In the diagram, 0, 1, 2, 3, 4, and 5 represent six indicators: PCI, RQI, RDI, PBI, PWI, and RMR, respectively. The results show that RMR has the greatest impact on model prediction, followed by PBI and PCI. RMR and PBI are not high-weight indicators in the specifications, but rather are based on expert experience related to the main defects of the surveyed subjects and maintenance standards. This indicates that the pavement performance evaluation model has effectively learned the content of the expert constraint function and can make predictions by combining practical engineering experience. PCI, as the indicator with the highest weight in the "Highway Technical Condition Assessment Standard," reflects the model's compliance with the standard due to its high contribution ratio.

[0124] This paper compares the pavement service performance evaluation results based on the fusion of expert experience and an ensemble learning model with those based on standards. For road sections with inconsistent evaluation results, a detailed analysis is conducted, incorporating information on the sub-indicators of each road section, to discuss the rationality of the evaluation method. The predicted values ​​of the model's test set are compared with the calculated values ​​in the "Highway Technical Condition Assessment Standard" using scatter plots. Figure 8 As shown, most of the predicted values ​​are distributed near the fitted line, indicating that the model's prediction results are highly consistent with the evaluation results of the "Highway Technical Condition Assessment Standard". Statistical analysis shows that among the 857 test set road segments, 843 road segments have prediction grades consistent with the "Highway Technical Condition Assessment Standard", achieving a consistency rate of 98%, while 14 road segments have prediction grades inconsistent with the "Highway Technical Condition Assessment Standard". The road segments with inconsistent evaluations are all those whose "Highway Technical Condition Assessment Standard" grade is "Excellent" while the predicted grade is "Good". These road segments can be divided into three types: (1) PCI is below the repair and maintenance threshold (85); (2) PBI is below the repair and maintenance threshold (75); (3) PWI is below the repair and maintenance threshold (75). When one or more of the main defects of the research object reach the repair and maintenance value range, it indicates that the key single indicator has deteriorated to the stage where immediate maintenance and repair are required. If the PQI grade is still "Excellent", it is obviously inconsistent with the actual road conditions. This demonstrates that the pavement service performance evaluation model can provide evaluation results that are more consistent with the actual road conditions of the road section than the "Highway Technical Condition Evaluation Standard" at the dividing line between the "Excellent" and "Good" grades.

[0125] In this implementation, by inputting multi-source parameters into the evaluation model's output predicted values, the problem of traditional static weight models being unable to dynamically adapt to regional differences is solved, achieving real-time accuracy of the evaluation results. Based on standardized calculation of standard values ​​and comparison with predicted values, the problem of data-driven models being out of sync with industry standards is addressed, establishing a quantitative alignment mechanism between subjective and objective evaluations. Combining maintenance thresholds for verification analysis resolves the misalignment between evaluation results and actual maintenance needs, achieving closed-loop verification of evaluation and decision-making.

[0126] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the pavement comprehensive service performance evaluation method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0127] This application also provides a pavement comprehensive service performance evaluation device; please refer to... Figure 9 The road surface comprehensive service performance evaluation device includes:

[0128] The parameter acquisition module 10 is used to acquire the ensemble learning dataset, which includes road performance parameters, environmental parameters, and traffic parameters.

[0129] Feature extraction module 20 is used to perform statistical analysis on the pavement performance parameters based on the environmental parameters and traffic parameters, and extract the features of the defects.

[0130] The rule generation module 30 is used to combine the preset maintenance standards with the disease characteristics to generate expert experience rules;

[0131] The performance evaluation module 40 is used to train an ensemble learning model based on the ensemble learning dataset and the expert experience rules to obtain a road surface performance evaluation model.

[0132] The comparison and verification module 50 is used to input the current pavement performance parameters, environmental parameters and traffic parameters into the pavement performance evaluation model to obtain the prediction results, and compare and verify the calculation results of the standard with the prediction results to obtain the evaluation results and verification results.

[0133] The pavement comprehensive service performance evaluation device provided in this application, employing the pavement comprehensive service performance evaluation method in the above embodiments, can solve the technical problems of low accuracy and insufficient adaptability in pavement service performance evaluation. Compared with the prior art, the beneficial effects of the pavement comprehensive service performance evaluation device provided in this application are the same as those of the pavement comprehensive service performance evaluation method provided in the above embodiments, and other technical features in the pavement comprehensive service performance evaluation device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0134] This application provides a pavement comprehensive service performance evaluation device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the pavement comprehensive service performance evaluation method in the above embodiment 1.

[0135] The following is for reference. Figure 10The diagram illustrates a structural schematic suitable for implementing the road surface comprehensive service performance evaluation device in the embodiments of this application. The road surface comprehensive service performance evaluation device in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), vehicle terminals (e.g., vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 10 The road surface comprehensive service performance evaluation equipment shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0136] like Figure 10 As shown, the road surface comprehensive service performance evaluation device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the road surface comprehensive service performance evaluation device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the road surface integrated service performance evaluation equipment to exchange data wirelessly or via wired communication with other devices. Although Figure 10 The diagram illustrates a comprehensive road surface service performance evaluation system with various systems; however, it should be understood that implementation or possession of all the systems shown is not required. Additional or fewer systems may be implemented alternatively.

[0137] The pavement comprehensive service performance evaluation equipment provided in this application, employing the pavement comprehensive service performance evaluation method described in the above embodiments, can solve the technical problems of low accuracy and insufficient adaptability in pavement service performance evaluation. Compared with the prior art, the beneficial effects of the pavement comprehensive service performance evaluation equipment provided in this application are the same as those of the pavement comprehensive service performance evaluation method provided in the above embodiments, and other technical features of this pavement comprehensive service performance evaluation equipment are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0138] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0139] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the pavement comprehensive service performance evaluation method in the above embodiments.

[0140] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0141] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described pavement comprehensive service performance evaluation method. This solves the technical problems of low accuracy and insufficient adaptability in pavement service performance evaluation. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the pavement comprehensive service performance evaluation method provided in the above embodiments, and will not be repeated here.

[0142] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for evaluating the comprehensive service performance of road surfaces, characterized in that, The method includes: Obtain an ensemble learning dataset, which includes road performance parameters, environmental parameters, and traffic parameters; Based on the environmental and traffic parameters, statistical analysis is performed on the pavement performance parameters to extract disease characteristics. The preset maintenance standards are combined with the disease characteristics to generate expert experience rules, including: matching maintenance thresholds in the preset maintenance standards according to the disease characteristics, wherein the maintenance thresholds correspond to sub-indicators; calculating the road section repair demand based on the maintenance thresholds; and generating expert experience rules according to the mapping relationship between the repair demand and the disease characteristics. The expert experience rules are transformed into constraint functions, which are used to penalize the output of predicted values ​​that deviate from the expert feasible region. The constraint function and the regression error function are linearly combined to obtain the composite loss function; The ensemble learning model is trained based on the ensemble learning dataset, and the composite loss function is defined as the optimization objective to generate a road surface performance evaluation model. The current pavement performance parameters, environmental parameters, and traffic parameters are input into the pavement performance evaluation model to obtain prediction results. The calculation results of the standard are then compared and verified with the prediction results to obtain evaluation results and verification results.

2. The method as described in claim 1, characterized in that, The step of obtaining the ensemble learning dataset, which includes road performance parameters, environmental parameters, and traffic parameters, includes: Using road segments as units, multi-dimensional testing of various road surface performance characteristics is conducted, and the test values ​​of various performance characteristics corresponding to each road segment are calculated according to the performance index calculation formula. Based on the road surface technical condition index, the test values ​​of the various performance aspects are weighted and calculated to obtain the comprehensive performance evaluation value of each road section. The environmental and traffic parameters corresponding to the area where the road is located are obtained based on the climate spatial distribution map and the average daily traffic flow.

3. The method as described in claim 2, characterized in that, The step of statistically analyzing the pavement performance parameters based on the environmental and traffic parameters to extract the characteristics of pavement defects includes: The pavement performance parameters are statistically analyzed according to the grading standards to obtain the grade distribution of each performance index and identify key deterioration indicators. Based on the aforementioned environmental and traffic parameters, the correlation between key deterioration indicators and environmental factors is analyzed. Based on the aforementioned correlation patterns, a preset disease type database is matched to extract disease characteristics under the current environmental traffic conditions.

4. The method as described in claim 1, characterized in that, The composite loss function is expressed as follows: + ； in, For composite loss function, Let be the regression error function. For constraint functions, For the first The weights corresponding to each sample For the first The true value of each sample For the first The predicted value for each sample, The total number of expert experience entries. The total number of samples, For the first The first sample corresponding to the first sample The weight of expert experience, For the first The logical judgment value of each sample.

5. The method according to any one of claims 1 to 4, characterized in that, The steps of inputting current pavement performance parameters, environmental parameters, and traffic parameters into the pavement performance evaluation model to obtain prediction results, and comparing and verifying the prediction results with the standard calculation results to obtain evaluation results and verification results include: Input the current pavement performance parameters, environmental parameters, and traffic parameters into the pavement performance evaluation model, and output the predicted value of the technical condition index; Calculate the standard value of the technical condition index for the same road section based on the standard specifications; The predicted value of the technical condition index is compared with the standard value of the technical condition index to obtain the evaluation result; The evaluation results were verified and analyzed based on the maintenance thresholds to obtain the verification results.

6. A pavement comprehensive service performance evaluation device, characterized in that, The device includes: The parameter acquisition module is used to acquire the ensemble learning dataset, which includes road performance parameters, environmental parameters, and traffic parameters. The feature extraction module is used to perform statistical analysis on the pavement performance parameters based on the environmental parameters and traffic parameters, and to extract the features of the defects. The rule generation module is used to combine the preset maintenance standards with the disease characteristics to generate expert experience rules, including: matching maintenance thresholds in the preset maintenance standards according to the disease characteristics, wherein the maintenance thresholds correspond to sub-indicators; calculating the road section repair demand based on the maintenance thresholds; and generating expert experience rules according to the mapping relationship between the repair demand and the disease characteristics. The performance evaluation module is used to transform the expert experience rules into constraint functions, which are used to penalize the output of predicted values ​​that deviate from the expert feasible region; to linearly combine the constraint functions with the regression error function to obtain a composite loss function; to train the ensemble learning model based on the ensemble learning dataset, and to define the composite loss function as the optimization objective to generate a road performance evaluation model; The comparison and verification module is used to input the current pavement performance parameters, environmental parameters, and traffic parameters into the pavement performance evaluation model to obtain prediction results, and compare and verify the calculation results of the standard with the prediction results to obtain evaluation results and verification results.

7. A road surface comprehensive service performance evaluation device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the road surface comprehensive service performance evaluation method as described in any one of claims 1 to 5.

8. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the pavement comprehensive service performance evaluation method as described in any one of claims 1 to 5.

Images