Deformation Monitoring and Control Method, System and Storage Medium for Road Pavement

Abstract

This application relates to the technical field of road surface monitoring and control, and discloses a method, system and storage medium for deformation monitoring and control of road surfaces. The method includes: collecting multi-level sensing data of the road surface; preprocessing the data to obtain a time series data set; inputting the data into an LSTM network to predict deformation risks; determining the dominant factors to generate a hierarchical strategy; converting it into an execution instruction and responding according to the priority; constructing a knowledge graph based on the feedback, modeling it as a Markov process, and training a deep Q network to achieve adaptive regulation. This application realizes the accurate identification of the road surface deformation state, the early prediction of the deformation trend, the accurate analysis of the deformation cause, and the adaptive optimization of the control strategy, thereby improving the road maintenance efficiency, extending the road service life, and ensuring driving safety.

Description

Technical Field

[0001] This application relates to the technical field of road surface monitoring and control, and particularly to a method, system, and storage medium for deformation monitoring and control of road surfaces. Background Art

[0002] The deformation monitoring and control of road surfaces is a key link in the maintenance and management of transportation infrastructure, and is of great significance for ensuring the service life of roads and driving safety. Traditional road surface deformation monitoring mainly relies on manual inspections and fixed sensor networks. Manual inspections usually use equipment such as level gauges and displacement gauges for regular measurements, with a low data acquisition frequency and difficulty in capturing the dynamic process of road surface deformation; fixed sensor networks monitor through devices such as strain sensors and displacement sensors buried under the road surface. Although the acquisition frequency is relatively high, the coverage range is limited, and the installation cost is high. In terms of data processing, traditional methods mainly use statistical analysis and threshold judgment, lacking the ability to fuse and process multi-source heterogeneous data; in terms of deformation control, it mainly relies on after-the-fact repairs and regular maintenance, lacking an active prevention and precise intervention mechanism. Scholars at home and abroad have carried out a large amount of research on road surface deformation monitoring and control, such as the pavement management system (PMS) developed by the US Federal Highway Administration and the intelligent road monitoring network (IRMN) of the European Union, which have initially realized the remote monitoring and evaluation of road surface conditions.

[0003] However, there are still many deficiencies in the existing technology for road surface deformation monitoring and control: First, there is a lack of collaborative monitoring of multi-level sensor networks in the data acquisition link, making it difficult to comprehensively capture the spatio-temporal characteristics of road surface deformation; second, the ability to fuse and analyze multi-source heterogeneous data in the data processing link is limited, and the data value cannot be fully explored; third, there is a lack of support from advanced deep learning algorithms in the deformation risk prediction link, with insufficient prediction accuracy and lead time; in addition, the control strategy generation link lacks a precise identification of the deformation cause and a hierarchical response mechanism, and the control measures lack pertinence; finally, the system generally lacks the ability of self-learning and continuous optimization, and cannot continuously improve the control effect based on historical experience. These deficiencies make it difficult for the existing road surface deformation monitoring and control systems to cope with complex and changing road conditions, unable to achieve precise prediction and active control of road surface deformation, and thus affecting road use safety and maintenance efficiency. Summary of the Invention

[0004] This application provides a method, system, and storage medium for deformation monitoring and control of road surfaces, which are used to achieve precise identification of road surface deformation states, early prediction of deformation trends, accurate analysis of deformation causes, and adaptive optimization of control strategies, thereby improving road maintenance efficiency, extending the service life of roads, and ensuring driving safety.

[0005] First aspect, the present application provides a method for deformation monitoring and control of road pavement. The method for deformation monitoring and control of road pavement includes: collecting multi-level sensor data of the road pavement to obtain an original deformation data set including fixed monitoring point data, mobile monitoring unit data, and road surface images; preprocessing the original deformation data set to obtain a time series data set of road surface states; inputting the time series data set of road surface states into a long short-term memory network for risk prediction to obtain a rating result of road surface deformation risk level; determining the dominant deformation factors and generating a hierarchical control strategy according to the rating result of road surface deformation risk level; converting the hierarchical control strategy into a standardized execution instruction set, judging the priorities of the instructions in the execution instruction set, when the instruction priority is 1, starting an emergency response mechanism and sending an emergency intervention notice within 15 minutes; when the instruction priority is 2 to 4, sending an execution notice 24 hours in advance, collecting execution feedback and calculating the execution deviation rate; based on the execution feedback and execution deviation rate, constructing a road surface health knowledge graph, modeling the road surface deformation control process as a Markov decision process, and training a deep Q network to achieve adaptive traffic flow regulation.

[0006] Second aspect, the present application provides a system for deformation monitoring and control of road pavement. The system for deformation monitoring and control of road pavement includes:

[0007] An acquisition module, configured to collect multi-level sensor data of the road pavement to obtain an original deformation data set including fixed monitoring point data, mobile monitoring unit data, and road surface images;

[0008] A processing module, configured to preprocess the original deformation data set to obtain a time series data set of road surface states;

[0009] A prediction module, configured to input the time series data set of road surface states into a long short-term memory network for risk prediction to obtain a rating result of road surface deformation risk level;

[0010] A grading module, configured to determine the dominant deformation factors and generate a hierarchical control strategy according to the rating result of road surface deformation risk level;

[0011] A judgment module, configured to convert the hierarchical control strategy into a standardized execution instruction set, judge the priorities of the instructions in the execution instruction set, when the instruction priority is 1, start an emergency response mechanism and send an emergency intervention notice within 15 minutes; when the instruction priority is 2 to 4, send an execution notice 24 hours in advance, collect execution feedback and calculate the execution deviation rate;

[0012] A regulation module, configured to construct a road surface health knowledge graph based on the execution feedback and the execution deviation rate, model the road surface deformation control process as a Markov decision process, and train a deep Q-network to achieve adaptive regulation of traffic flow.

[0013] In a third aspect, there is provided a deformation monitoring and control device for a road surface, including: a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory to enable the deformation monitoring and control device for the road surface to execute the above-mentioned deformation monitoring and control method for the road surface.

[0014] In a fourth aspect, there is provided a computer-readable storage medium, in which instructions are stored, and when the instructions run on a computer, the computer is enabled to execute the above-mentioned deformation monitoring and control method for the road surface.

[0015] In the technical solution provided by this application, through multi-level sensor data collection on the road surface, an original deformation data set including fixed monitoring point data, mobile monitoring unit data, and road surface images is obtained, realizing comprehensive monitoring of the road surface state, avoiding the limitations of traditional single data sources, and greatly improving the monitoring coverage and data integrity; preprocessing the original deformation data set to obtain a time series data set of road surface states, effectively eliminating data noise and discontinuity, and improving the accuracy of subsequent analysis; inputting the time series data set of road surface states into a long short-term memory network for risk prediction to obtain the evaluation result of the road surface deformation risk level. As a kind of deep learning algorithm, the long short-term memory network can effectively capture the long-term dependence relationship in time series data and accurately predict the development trend of road surface deformation. Compared with the traditional threshold judgment method, the early warning lead time is significantly improved, providing an adequate time window for preventive maintenance; according to the evaluation result of the road surface deformation risk level, determining the dominant deformation factors and generating a hierarchical control strategy, realizing in-depth analysis from symptoms to causes, making the control measures more targeted, and avoiding resource waste and ineffective intervention; converting the hierarchical control strategy into a standardized execution instruction set, judging the priorities of each instruction in the execution instruction set, realizing differential response based on the degree of urgency. The emergency-level instruction can trigger an emergency response within 15 minutes, greatly improving the system's response ability to emergencies; based on the execution feedback and execution deviation rate, constructing a road surface health knowledge graph, modeling the road surface deformation control process as a Markov decision process, and training a deep Q network to realize adaptive traffic flow regulation. This link fully reflects the unique value of artificial intelligence algorithms in specific application fields. The knowledge graph technology converts discrete monitoring data and control experience into structured knowledge. As a representative algorithm in the field of reinforcement learning, the deep Q network can gradually learn the optimal control strategy through continuous trial and error and optimization, realizing the transformation from "manual control" to "intelligent control". As the system is used for a longer time, the control effect is continuously improved, significantly reducing the road maintenance cost, extending the road service life, and improving the driving safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0017] Figure 1 It is a schematic diagram of an embodiment of the deformation monitoring and control method for road surfaces in an embodiment of this application;

[0018] Figure 2 It is a schematic diagram of an embodiment of the deformation monitoring and control system for road surfaces in an embodiment of this application;

[0019] Figure 3 It is a structural schematic block diagram of the deformation monitoring and control device for road pavement in the embodiment of the present invention. Specific implementation manners

[0020] The embodiments of the present application provide a deformation monitoring and control method, system and storage medium for road pavement. Terms such as "first", "second", "third", "fourth", etc. (if any) in the specification, claims and above-mentioned drawings of the present application are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments described here can be implemented in an order different from that shown or described here. In addition, the terms "include" or "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0021] For ease of understanding, the specific process of the embodiments of the present application will be described below. Please refer to Figure 1 An embodiment of the deformation monitoring and control method for road pavement in the embodiments of the present application includes:

[0022] Step S101: Collect multi-level sensor data of the road pavement to obtain an original deformation data set including fixed monitoring point data, mobile monitoring unit data and road surface images;

[0023] Step S102: Preprocess the original deformation data set to obtain a road surface state time series data set;

[0024] Step S103: Input the road surface state time series data set into a long short-term memory network for risk prediction to obtain a road surface deformation risk level assessment result;

[0025] Step S104: Determine the deformation dominant factors according to the road surface deformation risk level assessment result and generate a hierarchical control strategy;

[0026] Step S105: Convert the hierarchical control strategy into a standardized execution instruction set, judge the priorities of the instructions in the execution instruction set. When the instruction priority is 1, start the emergency response mechanism and send an emergency intervention notice within 15 minutes; when the instruction priority is 2 to 4, send an execution notice 24 hours in advance, collect the execution feedback and calculate the execution deviation rate;

[0027] Step S106: Based on the execution feedback and the execution deviation rate, construct a road surface health knowledge graph, model the road surface deformation control process as a Markov decision process, and train a deep Q-network to achieve adaptive traffic flow regulation.

[0028] It can be understood that the execution entity of this application can be a deformation monitoring and control system for road surfaces, or it can also be a terminal or a server. Specifically, it is not limited here. In this embodiment of the application, the server is used as the execution entity for illustration.

[0029] Specifically, multi-level sensor data is collected for the road surface to obtain an original deformation data set including fixed monitoring point data, mobile monitoring unit data, and road surface images. In this step, strain sensors, displacement sensors, pressure sensors, and temperature and humidity sensors are buried every 50 meters on the road to form fixed monitoring points, and data is collected at a frequency of once per hour. At the same time, an acceleration sensor, a gyroscope, and a GPS positioning module are installed on a specific vehicle to form a mobile monitoring unit, and the vertical acceleration change value of the vehicle is recorded at a sampling rate of 100 Hz. The road surface image acquisition device scans the road comprehensively once a day to obtain high-definition road surface images. During the data collection process, when the data of a single sensor deviates from the historical average value by more than 30%, this data point will be marked and a repeated collection instruction will be triggered to ensure the accuracy of the data. When adverse weather conditions are detected, such as a rainfall of more than 50 mm / day or a temperature below -10°C, the data collection frequency will be automatically adjusted to four times per hour. The original deformation data set is preprocessed to obtain a time series data set of road surface states. The wavelet transform method is applied to the fixed monitoring point data to remove high-frequency noise, and the Kalman filter algorithm is applied to the mobile monitoring unit data to eliminate the vibration interference of the vehicle itself. The missing values in the data are filled by the linear interpolation method of adjacent time series points, and when there are too many consecutive missing data, the average value of the historical same-time period data is used for substitution. The processed data is standardized to uniformly convert data from different sources and with different dimensions to a standard normal distribution. After standardization, the data is subjected to spatio-temporal alignment processing. The mobile monitoring unit data is mapped according to the GPS coordinates and the road mileage stake number to locate it in the road surface grid unit. Finally, the multi-source data fusion is carried out by the evidence theory method, weights are assigned to the data from different sources, and the time series data set of road surface states is integrated through the feature layer fusion algorithm.

[0030] The pavement condition time-series dataset is input into a long short-term memory network for risk prediction to obtain the evaluation results of the pavement deformation risk level. First, perform time-domain analysis on the pavement condition time-series dataset to extract basic characteristic parameters such as the mean and standard deviation of pavement displacement, perform frequency-domain conversion through fast Fourier transform to extract frequency-domain features, and apply the wavelet packet decomposition method to obtain time-frequency features. After extracting the main feature representation through principal component analysis, the dimensionality-reduced feature set is input into the long short-term memory network. This network includes an input layer, two memory cell layers, and an output layer. The first memory cell layer processes the input data through forget gates, input gates, and output gates to generate a hidden state vector; the second memory cell layer generates high-level time-series features through cell state update and output control gates; finally, it is mapped to the future deformation prediction value through a fully connected output layer. Based on the prediction results, the pavement deformation risk is classified into safety level, attention level, warning level, danger level, and emergency level. According to the evaluation results of the pavement deformation risk level, determine the dominant factors of deformation and generate a hierarchical control strategy. Construct a multi-factor pavement deformation cause analysis model, including an environmental factor sub-model, a material factor sub-model, and a load factor sub-model, to analyze the contribution of each factor to the deformation. Through Bayesian network integrated analysis, output the contribution percentage of the dominant deformation cause. According to the dominant factors of deformation and the risk level, retrieve matching items from the predefined control scheme knowledge base to generate an initial control scheme set. Conduct a life-cycle cost analysis of the control schemes to screen out the scheme with the optimal cost-benefit ratio. Through integer programming algorithm for global optimization, generate a hierarchical control strategy including the implementation schedule, resource requirement list, expected effect, and emergency plan corresponding to different risk levels. Convert the hierarchical control strategy into a standardized execution instruction set, judge the priority of each instruction in the execution instruction set, and take corresponding measures according to the priority. The standardized execution instruction set includes six elements: the execution entity, execution time, execution location, execution content, execution standard, and acceptance requirement. The instructions are assigned priorities according to the risk level, with the priority of the emergency-level instruction being 1, the priority of the danger-level instruction being 2, and so on. When the instruction priority is 1, send an emergency intervention notice to the relevant responsible unit and activate the emergency resource scheduling system; when the instruction priority is 2 to 4, send an execution notice 24 hours in advance. Collect execution status data through multiple channels, calculate the execution deviation rate, and send an exception alarm and adjustment suggestions when the deviation rate exceeds the preset threshold.

[0031] Based on execution feedback and execution deviation rate, a pavement health knowledge graph is constructed, the pavement deformation control process is modeled as a Markov decision process, and a deep Q-network is trained to achieve adaptive traffic flow regulation. The pavement health knowledge graph includes five types of nodes: section information, deformation characteristics, environmental factors, intervention measures, and effect evaluation. The nodes are connected by causal relationships and temporal relationships. The state space of the Markov decision process is the pavement deformation feature vector, the action space is the set of optional control strategies, the transition probability is the change relationship between states, and the reward function is the comprehensive score of deformation control effect and resource consumption. The deep Q-network consists of three fully connected layers, with the number of neurons in each layer being 128, 64, and 32 respectively. The input layer receives the pavement state feature vector, and the output layer generates the Q-values of each control strategy. After training, the deep Q-network can analyze the current pavement deformation state and prediction trend, calculate the long-term benefits of different traffic flow allocation schemes, obtain the optimal traffic regulation strategy, and adjust the passing time, speed limit value, and vehicle passing ratio through traffic signal control and variable message signs to achieve adaptive traffic flow regulation for key deformed sections.

[0032] In the embodiments of the present application, by collecting multi-level sensor data of the road surface, an original deformation data set including fixed monitoring point data, mobile monitoring unit data, and road surface images is obtained, realizing comprehensive monitoring of the road surface state, avoiding the limitations of traditional single data sources, and greatly improving the monitoring coverage and data integrity; preprocessing the original deformation data set to obtain a time series data set of the road surface state, effectively eliminating data noise and discontinuity, and improving the accuracy of subsequent analysis; inputting the time series data set of the road surface state into a long short-term memory network for risk prediction to obtain the evaluation result of the road surface deformation risk level. As a kind of deep learning algorithm, the long short-term memory network can effectively capture the long-term dependence relationship in time series data and accurately predict the development trend of road surface deformation. Compared with the traditional threshold judgment method, the early warning lead time is significantly improved, providing a sufficient time window for preventive maintenance; according to the evaluation result of the road surface deformation risk level, determining the main deformation factors and generating a hierarchical control strategy, realizing in-depth analysis from symptoms to causes, making the control measures more targeted, and avoiding resource waste and ineffective intervention; converting the hierarchical control strategy into a standardized execution instruction set and judging the priorities of the instructions in the execution instruction set, realizing differential response based on the urgency. The emergency-level instructions can trigger an emergency response within 15 minutes, greatly improving the system's ability to respond to emergencies; based on the execution feedback and execution deviation rate, constructing a road surface health knowledge graph, modeling the road surface deformation control process as a Markov decision process, and training a deep Q network to realize adaptive traffic flow regulation. This link fully reflects the unique value of artificial intelligence algorithms in specific application fields. The knowledge graph technology converts discrete monitoring data and control experience into structured knowledge. As a representative algorithm in the field of reinforcement learning, the deep Q network can gradually learn the optimal control strategy through continuous trial and error and optimization, realizing the transformation from "manual control" to "intelligent control". As the system is used for a longer time, the control effect is continuously improved, significantly reducing the road maintenance cost, extending the service life of the road, and improving the driving safety.

[0033] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0034] Strain sensors, displacement sensors, pressure sensors, and temperature and humidity sensors are buried every 50 meters on the road to form fixed monitoring points, and sensor data is collected at a frequency of once per hour to obtain fixed monitoring point data;

[0035] An acceleration sensor, a gyroscope, and a GPS positioning module are installed on a specific vehicle to form a mobile monitoring unit, and the vertical acceleration change value of the vehicle is recorded at a sampling rate of 100 Hz to obtain mobile monitoring unit data;

[0036] Use a road surface image acquisition device to conduct a full - range scan of the road once a day, obtain high - definition images with a resolution of not less than 4K, and form road surface images;

[0037] Through the 4G / 5G wireless transmission module, upload the fixed monitoring point data, mobile monitoring unit data, and road surface images to the central data processing server in real - time;

[0038] Perform deviation detection on the single - sensor data in the original deformation dataset. When the data deviates from the historical average by more than 30%, mark the deviated data points and trigger a repeated acquisition instruction to ensure data validity;

[0039] Use an automatic weather station to record environmental parameters in real - time. When the rainfall exceeds 50mm / day or the temperature is lower than - 10°C, adjust the data acquisition frequency to four times per hour to obtain intensive monitoring data under harsh conditions.

[0040] Specifically, strain sensors, displacement sensors, pressure sensors, and temperature - humidity sensors are buried every 50 meters on the road to form fixed monitoring points. The strain sensor measures the deformation of the road surface material under the action of external forces. The displacement sensor directly measures the vertical displacement change of the road surface. The pressure sensor detects the pressure distribution borne by the road surface. The temperature - humidity sensor records the environmental temperature and humidity changes. These sensors collect data at a frequency of once per hour to form a time series. The data generated by each monitoring point includes a timestamp, a monitoring point ID, a measurement value, and a quality flag. The fixed monitoring point data directly reflects the deformation state of the road surface under static and dynamic loads and is the core data source for deformation monitoring. An acceleration sensor, a gyroscope, and a GPS positioning module are installed on a specific vehicle to form a mobile monitoring unit. The acceleration sensor is installed on the vehicle chassis, perpendicular to the road surface direction, and records the vertical acceleration change of the vehicle during driving at a sampling rate of 100Hz. This high - frequency sampling can capture subtle road surface unevenness, and the acceleration fluctuations generated when the vehicle passes through uneven roads directly reflect the road surface conditions. The gyroscope monitors the attitude change of the vehicle in three - dimensional space to supplement the acceleration data. The GPS positioning module records the precise position information of the vehicle in real - time and associates the acceleration data with the geographical location. The advantage of the mobile monitoring unit data is its wide coverage range, which can dynamically collect data for the entire road network and make up for the spatial limitations of fixed monitoring points.

[0041] The road surface image acquisition device is a high-resolution camera group installed on a special vehicle, which conducts an all-round scan of the road once a day to obtain high-definition images with a resolution of not less than 4K. During the image acquisition process, the camera works synchronously with the GPS to add accurate geographical location tags to each image. These high-definition images can identify surface deformation features such as cracks, potholes, and ruts after being processed, forming road surface image data. The road surface image data provides an intuitive form of deformation performance and complements the sensor data. The fixed monitoring point data, mobile monitoring unit data, and road surface images are uploaded to the central data processing server in real time through the 4G / 5G wireless transmission module. During the transmission process, data compression and block transmission technologies are adopted to ensure the efficient transmission of a large amount of data. For the mobile monitoring unit and the image acquisition device, the data is first stored in the local cache and then centrally uploaded when the vehicle enters the network coverage area to avoid data loss. After receiving the data, the server immediately stores and conducts a preliminary verification to prepare for subsequent processing.

[0042] Detecting the deviation of individual sensor data in the original deformation data set is an important link to ensure data quality. The deviation detection evaluates the data credibility by calculating the difference between the current data and the mean value of historical data at the same time period and the same location. The specific calculation method is to take the average value of the data in the same time period in the recent 30 days as the reference value and calculate the deviation percentage between the current data and the reference value. When the data deviates from the historical average by more than 30%, the system will automatically mark this data point as an abnormal point and trigger a repeated acquisition instruction. The repeated acquisition will conduct short-time intensive sampling of the abnormal sensor at a speed three times the normal frequency to verify whether the data abnormality is caused by changes in road surface conditions or sensor failures.

[0043] The automatic weather station is a special meteorological monitoring device set along the road, which records environmental parameters such as rainfall, temperature, humidity, and wind speed in real time. When extreme weather conditions are detected, such as rainfall exceeding 50 mm / day or temperature below -10°C, the system will automatically adjust the data acquisition frequency of all sensors from once per hour to four times per hour. This is because under extreme weather conditions, the risk of road surface deformation increases significantly, and more intensive monitoring is required to capture the change trend. This mechanism of dynamically adjusting the sampling frequency not only ensures the data density during critical periods but also avoids generating too much redundant data under normal circumstances.

[0044] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0045] Apply the wavelet transform method to the fixed monitoring point data to remove high-frequency noise at a specific frequency and obtain the filtered fixed monitoring point signal;

[0046] Apply the Kalman filter algorithm to the mobile monitoring unit data to eliminate the interference caused by the vehicle's own vibration and obtain a pure road surface response signal;

[0047] For the missing values in the filtered fixed monitoring point signals and pure road surface response signals, the linear interpolation method of time-series adjacent points is used for completion. When the continuous missing data exceeds the preset threshold, the mean value of the historical data in the same period is used for substitution to obtain a complete data sequence;

[0048] The complete data sequence is processed by the standardization method to uniformly convert data from different sources to the standard normal distribution, obtaining a standardized data set;

[0049] Perform spatio-temporal alignment processing on the standardized data set, establish a mapping relationship between the mobile monitoring unit data according to the GPS coordinates and the road mileage stake number, and locate it to the road surface grid unit of a specific size to obtain an aligned data set;

[0050] Apply the evidence theory method to the aligned data set for multi-source data fusion, assign different weights to the fixed monitoring point data, mobile monitoring unit data, and road surface images respectively, and integrate them through the feature layer fusion algorithm to form a time-series data set of road surface conditions.

[0051] Specifically, the preprocessing of the original deformation data set is a key link in the road surface deformation monitoring and control method. First, the wavelet transform method is applied to the fixed monitoring point data. Wavelet transform is a time-frequency analysis tool that can remove high-frequency noise while retaining the main features of the signal. In this method, with a cut-off frequency of 40 Hz, the Daubechies wavelet function is used to decompose the fixed monitoring point data. The specific operation is to decompose the original signal into approximation coefficients and detail coefficients. The approximation coefficients represent the low-frequency part of the signal, and the detail coefficients represent the high-frequency part. By setting a threshold, the approximation coefficients and some detail coefficients are retained, the noise components higher than the cut-off frequency are filtered out, and then the signal is reconstructed to obtain the filtered fixed monitoring point signal. This processing process effectively removes the high-frequency noise caused by environmental vibration, electrical interference, etc. that are not caused by road surface deformation, making the sensor data more accurately reflect the actual deformation of the road surface.

[0052] Applying the Kalman filter algorithm to the mobile monitoring unit data is to eliminate the interference caused by the vehicle's own vibration. The Kalman filter is a recursive state estimation algorithm, which is particularly suitable for processing dynamic systems containing random noise. During the processing, first establish the state space model of the vehicle-road surface system, and the state variables include road surface displacement, speed, and acceleration. Through the prediction step, the state at the current moment is estimated based on the state at the previous moment; then through the update step, the prediction value is corrected by combining the actual measurement value. The Kalman gain is dynamically adjusted according to the ratio of the prediction error and the measurement error. When the vehicle is driving on a smooth road surface, the model can learn and filter out the vibration characteristics of the vehicle itself, and only retain the response caused by the road surface unevenness, so as to obtain a pure road surface response signal.

[0053] The missing values in the filtered fixed monitoring point signals and the pure road surface response signals are filled in by the linear interpolation method of time-series adjacent points. This method uses the effective data at the time points before and after the missing point and calculates the interpolation according to the time weight. The specific operation is as follows: for a single missing point, take one effective point before and after it and perform weighted average according to the inverse of the time distance; for multiple consecutive missing points, use the effective points at both ends of the missing interval for linear interpolation. However, when the consecutive missing data exceeds the preset threshold (set to 5 data points), simple linear interpolation may lead to large errors, and at this time, the mean value of the historical data in the same period is used for substitution instead. That is, extract the data of the same monitoring point and the same time period (such as the same hour of the same day) from the historical database, calculate its mean value as the filling value of the missing interval. This method makes full use of the periodic characteristics of road deformation and obtains a more reasonable complete data sequence.

[0054] The complete data sequence is processed by the standardization method to solve the problem of inconsistent dimensions of different types of data. The standardization process uses the Z-score method to convert various types of data into a standard normal distribution with a mean of 0 and a standard deviation of 1. Calculate the mean and standard deviation for each data type separately, and then perform the conversion according to the formula (original value - mean) / standard deviation. For example, the unit of displacement sensor data is millimeters, and the unit of acceleration sensor data is m / s². After standardization, both types of data become dimensionless standard scores, which is convenient for subsequent comprehensive analysis. Standardization not only unifies the scale of the data but also highlights the outliers, making them easier to be identified in the subsequent analysis.

[0055] The spatio-temporal alignment process is carried out on the standardized data set, focusing on solving the spatial mapping problem between the fixed monitoring point data and the mobile monitoring unit data. First, the road is divided into grid cells of 1 meter × 1 meter as the basic spatial reference unit. The fixed monitoring point data is directly mapped to the grid cell where it is located; while the mobile monitoring unit data needs to establish a mapping relationship through the GPS coordinates and the road mileage stake number. In the specific process, use the longitude and latitude coordinates recorded by the vehicle-mounted GPS, combined with the road GIS database, to locate each measurement point to the nearest road grid cell. When multiple mobile measurement points are mapped to the same grid cell, take their average value as the representative value of this cell. In this way, the fixed point data and the mobile data are aligned with the unified spatial reference system, forming an aligned data set that is continuous in time and consistent in space.

[0056] Applying the Dempster-Shafer evidence theory method to the alignment dataset for multi-source data fusion is a crucial step in integrating information from different sources. The Dempster-Shafer evidence theory is suitable for handling uncertain and incomplete data. By defining the basic probability assignment function, belief function, and plausibility function, evidence from different sensors can be combined. In this method, based on the reliability and accuracy of each data source, basic weights of 0.6, 0.3, and 0.1 are assigned to the fixed monitoring point data, mobile monitoring unit data, and road surface images respectively. Then, through the feature-level fusion algorithm, the three types of data are integrated in the feature space to generate a feature vector containing 20 dimensions, covering key parameters such as road surface vertical displacement, displacement change rate, vibration frequency characteristics, etc. Finally, the fused feature vectors are reorganized according to the time series and processed using the sliding window method with a window size of 24 hours and a sliding step of 1 hour to form a continuous time series dataset of road surface states.

[0057] Taking the road surface monitoring of a certain highway as an example, the original data of the displacement sensor at a fixed monitoring point contains about 15% high-frequency noise components. After wavelet transform, the smoothness of the signal is significantly improved, and it can more clearly reflect the long-term settlement trend of the road surface. During the driving of the mobile monitoring vehicle, about 60% of the vertical vibration data recorded by its acceleration sensor comes from the vibration of the vehicle's own suspension system. After Kalman filtering, this part of the interference is effectively eliminated, highlighting the characteristic vibration caused by road surface potholes. During the data collection process, due to a communication interruption, there were 4 consecutive missing data points at a certain time. Through the linear interpolation method of adjacent points in time series, based on the values before and after the missing data, which were 2.3 mm and 2.5 mm respectively, the filled values of the missing points were calculated as 2.35 mm, 2.4 mm, 2.45 mm, and 2.475 mm, maintaining the continuity and trend of the data. For data from different sources, such as the displacement data range of the fixed point is 0 - 10 mm, and the acceleration data range of the mobile unit is 0 - 3 g. After standardization processing, all data are mapped to a similar distribution range, facilitating comprehensive comparison. In the data fusion stage, at the same location on a certain road section, the fixed monitoring point shows a displacement of 3.2 mm (1.5 after standardization), the mobile monitoring shows abnormal vibration acceleration (1.8 after standardization), and the road surface image shows slight cracks (1.0 after standardization). After fusion using the evidence theory method, the comprehensive score is 1.5, indicating a medium-level deformation risk at this location and requiring close monitoring.

[0058] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0059] Perform time-domain analysis on the time series dataset of road surface states, extract basic feature parameters including the mean value, standard deviation, peak value, valley value, peak-to-valley difference, and their change rates of road surface displacement to obtain the time-domain feature set;

[0060] The time-domain feature set is subjected to frequency-domain conversion through fast Fourier transform to extract the main frequency component, power spectral density, and frequency bandwidth, obtaining the frequency-domain feature set;

[0061] The wavelet packet decomposition method is applied to the frequency-domain feature set, with the decomposition level set to five layers, to obtain the energy distribution and its proportional relationship in each frequency band, obtaining the time-frequency feature set;

[0062] The main feature representation is extracted from the time-frequency feature set through the principal component analysis method, and the first five principal components are selected as the morphological feature description of the road surface deformation, obtaining the dimensionality-reduced feature set;

[0063] The dimensionality-reduced feature set is input into the input layer of the long short-term memory network in the order of time series. After being processed by the first memory unit layer including the forget gate, input gate, and output gate, a hidden state vector is obtained. Then, the cell state update and output control gate of the second memory unit layer are used to generate high-level temporal features. Finally, the future 24-hour deformation prediction value is mapped through the fully connected output layer, obtaining the deformation prediction result;

[0064] Based on the deformation prediction result, the road surface deformation risk is classified into levels. According to the comprehensive score in three dimensions of displacement, deformation rate, and prediction trend, the risk levels are divided into safe level, attention level, warning level, danger level, and emergency level, forming the road surface deformation risk level assessment result.

[0065] Specifically, perform time-domain analysis on the time-series dataset of road surface conditions to extract basic characteristic parameters. Time-domain analysis directly calculates statistical characteristics from time-series data, including the mean value of road surface displacement, standard deviation, peak value, valley value, peak-valley difference, and its change rate. The mean value of road surface displacement reflects the overall deformation level of the road surface and is obtained by taking the arithmetic mean of 24-hour continuous data; the standard deviation measures the degree of deformation fluctuation and characterizes the road surface stability; the peak value and valley value are the maximum and minimum displacement values within the time window respectively, reflecting extreme deformation conditions; the peak-valley difference is the difference between the peak value and the valley value, indicating the deformation amplitude; the change rate is the difference between the mean values of adjacent time windows divided by the time interval, reflecting the deformation development speed. These basic characteristic parameters together constitute the time-domain feature set, which intuitively describes the static and dynamic characteristics of road surface deformation. Applying fast Fourier transform to the time-domain feature set for frequency-domain conversion is to reveal the periodic characteristics hidden in the deformation data. Fast Fourier transform decomposes the time-domain signal into the superposition of sine waves of different frequencies and can identify deformation components of different frequencies. During the processing, apply fast Fourier transform to the 24-hour time-series data of each monitoring point to obtain a spectrogram, and then extract three types of key frequency-domain characteristics from it: the main frequency component, that is, the frequency point with the most concentrated energy, which is usually related to the main deformation factors; the power spectral density, which represents the energy distribution of each frequency component and reflects the deformation intensity; the frequency bandwidth, which represents the range of effective frequency components and reflects the deformation complexity. These frequency-domain characteristics form the frequency-domain feature set, which can distinguish different types of deformation patterns, such as high-frequency vibrations caused by traffic loads and low-frequency expansion and contraction caused by temperature changes.

[0066] Applying the wavelet packet decomposition method to the frequency-domain feature set is to obtain more refined time-frequency characteristics. Wavelet packet decomposition can analyze the signal at different scales, taking into account both time-domain and frequency-domain information. Set the decomposition level to five layers, which means dividing the spectrum into 32 equally wide frequency bands. Calculate the energy of each frequency band, that is, the sum of the squares of all coefficients within that frequency band, and then calculate the proportion of the energy of each frequency band in the total energy. This decomposition method can capture deformation characteristics at different scales and different frequency bands and is particularly suitable for analyzing non-stationary signals, such as the response characteristics of the road surface under different traffic flows and temperature conditions. The energy distribution and proportional relationship of each frequency band obtained by wavelet packet decomposition constitute the time-frequency feature set, which provides a richer description of deformation characteristics than simple time-domain or frequency-domain analysis.

[0067] Extracting the main feature representation from the time-frequency feature set through principal component analysis is a dimensionality reduction technique aimed at reducing the number of features while retaining the most crucial information. Principal component analysis transforms the original features into a set of linearly independent new features (principal components) through an orthogonal transformation. These principal components are sorted according to the variance magnitude, and the principal components with larger variances carry more information. In this method, principal component analysis is performed on the high-dimensional data set containing multiple time-frequency features, the eigenvalues and eigenvectors of the feature covariance matrix are calculated, and then the eigenvectors corresponding to the largest five eigenvalues are selected as the principal components. These five principal components cumulatively explain more than 85% of the variance of the original data. In this way, the high-dimensional time-frequency feature set is compressed into a five-dimensional dimensionality reduction feature set, and each principal component is a linear combination of the original features, representing a key aspect of pavement deformation.

[0068] Inputting the dimensionality reduction feature set into a long short-term memory network for risk prediction is to use deep learning technology to capture the temporal patterns of deformation. The long short-term memory network is a special type of recurrent neural network that is good at processing sequential data with long-term dependencies. The core of this network is the memory cell, which includes three control mechanisms: the forget gate, the input gate, and the output gate. The dimensionality reduction feature set is input into the input layer of the network in chronological order and then processed through the first memory cell layer: the forget gate determines which historical information needs to be discarded, the input gate determines which new information needs to be added, and the output gate controls the information output of the current state. These three gates jointly generate the hidden state vector. This vector is further processed through the second memory cell layer. The cell state update mechanism integrates long-term memory and short-term input, and the output control gate filters useful information to generate high-level temporal features. Finally, these features are converted into specific predicted values, that is, the pavement deformation trend in the next 24 hours, through the fully connected output layer. The entire network is trained through the backpropagation algorithm to minimize the error between the predicted value and the actual observed value.

[0069] Classifying the pavement deformation risk based on the deformation prediction results is a crucial step in transforming quantitative analysis into a decision-making basis. The risk classification considers three dimensions: the displacement amount, which directly represents the current deformation degree; the deformation rate, which represents the development speed of the deformation; and the prediction trend, which represents the future development direction of the deformation. Each of these three dimensions is scored, and then the weighted sum is obtained to get the comprehensive score. According to the comprehensive score, the risk level is divided into five levels: the safe level (green), indicating that the deformation is within the normal range and no intervention is required; the attention level (blue), indicating that the deformation is slightly abnormal and the monitoring frequency needs to be increased; the warning level (yellow), indicating that the deformation is obvious and preventive maintenance needs to be arranged; the dangerous level (orange), indicating that the deformation is severe and repair needs to be implemented as soon as possible; the emergency level (red), indicating that the deformation is extreme and emergency measures need to be taken immediately. This multi-level risk assessment method can intuitively reflect the pavement condition and provide clear guidance for the formulation of subsequent control strategies.

[0070] Taking the monitoring of a certain section of a national road as an example, the data processing process of a bridgehead bumping section is as follows: First, extract the time-domain features from the displacement sensor data for 72 consecutive hours. The calculated displacement mean is 2.8 mm, the standard deviation is 0.5 mm, the peak value is 3.9 mm (appearing during the afternoon traffic peak), the valley value is 1.8 mm (appearing during the early morning low-flow period), the peak-to-valley difference is 2.1 mm, and the displacement change rate is 0.2 mm / day, indicating that the deformation at this location is slowly intensifying. Convert these time-domain data to the frequency domain through fast Fourier transform. It is found that the main frequency component is located at 0.5 Hz, which coincides with the passing frequency of heavy vehicles. The power spectral density reaches the maximum value at this frequency point, and the frequency bandwidth is concentrated in the range of 0.3 - 0.8 Hz, indicating that the deformation is mainly caused by vehicle loads. Through five-layer wavelet packet decomposition, it is found that the energy is mainly concentrated in the low-frequency and medium-frequency bands. The energy proportion of the low-frequency band (0 - 0.2 Hz) is 35%, reflecting the characteristics of foundation settlement; the energy proportion of the medium-frequency band (0.2 - 0.6 Hz) is 45%, reflecting the influence of traffic loads; the energy of the high-frequency band is relatively low, only accounting for 20%. Perform principal component analysis on these time-frequency characteristics. The first five principal components respectively reflect five aspects: load response, temperature effect, material fatigue, foundation settlement, and seasonal variation. Input the time-series data formed by these five principal components into a long short-term memory network. After the network is trained with historical data, it is predicted that the displacement will increase to 3.5 mm within the next 24 hours, and the deformation rate will accelerate to 0.3 mm / day. Considering the three dimensions of the current displacement, deformation rate, and prediction trend for scoring, the risk level of this section of the road is rated as dangerous (orange), triggering the subsequent control strategy generation process, and arranging a pavement structural repair plan.

[0071] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0072] Construct a multi-factor pavement deformation cause analysis model, including an environmental factor sub-model, a material factor sub-model, and a load factor sub-model. Analyze the relationship between environmental parameters and deformation through multiple linear regression to obtain the contribution degree of environmental factors;

[0073] Evaluate the contribution degree of environmental factors, the degree of material aging, and the influence of traffic loads through a fuzzy inference system, calculate the functional relationship between the cumulative equivalent axle load repetitions and the deformation development rate, and obtain the action intensity of each factor;

[0074] Input the action intensity of each factor into a Bayesian network for integrated analysis, output the contribution percentage of the dominant deformation cause, and obtain the dominant deformation factor;

[0075] According to the dominant deformation factor and the pavement deformation risk level assessment result, retrieve matching items from the knowledge base containing predefined control schemes, generate corresponding control types for sections with different risk levels, and obtain an initial control scheme set;

[0076] Conduct a life cycle cost analysis on the initial control scheme set, calculate the long-term cost-benefit ratio of each scheme, screen out the scheme with the optimal cost-benefit ratio, and obtain the candidate control strategies;

[0077] Globally optimize the candidate control strategies through the integer programming algorithm, balance the allocation of the entire road network resources on the premise of meeting safety requirements, and generate a hierarchical control strategy that includes an implementation schedule, a resource requirement list, expected effects, and an emergency plan corresponding to different risk levels.

[0078] Specifically, construct a multi-factor road surface deformation cause analysis model, which includes three sub-models: an environmental factor sub-model, a material factor sub-model, and a load factor sub-model. The environmental factor sub-model considers parameters such as temperature cyclic variation, precipitation, and freeze-thaw cycle times, and establishes a quantitative relationship between environmental parameters and deformation through the multiple linear regression method. Multiple linear regression analysis estimates the regression coefficients by the least squares method, establishes a linear relationship between multiple environmental variables and deformation, and calculates the influence degree of each environmental factor on deformation. For example, the deformation increment corresponding to every 10-degree increase in temperature, the deformation increment corresponding to every 10-mm increase in precipitation, etc. These regression coefficients directly reflect the influence intensity of each environmental factor, so as to obtain the contribution degree of environmental factors. Conduct a fuzzy inference system evaluation on the contribution degree of environmental factors, material aging degree, and traffic load influence to handle the uncertainty and fuzziness existing in these factors. The fuzzy inference system first converts the accurate numerical values of each factor into fuzzy sets. For example, the material aging degree is divided into three fuzzy levels: "slight", "medium", and "severe". Then, it conducts inferences through a predefined fuzzy rule base. The rule base contains rules in the form of "if... then...", such as "if the temperature change is large and the material aging is severe, then the comprehensive influence of the environment and materials is strong". For the traffic load influence, it quantifies by calculating the functional relationship between the cumulative equivalent axle load repetitions and the deformation development rate. The cumulative equivalent axle load repetitions is the cumulative number of times that different types of vehicles are converted into the standard axle load, and it is calculated according to the vehicle type, axle weight, and passing frequency. Establish a functional relationship between the cumulative equivalent axle load repetitions and the observed deformation development rate to obtain the sensitivity coefficient of the load intensity and deformation. The fuzzy inference system finally outputs the action intensity of each factor, expressed in clear quantitative values.

[0079] Integrating the action intensities of various factors through inputting them into a Bayesian network for analysis is a probabilistic graphical model method that can express the conditional dependence relationships between variables. A Bayesian network consists of nodes and directed edges. The nodes represent random variables (each influencing factor), and the directed edges represent causal relationships. The network contains two parts of information: the network structure (graph) and the conditional probability table. The network structure is predefined based on domain knowledge. For example, environmental factors affect material properties, and material properties and loads jointly affect deformation, etc. The conditional probability table defines the probability distribution of the child node given the states of the parent nodes and is obtained through training with historical data. After inputting the observed action intensities of various factors, the Bayesian network calculates the posterior probability of each factor on deformation through probabilistic inference, and finally outputs the contribution percentage of the dominant cause of deformation, clearly indicating which factor is the dominant cause of deformation. According to the dominant cause of deformation and the evaluation result of the pavement deformation risk level, matching items are retrieved from the predefined control scheme knowledge base to generate corresponding control types for road sections with different risk levels. The control scheme knowledge base contains 250 predefined schemes, covering combinations of different deformation causes, different deformation degrees, and different urgency levels. The retrieval process uses a multi-condition matching algorithm to calculate the similarity between the dominant cause of deformation and the risk level of the current road section and the cases in the knowledge base, and selects several schemes with the highest similarity. For example, for slight cracks (attention level) caused by temperature stress, the retrieval results include sealing treatment schemes; for moderate deformation (warning level) caused by foundation settlement, the retrieval results include base reinforcement schemes; for severe cracking (hazard level) caused by material aging, the retrieval results include milling and resurfacing schemes. The retrieved schemes form the initial control scheme set, providing candidates for subsequent optimization.

[0080] Perform a life cycle cost analysis on the initial control scheme set and calculate the long-term cost-benefit ratio of each scheme. The life cycle cost analysis considers the total life cycle cost of the scheme, including four parts: the initial construction cost, the maintenance cost, the user delay cost, and the risk cost. The initial construction cost is the direct cost of implementing the scheme; the maintenance cost is the cumulative cost of subsequent regular maintenance; the user delay cost is the economic loss caused by traffic delays due to construction; the risk cost is the expected value of the losses that may be caused by the failure of the scheme. The future costs are converted into present values through a discount rate to calculate the total life cycle cost. At the same time, evaluate the benefits of the scheme, including extending the service life, improving the service quality, etc., and quantify them as economic benefits. Divide the benefits by the costs to obtain the cost-benefit ratio, and select the scheme with the optimal cost-benefit ratio as the candidate control strategy, taking into account both technical feasibility and economic rationality.

[0081] Globally optimizing the candidate control strategy through integer programming algorithms is to reasonably allocate limited maintenance resources on the premise of meeting the safety requirements of the entire road network. Integer programming is a mathematical optimization method where decision variables can only take integer values and is suitable for dealing with resource allocation problems. In this solution, an integer programming model is established. The objective function is to minimize the total risk or total cost, and the constraint conditions include budget limitations, human resource limitations, equipment resource limitations, time limitations, etc. By solving this model, the optimal control strategy combination for each section of the entire road network is obtained, including the implementation sequence, resource allocation, etc. The finally generated hierarchical control strategy includes four parts: an implementation schedule (when to implement what measures), a resource requirement list (required manpower, materials, equipment), expected effects (performance indicators after repair), and an emergency plan (alternative plan when the main plan fails).

[0082] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0083] Convert the hierarchical control strategy into a standardized execution instruction set containing six elements: the execution entity, execution time, execution location, execution content, execution standard, and acceptance requirements. Among them, the execution entity is divided into two categories: an automatic execution system and manual intervention;

[0084] Assign priorities to the standardized execution instruction set according to the risk level. Set the priority of emergency-level instructions to 1, the priority of dangerous-level instructions to 2, the priority of warning-level instructions to 3, and the priority of attention-level instructions to 4 to obtain the priority sorting result;

[0085] Judge the instruction type according to the priority sorting result. When the instruction priority is 1, send an emergency intervention notice containing the intervention location, intervention content, and intervention standard to the relevant responsible unit, and activate the emergency resource scheduling system to obtain an emergency response record;

[0086] Judge the instruction type according to the priority sorting result. When the instruction priority is 2 to 4, send an execution notice containing the execution plan, resource requirements, and acceptance criteria to the relevant responsible unit according to the scheduled time plan to obtain a regular execution record;

[0087] Collect the execution status data of the emergency response record and the regular execution record through multiple channels, including the automatic execution status data and the manual intervention progress data, and compare and analyze the execution progress with the planned progress to obtain an execution feedback;

[0088] Calculate the ratio of the actual execution time to the planned execution time based on the execution feedback to obtain an execution deviation rate, and when the execution deviation rate exceeds the preset threshold, send an exception alarm and adjustment suggestions to the management personnel.

[0089] Specifically, converting the hierarchical control strategy into a standardized execution instruction set is the execution link in the road pavement deformation monitoring and control method. The standardized execution instruction set includes six core elements: the execution entity, execution time, execution location, execution content, execution standard, and acceptance requirements. The execution entity is divided into two categories: the automatic execution system and manual intervention. The automatic execution system includes non-physical intervention devices such as traffic flow control equipment, information release platforms, and environmental parameter monitoring devices; manual intervention includes maintenance construction teams, professional inspection personnel, road administration management personnel, etc., who are responsible for specific physical maintenance work. The execution time defines the start time and completion deadline of the instruction, accurate to the hour level. The execution location is accurately positioned through highway numbers, mileage markers, and GPS coordinates, and is accompanied by electronic map markings. The execution content details the specific measures to be taken, such as "milling 3 cm of the asphalt surface course and repaving with modified asphalt mixture", etc. The execution standard stipulates the operation specifications and quality requirements, such as "the deviation of milling flatness does not exceed 3 mm / 3 m, and the compaction degree of the newly paved surface layer is not less than 98%", etc. The acceptance requirements clarify the inspection methods and qualified standards after completion, such as "testing with a deflectometer, and the deflection value does not exceed 0.4 mm", etc. The conversion process maps the maintenance plan elements in the hierarchical control strategy into the standardized instruction format through a structured template to form a complete execution instruction set. The priority assignment adopts a direct mapping method, converting the five-level risk level into a four-level priority. The priority of the emergency-level instruction (corresponding to the red risk) is set to 1, indicating the highest priority and requiring immediate handling; the priority of the dangerous-level instruction (corresponding to the orange risk) is set to 2, indicating a high priority and requiring prompt handling; the priority of the warning-level instruction (corresponding to the yellow risk) is set to 3, indicating a medium priority and being processed according to the plan; the priority of the attention-level instruction (corresponding to the blue risk) is set to 4, indicating a low priority and can be postponed for handling. No execution instructions are generated for the safe-level sections. The priority assignment process also considers the importance coefficient of the section, and appropriately raises the priority of the instructions for key locations such as important arterial roads and transportation hubs. After the assignment is completed, the instructions are sorted from high to low according to the priority to obtain the priority sorting result, providing a basis for subsequent differential processing.

[0090] Judge the instruction type according to the priority sorting result, and adopt an emergency response mechanism for instructions with a priority of 1. When an instruction with a priority of 1 is recognized, immediately initiate the emergency intervention process and send an emergency intervention notice to the relevant responsible units. The emergency intervention notice is sent simultaneously in multiple ways such as text messages, phone calls, and work platform pushes to ensure that the information is quickly delivered. The notice content includes three core elements: the intervention location (accurately describe the location of the problem section), the intervention content (detailed description of the emergency measures to be taken), and the intervention standard (specify the quality requirements for emergency disposal). At the same time, activate the emergency resource scheduling process, mobilize personnel, equipment, and materials from the nearest maintenance base, and give priority to ensuring the resource requirements for emergency tasks. The entire notification process is completed within 15 minutes after detecting an emergency problem to ensure a quick response. The entire emergency response process is recorded, including the notice sending time, receipt confirmation time, resource scheduling situation, etc., to form an emergency response record for subsequent tracking and evaluation.

[0091] For instructions with priorities from 2 to 4, adopt the regular execution process for processing. According to the priority sorting result, send an execution notice to the relevant responsible units 24 hours in advance according to the pre-scheduled time plan. The execution notice content is more comprehensive, including the execution plan (detailed operation schedule and step arrangement), resource requirements (specific quantity and specifications of the required personnel, equipment, and materials), and acceptance criteria (indicators and methods for completion acceptance). The regular execution notice is sent through the work platform, email, etc. to ensure the complete transmission of information. After receiving the notice, the responsible unit makes resource preparations and operation arrangements as required and starts execution at the planned time. Each notice and the corresponding execution process form an execution record, including key information such as the notice time, planned execution time, actual start time, etc., to form a regular execution record, providing a data basis for execution monitoring. Collecting execution status data through multiple channels is an important link to ensure the effective implementation of control measures. The data collection channels are divided into two categories: automatic execution status data is automatically uploaded by Internet of Things devices, including the working status of traffic control devices, information update of electronic displays, etc.; manual intervention progress data is uploaded in real time by on-site personnel through a mobile terminal application, including text descriptions, on-site photos, measured values of key parameters, etc. On-site personnel need to update the progress information at each key node (such as start, half completion, completion, etc.), and increase the reporting frequency in special cases. All the collected data is summarized to the execution monitoring platform and automatically compared and analyzed with the planned progress. The comparison process calculates the completion time difference and quality index deviation at each node to form structured execution feedback data, providing a basis for subsequent evaluation and adjustment.

[0092] Calculating the execution deviation rate based on the execution feedback is the core step in quantifying the execution effect. The execution deviation rate is mainly calculated by the ratio of the actual execution time to the planned execution time, indicating the degree of lag or advance in the work progress. The specific calculation method is as follows: for each execution node, subtract the planned completion time from the actual completion time to obtain the time difference, then divide it by the total planned execution duration, and multiply by 100% to get the execution deviation rate of this node. By comprehensively considering the deviation rates of all nodes and performing weighted averaging, the overall execution deviation rate is obtained. When the execution deviation rate exceeds the preset threshold (usually set at 20%), an abnormal alarm mechanism is triggered, and an alarm message containing an analysis of the abnormal reasons and adjustment suggestions is sent to the road surface management personnel. The adjustment suggestions include resource supplementation plans, construction period adjustment plans, technical method optimization plans, etc., to help the management personnel make quick decisions. The abnormal alarm is sent through instant messaging methods such as text messages and phone calls to ensure that the problem is handled in a timely manner.

[0093] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0094] Based on the execution feedback and the execution deviation rate, construct a road surface health knowledge graph containing five types of nodes: road section information, deformation characteristics, environmental factors, intervention measures, and effect evaluation. The nodes are connected by edges representing causal relationships and temporal relationships to obtain a deformation control knowledge base;

[0095] Model the road surface deformation control process as a Markov decision process, where the state space is the road surface deformation feature vector, the action space is the set of optional control strategies, the transition probability is the change relationship between states, and the reward function is the comprehensive score of the deformation control effect and resource consumption, to obtain a decision process model;

[0096] Construct a deep Q-network containing three fully connected layers, with the number of neurons in each layer being one hundred and twenty-eight, sixty-four, and thirty-two respectively. The input layer receives the road surface state feature vector, and the output layer generates the Q-values of each control strategy to obtain a reinforcement learning model structure;

[0097] Adopt an experience replay mechanism to train the reinforcement learning model structure, set the size of the replay buffer to ten thousand records, update the target network parameters every five hundred interactions, and optimize the Q-value estimation through the gradient descent algorithm to obtain a trained deep Q-network;

[0098] Based on the trained deep Q-network, perform dynamic analysis on the traffic flow. According to the current road surface deformation state and the predicted trend, calculate the long-term benefits of different traffic flow allocation schemes to obtain the optimal traffic regulation strategy;

[0099] Send the optimal traffic regulation strategy to the traffic signal control system and variable message signs to dynamically adjust the passing time, speed limit value, and vehicle passing ratio, and achieve adaptive regulation of the traffic flow on key deformed road sections.

[0100] Specifically, the road surface health knowledge graph consists of five types of core nodes: road section information nodes, deformation feature nodes, environmental factor nodes, intervention measure nodes, and effect evaluation nodes. Road section information nodes record the basic attributes of the road, including road grade, paving material, service life, and historical maintenance records; deformation feature nodes store technical indicators such as the displacement amount, deformation rate, and vibration frequency characteristics of the road surface; environmental factor nodes contain environmental data such as temperature cyclic changes, precipitation, and freeze-thaw cycle times; intervention measure nodes record the executed road surface maintenance operations, such as specific technical means like crack sealing and road surface milling and resurfacing; effect evaluation nodes save the state comparison data before and after the intervention, such as improvement rate and stability index. In the knowledge graph, the nodes are connected by two types of edges: causal relationship edges and temporal relationship edges. Causal relationship edges describe the influence mechanism between different nodes. For example, a "causes" relationship edge is established between the environmental factor node with precipitation exceeding 50 mm / day and the deformation feature node with the road surface displacement increasing by 0.8 mm / day; temporal relationship edges record the sequence of events. For example, a "preceding" relationship edge is established between the intervention measure node of road surface milling and resurfacing and the effect evaluation node with a displacement reduction of 85%. Through this structured representation, the knowledge graph transforms discrete monitoring data and control experience into a structured deformation control knowledge base.

[0101] Modeling the road surface deformation control process as a Markov decision process is a key step in achieving adaptive control. In this model, the state space is composed of the road surface deformation feature vector, which contains multi-dimensional parameters such as displacement amount, deformation rate, and vibration characteristics; the action space corresponds to the set of optional control strategies, covering all technical means from mild interventions (such as crack sealing) to severe interventions (such as road surface reconstruction); the transition probability represents the change law of the road surface state after adopting a specific control strategy; the reward function comprehensively evaluates the effectiveness of the control strategy based on the trade-off between deformation control effect and resource consumption through the weighted summation method, with the weight ratio being 70% for the effect and 30% for resource consumption.

[0102] The deep Q-network is the core algorithm in reinforcement learning for learning the optimal control strategy. The network structure includes three fully connected layers. The first layer has 128 neurons, which receive the road surface state feature vector as input; the second layer has 64 neurons, responsible for feature extraction and transformation; the third layer has 32 neurons, which generate the Q-values of each control strategy. Each neuron uses the ReLU activation function to improve the non-linear expression ability of the network. The dimension of the road surface state feature vector received by the input layer is 20, including key indicators such as displacement amount, deformation rate, and vibration frequency; the output layer generates the Q-values of each control strategy, and the Q-value represents the long-term expected return of selecting a specific control strategy in the current state.

[0103] The experience replay mechanism is a core technology for training the Deep Q-Network. It improves the learning efficiency by storing and reusing historical interaction data. The size of the replay buffer is set to 10,000 records, and each record contains four elements: state, action, reward, and next state. During the training process, the system updates the target network parameters every 500 interactions to ensure the stability of the learning process. The Q-value estimation is optimized by the gradient descent algorithm, with a learning rate of 0.001, a batch size of 32, and 10,000 training epochs. Finally, a Deep Q-Network that can select the optimal control strategy under various road conditions is obtained.

[0104] Based on the trained Deep Q-Network, the system can dynamically analyze and regulate traffic flow. When a deformation risk is detected on the road surface, the system will calculate the long-term benefits of each traffic flow allocation plan according to the current road surface deformation state and prediction trend, while considering the impact of different traffic flow allocation plans on the road surface. For plans with a high proportion of heavy vehicles and significant damage to the road surface, the system will assign a lower Q-value; for plans that can reduce the road surface load and delay the development of deformation, the system will assign a higher Q-value. By comparing the Q-values of different plans, the system selects the optimal traffic regulation strategy.

[0105] Issuing the optimal traffic regulation strategy to traffic control devices is the last link to achieve adaptive regulation. The system sends regulation instructions through the interfaces with the traffic signal control system and variable message signs to dynamically adjust the passing time, speed limit value, and vehicle passing proportion. For sections with a high deformation risk, the system will reduce the passing frequency of heavy vehicles to reduce the road surface load; for sections where the deformation has reached the warning level, the system will implement a speed limit measure to limit the vehicle speed within a range that has less impact on the road surface; for sections approaching the dangerous level, the system will guide some traffic flows to detour to reduce the road surface load.

[0106] In practical applications, after continuous rainfall, the road surface displacement sensor on section K1 of a certain main road shows that the deformation rate reaches 0.7 mm / day. The system retrieves historical cases of similar road conditions through the knowledge graph and finds that the deformation is highly correlated with the passing frequency of heavy vehicles. After analysis by the Deep Q-Network, it is calculated that reducing the passing volume of heavy vehicles by 30% can control the deformation rate below 0.3 mm / day while having the least impact on the overall traffic efficiency. Based on this analysis result, the system sends an instruction to the traffic signal control system to shorten the green light time for heavy vehicles on this section by 15 seconds, and at the same time displays a detour suggestion for heavy vehicles on the variable message sign, achieving precise traffic flow regulation for the deformed section, effectively delaying the development of road surface deformation, avoiding serious damage, and extending the service life of the road.

[0107] The above describes the method for deformation monitoring and control of road surfaces in the embodiments of the present application. Next, the deformation monitoring and control system for road surfaces in the embodiments of the present application will be described. Please refer toFigure 2 , in an embodiment of the deformation monitoring and control system for road pavement in the embodiments of the present application, it includes:

[0108] A collection module, configured to collect multi-level sensor data of the road pavement to obtain an original deformation data set including fixed monitoring point data, mobile monitoring unit data, and road surface images;

[0109] A processing module, configured to preprocess the original deformation data set to obtain a time series data set of road surface states;

[0110] A prediction module, configured to input the time series data set of road surface states into a long short-term memory network for risk prediction to obtain a rating result of road surface deformation risk level;

[0111] A grading module, configured to determine the dominant deformation factors and generate a hierarchical control strategy according to the rating result of the road surface deformation risk level;

[0112] A judgment module, configured to convert the hierarchical control strategy into a standardized execution instruction set, judge the priority of each instruction in the execution instruction set. When the instruction priority is 1, start an emergency response mechanism and send an emergency intervention notice within 15 minutes; when the instruction priority is 2 to 4, send an execution notice 24 hours in advance, collect execution feedback and calculate the execution deviation rate;

[0113] A regulation module, configured to construct a road surface health knowledge graph based on the execution feedback and the execution deviation rate, model the road surface deformation control process as a Markov decision process, and train a deep Q-network to achieve adaptive regulation of traffic flow.

[0114] Through the collaborative cooperation of the above-mentioned various components, by collecting multi-level sensor data on the road surface, an original deformation dataset including fixed monitoring point data, mobile monitoring unit data, and road surface images is obtained, realizing the comprehensive monitoring of the road surface state, avoiding the limitations of traditional single data sources, and greatly improving the monitoring coverage and data integrity; preprocessing the original deformation dataset to obtain a time-series dataset of the road surface state, effectively eliminating data noise and discontinuity, and improving the accuracy of subsequent analysis; inputting the time-series dataset of the road surface state into a long short-term memory network for risk prediction to obtain the evaluation result of the road surface deformation risk level. As a type of deep learning algorithm, the long short-term memory network can effectively capture long-term dependencies in time-series data and accurately predict the development trend of road surface deformation. Compared with traditional threshold judgment methods, the early warning lead time is significantly improved, providing an adequate time window for preventive maintenance; according to the evaluation result of the road surface deformation risk level, determining the dominant deformation factors and generating a hierarchical control strategy, realizing in-depth analysis from symptoms to causes, making control measures more targeted, and avoiding resource waste and ineffective intervention; converting the hierarchical control strategy into a standardized execution instruction set and judging the priorities of each instruction in the execution instruction set, realizing differential response based on the degree of urgency. Emergency-level instructions can trigger an emergency response within 15 minutes, greatly improving the system's ability to respond to emergencies; based on execution feedback and execution deviation rate, constructing a road surface health knowledge graph, modeling the road surface deformation control process as a Markov decision process, and training a deep Q-network to achieve adaptive traffic flow regulation. This link fully reflects the unique value of artificial intelligence algorithms in specific application fields. The knowledge graph technology converts discrete monitoring data and control experience into structured knowledge. As a representative algorithm in the field of reinforcement learning, the deep Q-network can gradually learn the optimal control strategy through continuous trial and error and optimization, realizing the transformation from "manual control" to "intelligent control". As the system is used for a longer time, the control effect is continuously improved, significantly reducing the road maintenance cost, extending the road service life, and improving driving safety.

[0115] Above Figure 2 From the perspective of modular functional entities, the deformation monitoring and control system for road surfaces in the embodiments of the present invention is described in detail. Next, the deformation monitoring and control device for road surfaces in the embodiments of the present invention is described in detail from the perspective of hardware processing.

[0116] Figure 3FIG. 0 is a schematic structural diagram of a deformation monitoring and control device for a road surface provided by an embodiment of the present invention. The deformation monitoring and control device 300 for the road surface may vary greatly due to different configurations or performances, and may include one or more processors (central processing units, CPUs) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 for storing application programs 333 or data 332 (for example, one or more mass storage device terminals). Among them, the memory 320 and the storage media 330 may be transient storage or persistent storage. The program stored in the storage media 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations for the deformation monitoring and control device 300 for the road surface. Further, the processor 310 may be configured to communicate with the storage media 330 and execute a series of instruction operations in the storage media 330 on the deformation monitoring and control device 300 for the road surface to implement the steps of the above-mentioned deformation monitoring and control method for the road surface.

[0117] The deformation monitoring and control device 300 for the road surface may further include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art can understand that Figure 3 the shown structural diagram of the deformation monitoring and control device for the road surface does not constitute a limitation on the deformation monitoring and control device for the road surface provided by the present invention, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0118] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions run on a computer, the computer is caused to execute the steps of the deformation monitoring and control method for the road surface.

[0119] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, systems, and units may refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0120] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions to enable a deformation monitoring and control device for road pavement (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0121] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of various embodiments of the present invention.

Claims

1. A method for deformation monitoring and control of road surfaces, characterized in that, The method includes: Performing multi-level sensor data collection on the road surface to obtain an original deformation data set including fixed monitoring point data, mobile monitoring unit data, and road surface images; Preprocessing the original deformation data set to obtain a road surface state time series data set, including: applying the wavelet transform method to the fixed monitoring point data to remove high-frequency noise at a specific frequency to obtain a filtered fixed monitoring point signal; applying the Kalman filtering algorithm to the mobile monitoring unit data to eliminate interference caused by the vehicle's own vibration to obtain a pure road surface response signal; using the linear interpolation method of time series adjacent points to complement the missing values in the filtered fixed monitoring point signal and the pure road surface response signal, and when the continuous missing data exceeds a preset threshold, using the mean value of historical data in the same time period to replace it to obtain a complete data sequence; processing the complete data sequence through a standardization method to uniformly convert data from different sources to a standard normal distribution to obtain a standardized data set; performing spatio-temporal alignment processing on the standardized data set, establishing a mapping relationship between the mobile monitoring unit data according to the GPS coordinates and the road mileage stake number, and positioning it to a road surface grid unit of a specific size to obtain an aligned data set; applying the evidence theory method to the aligned data set for multi-source data fusion, assigning different weights to the fixed monitoring point data, mobile monitoring unit data, and road surface images respectively, and integrating and forming the road surface state time series data set through a feature layer fusion algorithm; Inputting the road surface state time series data set into a long short-term memory network for risk prediction to obtain a road surface deformation risk level assessment result; Determining the deformation dominant factors and generating a hierarchical control strategy according to the road surface deformation risk level assessment result; Converting the hierarchical control strategy into a standardized execution instruction set, judging the priorities of the instructions in the execution instruction set, and when the instruction priority is 1, starting an emergency response mechanism and sending an emergency intervention notice within 15 minutes; When the instruction priority is 2 to 4, sending an execution notice 24 hours in advance, collecting execution feedback and calculating the execution deviation rate; Based on the execution feedback and execution deviation rate, construct a pavement health knowledge graph, model the pavement deformation control process as a Markov decision process, train a deep Q-network to achieve adaptive traffic flow regulation. Based on the execution feedback and the execution deviation rate, construct a pavement health knowledge graph including five types of nodes: road section information, deformation characteristics, environmental factors, intervention measures, and effect evaluation. The nodes are connected by edges of causal relationships and temporal relationships to obtain a deformation control knowledge base; model the pavement deformation control process as a Markov decision process, with its state space being the pavement deformation feature vector, the action space being the set of optional control strategies, the transition probability being the change relationship between states, and the reward function being the comprehensive score of deformation control effect and resource consumption, to obtain a decision process model; construct a deep Q-network including three fully connected layers, with the number of neurons in each layer being 128, 64, and 32 respectively. The input layer receives the pavement state feature vector, and the output layer generates the Q-values of each control strategy to obtain a reinforcement learning model structure; use the experience replay mechanism to train the reinforcement learning model structure, set the size of the replay buffer to 10,000 records, update the target network parameters every 500 interactions, and optimize the Q-value estimation through the gradient descent algorithm to obtain a trained deep Q-network; based on the trained deep Q-network, conduct dynamic analysis of traffic flow, calculate the long-term benefits of different traffic flow allocation schemes according to the current pavement deformation state and prediction trend, and obtain the optimal traffic regulation strategy. Among them, based on the trained deep Q-network, when a pavement deformation risk is detected, according to the current pavement deformation state and prediction trend, and considering the impact of different traffic flow allocation schemes on the pavement at the same time, calculate the long-term benefits of each scheme. For the scheme with a high proportion of heavy vehicles and large damage to the pavement, give a lower Q-value; for the scheme that can reduce the pavement load, give a higher Q-value. By comparing the Q-values of different schemes, select the optimal traffic regulation strategy; send the optimal traffic regulation strategy to the traffic signal control system and variable message signs to dynamically adjust the passing time, speed limit value, and vehicle passing proportion, and achieve adaptive traffic flow regulation for key deformed road sections.

2. The deformation monitoring and control method for road pavement according to claim 1, characterized in that, The multi-level sensor data collection for the road pavement to obtain the original deformation data set including fixed monitoring point data, mobile monitoring unit data, and pavement images, includes: Bury strain sensors, displacement sensors, pressure sensors, and temperature and humidity sensors every 50 meters on the road to form fixed monitoring points, and collect sensor data at a frequency of once per hour to obtain the fixed monitoring point data; Install an acceleration sensor, a gyroscope, and a GPS positioning module on a specific vehicle to form a mobile monitoring unit, and record the vertical acceleration change value of the vehicle at a sampling rate of 100 Hz to obtain the mobile monitoring unit data; Use a pavement image acquisition device to conduct a full-range scan of the road once a day to obtain high-definition images with a resolution of not less than 4K to form the pavement images; The fixed monitoring point data, the mobile monitoring unit data, and the road surface images are uploaded to the central data processing server in real time through the 4G / 5G wireless transmission module; Deviation detection is performed on the single-sensor data in the original deformation dataset. When the data deviates from the historical average by more than 30%, the deviated data points are marked and a repeated acquisition instruction is triggered to ensure data validity; An automatic weather station is used to record environmental parameters in real time. When the rainfall exceeds 50 mm / day or the temperature is lower than -10°C, the data acquisition frequency is adjusted to four times per hour to obtain intensive monitoring data under harsh conditions.

3. The deformation monitoring and control method for road pavement according to claim 1, characterized in that, The road surface state time series dataset is input into a long short-term memory network for risk prediction to obtain the road surface deformation risk level assessment result, including: Perform time-domain analysis on the road surface state time series dataset, extract basic characteristic parameters including the mean value, standard deviation, peak value, valley value, peak-valley difference, and their change rates of road surface displacement to obtain the time-domain feature set; Perform frequency-domain conversion on the time-domain feature set through fast Fourier transform, extract the main frequency component, power spectral density, and frequency bandwidth to obtain the frequency-domain feature set; Apply the wavelet packet decomposition method to the frequency-domain feature set, set the decomposition layer to five layers, and obtain the energy distribution and its proportional relationship of each frequency band to obtain the time-frequency feature set; Extract the main feature representation from the time-frequency feature set through the principal component analysis method, select the first five principal components as the morphological feature description of road surface deformation to obtain the dimensionality-reduced feature set; Input the dimensionality-reduced feature set into the input layer of the long short-term memory network in the order of time series. After being processed by the first memory unit layer including the forget gate, input gate, and output gate, a hidden state vector is obtained. Then, the cell state of the second memory unit layer is updated and the output control gate generates high-level time series features. Finally, it is mapped to the deformation prediction value for the next 24 hours through the fully connected output layer to obtain the deformation prediction result; Based on the deformation prediction result, grade the road surface deformation risk. According to the comprehensive score of three dimensions: displacement, deformation rate, and prediction trend, the risk level is divided into safe level, attention level, warning level, danger level, and emergency level to form the road surface deformation risk level assessment result.

4. The deformation monitoring and control method for road pavement according to claim 1, characterized in that, According to the road surface deformation risk level assessment result, determine the deformation dominant factors and generate a hierarchical control strategy, including: Construct a multi-factor road surface deformation cause analysis model, including an environmental factor sub-model, a material factor sub-model, and a load factor sub-model. Analyze the relationship between environmental parameters and deformation through multiple linear regression to obtain the contribution degree of environmental factors; Evaluate the contribution degree of environmental factors, material aging degree, and traffic load impact through a fuzzy inference system, calculate the functional relationship between the cumulative equivalent axle load times and the deformation development rate to obtain the action intensity of each factor; Input the action intensity of each factor into the Bayesian network for integrated analysis, and output the contribution percentage of the dominant deformation cause to obtain the deformation dominant factors; Retrieve matching items from a knowledge base containing predefined control schemes according to the deformation dominant factors and the evaluation results of the road surface deformation risk levels, generate corresponding control types for road sections with different risk levels, and obtain an initial control scheme set; Conduct a life cycle cost analysis on the initial control scheme set, calculate the long-term cost-benefit ratio of each scheme, and screen out the scheme with the optimal cost-benefit ratio to obtain a candidate control strategy; Globally optimize the candidate control strategy through an integer programming algorithm, balance the allocation of resources across the entire road network on the premise of meeting safety requirements, and generate a hierarchical control strategy including an implementation schedule, a resource demand list, expected effects, and an emergency plan corresponding to different risk levels.

5. The deformation monitoring and control method for road pavement according to claim 1, characterized in that, Convert the hierarchical control strategy into a standardized execution instruction set, determine the priorities of the instructions in the execution instruction set. When the instruction priority is 1, activate the emergency response mechanism and send an emergency intervention notice within 15 minutes; when the instruction priority is 2 to 4, send an execution notice 24 hours in advance, collect execution feedback and calculate the execution deviation rate, including: Convert the hierarchical control strategy into a standardized execution instruction set including six elements: the execution entity, execution time, execution location, execution content, execution standard, and acceptance requirements. The execution entity is divided into two categories: an automatic execution system and manual intervention; Assign priorities to the standardized execution instruction set according to the risk levels, set the priority of emergency-level instructions to 1, the priority of dangerous-level instructions to 2, the priority of warning-level instructions to 3, and the priority of attention-level instructions to 4 to obtain the priority ranking result; Judge the instruction type according to the priority ranking result. When the instruction priority is 1, send an emergency intervention notice including the intervention location, intervention content, and intervention standard to the relevant responsible unit, and activate the emergency resource scheduling system to obtain an emergency response record; Judge the instruction type according to the priority ranking result. When the instruction priority is 2 to 4, send an execution notice including the execution plan, resource requirements, and acceptance criteria to the relevant responsible unit according to the scheduled time plan to obtain a regular execution record; Collect the execution status data of the emergency response record and the regular execution record through multiple channels, including automatic execution status data and manual intervention progress data, and compare and analyze the execution progress with the planned progress to obtain the execution feedback; Calculate the ratio of the actual execution time to the planned execution time based on the execution feedback to obtain the execution deviation rate, and send an abnormal alarm and adjustment suggestions to the management personnel when the execution deviation rate exceeds the preset threshold.

6. A deformation monitoring and control system for road surfaces, characterized in that, For implementing the deformation monitoring and control method for road surfaces as described in any one of claims 1-5, the deformation monitoring and control system for road surfaces includes: An acquisition module for collecting multi-level sensor data of the road surface to obtain an original deformation data set including fixed monitoring point data, mobile monitoring unit data, and road surface images; A processing module for preprocessing the original deformation data set to obtain a time series data set of the road surface state; A prediction module for inputting the time series data set of the road surface state into a long short-term memory network for risk prediction to obtain the evaluation result of the road surface deformation risk level; A grading module for determining the dominant deformation factor and generating a grading control strategy according to the evaluation result of the road surface deformation risk level; A judgment module for converting the grading control strategy into a standardized execution instruction set, judging the priorities of the instructions in the execution instruction set, and when the instruction priority is 1, starting an emergency response mechanism and sending an emergency intervention notice within 15 minutes; when the instruction priority is 2 to 4, sending an execution notice 24 hours in advance, collecting execution feedback and calculating the execution deviation rate; A regulation module for constructing a road surface health knowledge graph based on the execution feedback and the execution deviation rate, modeling the road surface deformation control process as a Markov decision process, and training a deep Q network to achieve adaptive regulation of traffic flow.

7. A deformation monitoring and control device for road pavement, characterized in that, It includes a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the computer program, it implements the deformation monitoring and control method for road surfaces according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, the processor is caused to execute the deformation monitoring and control method for road surfaces according to any one of claims 1 to 5.

Images