Real-time compactness monitoring system for highway filling construction

Abstract

The invention relates to the technical field of compactness monitoring, in particular to a real-time compactness monitoring system for highway filling construction, which comprises a vibration and pressure data synchronization module for activating a vibration frequency sensor and a pressure sensor equipped on an unmanned aerial vehicle, synchronously collecting vibration and pressure data of a construction site, carrying out real-time data calibration, and sending the data to a data processing module; and keeping time consistency, and unifying the format of data collected by the unmanned aerial vehicle to generate a synchronous data set. According to the method, the change trend in the data is identified through the self-organizing mapping algorithm, the deep learning compactness analysis module can effectively conclude and predict the change mode of the compactness, the key change trend is quickly captured from a large amount of dynamic data, and the accuracy and efficiency of data analysis are remarkably improved; a multi-layer perceptron network is adopted to analyze the deviation with the current construction standard, the construction parameters are adjusted, the monitoring and adjusting process is optimized, the deviation in construction is corrected, the stability of the road is improved, and the service life of the road is prolonged.

Description

Technical Field

[0001] The invention relates to the technical field of density monitoring, and in particular to a real-time density monitoring system for highway filling construction. Background Art

[0002] The field of density monitoring technology focuses on measuring and monitoring the compression of construction materials such as soil, asphalt or concrete during the construction process, which is a key technical link to ensure the quality of construction and the safety of subsequent use. It involves the use of various sensors and monitoring equipment to collect data on the density of materials in real time, and ensure that the materials meet the engineering specifications through data analysis.

[0003] The main purpose of the real-time density monitoring system for highway filling construction is to ensure the construction quality by real-time monitoring of the density of filling materials, thereby improving the stability and life of the road, and providing instant feedback of density data to help engineers and construction personnel make adjustments to ensure that each part of the construction meets safety standards and design specifications, ultimately achieving the effect of extending the service life of the road and reducing subsequent maintenance costs.

[0004] Traditional compaction monitoring systems rely on intermittent manual detection and basic real-time monitoring technology, lacking complex data analysis and real-time response capabilities, resulting in significant lags in monitoring data processing and response, and being unable to adjust construction parameters in time to adapt to real-time changes in conditions, affecting construction quality, resulting in road areas that are not compacted according to design standards, increasing maintenance costs and causing safety issues. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a real-time monitoring system for density of highway filling construction.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: A real-time monitoring system for density of highway filling construction comprises:

[0007] Vibration and pressure data synchronization module: activates the vibration frequency sensor and pressure sensor equipped on the drone, synchronously collects vibration and pressure data at the construction site, performs real-time data calibration, maintains time consistency, and unifies the data format collected by the drone to generate a synchronized data set;

[0008] Real-time monitoring and analysis module: Based on the synchronized data set, monitor the continuous data stream, identify and record abnormal fluctuations in the data, extract key data suitable for further analysis through data screening logic, and obtain the screened data set;

[0009] Deep learning density analysis module: based on the screened data set, convert the data into an analyzable format, use a self-organizing mapping algorithm to identify the change trend in the data, summarize the density change pattern, and convert the result into a prediction indicator to generate a density analysis result;

[0010] Real-time adjustment module for construction parameters: Based on the density analysis results, a multi-layer perceptron network is used to analyze the deviation from the current construction standard, determine whether the construction parameters need to be adjusted, adjust the compaction strength and speed through control logic, and generate an adjustment strategy plan;

[0011] Density change monitoring module: using the data after the execution of the adjustment strategy scheme to conduct continuous monitoring, compare the consistency of the density data with the preset standard, evaluate the density change trend, and generate a density monitoring report based on the comparison results;

[0012] Density response optimization module: through the density monitoring report, analyze the changes in density during the monitoring period, adjust the parameters and response strategy of the prediction model, and obtain a parameter optimization plan based on the adjustment results.

[0013] As a further solution of the present invention, the vibration and pressure data synchronization module includes a data acquisition submodule, a data calibration submodule and a data unification submodule, wherein:

[0014] Data acquisition submodule: The vibration frequency sensor and pressure sensor equipped by the drone are used to collect vibration and pressure data at the construction site, and synchronize the data to obtain the original synchronized data set;

[0015] Data calibration submodule: through the original synchronized data set, correct the data to maintain the consistency of the time line, synchronize the time mark, and obtain the calibrated data set;

[0016] Data unification submodule: Based on the calibrated data set, the data collected by the drone is processed in a unified format to ensure that all data follow the same format standard and generate a unified format data set.

[0017] As a further solution of the present invention, the real-time monitoring and analysis module includes a data monitoring submodule, an abnormality identification submodule and a key data extraction submodule, wherein:

[0018] Data monitoring submodule: based on the unified format data set, monitor the continuity and integrity of the data stream, maintain the real-time update of the data stream, and obtain a continuously monitored data stream;

[0019] Abnormal identification submodule: using the continuous monitoring data stream to identify unexpected fluctuations, record them, and obtain abnormal fluctuation records;

[0020] Key data extraction submodule: through the abnormal fluctuation records, key data points are screened out, and processed and analyzed to obtain key data sets.

[0021] As a further solution of the present invention, the deep learning density analysis module includes a data format conversion submodule, a trend identification submodule and a prediction index generation submodule, wherein:

[0022] Data format conversion submodule: based on the screened data set, eliminate noise and outliers through data cleaning, unify the data scale and format, perform structured processing on the data, improve the data availability, and generate a formatted data set;

[0023] Trend identification submodule: based on the formatted data set, using the self-organizing map algorithm, drawing a trend graph and performing regression analysis, extracting the main change trends and periodic patterns in the data, and obtaining a trend analysis report;

[0024] Prediction index generation submodule: Based on the trend analysis report, a prediction model is constructed through a modeling tool, model training and verification are performed, prediction criteria and indicators are established, and density prediction indicators are generated.

[0025] As a further solution of the present invention, the self-organizing map algorithm is according to the formula:

[0026] W(v+1)=W(v)+θ(u,v)·α·β·(D(t)-W(v))+γ·grad(E(W))

[0027] Where: W(v) is the weight vector of the iteration step, W′(v+1) is the weight vector of the next iteration step, θ(u, v) is the neighborhood function, α is the learning rate, β is the dynamic adjustment factor, D(t) is the input vector at time point t, γ is the historical influence coefficient, and H(t, τ) is the weighted average of historical data based on the time window τ.

[0028] As a further solution of the present invention, the construction parameter real-time adjustment module includes a deviation analysis submodule, a parameter decision submodule and a strategy implementation submodule, wherein:

[0029] Deviation analysis submodule: Based on the density prediction index, the deviation between the current construction parameters and the standard parameters is evaluated through comparative analysis tools, key deviation points are identified, and deviation analysis results are generated;

[0030] Parameter decision submodule: using the deviation analysis results, examining the impact of the deviation through the decision support system, formulating a reasonable parameter adjustment plan, establishing a new construction parameter standard, and generating an optimized adjustment plan;

[0031] Strategy implementation submodule: According to the optimization adjustment plan, a multi-layer perceptron network is used to adjust the compaction strength and running speed of the compactor through the control system, implement construction parameter adjustment, and obtain an adjustment strategy plan.

[0032] As a further solution of the present invention, the multi-layer perceptron network is according to the formula:

[0033] output=f(W′·[h, v, s, t, c] + b′)

[0034] Where: f is the activation function, W′ is the weight matrix, b′ is the bias term, h represents the ground hardness, v is the current speed of the compactor, s is the expected compaction standard, t is the ambient temperature, and c is the weight of the compactor.

[0035] As a further solution of the present invention, the density change monitoring module includes a data comparison submodule, a trend evaluation submodule and a report generation submodule, wherein:

[0036] Data comparison submodule: Based on the density data obtained after executing the adjustment strategy, continuous comparison is performed through time series analysis to identify deviations and perform data consistency verification to generate data consistency results;

[0037] Trend evaluation submodule: using the data consistency results, draw trend lines through statistical analysis tools, analyze trend fluctuations, and obtain trend change analysis results;

[0038] Report generation submodule: through the trend change analysis results, integrate the analysis data, compile a structured monitoring report, and obtain a density monitoring report.

[0039] As a further solution of the present invention, the density response optimization module includes a parameter adjustment submodule, a strategy optimization submodule and a configuration generation submodule, wherein:

[0040] Parameter adjustment submodule: Based on the density monitoring report, a statistical method is used to analyze the influence of response parameters, fine-tune parameters, and generate parameter optimization results;

[0041] Strategy optimization submodule: using the parameter optimization results, performing strategy simulation through experimental design methods, calibrating the strategy to improve model efficiency, and obtaining strategy optimization results;

[0042] Configuration generation submodule: Through the optimization results of the strategy, the optimization data is comprehensively optimized to build the final parameter configuration file, which is completely consistent with the final result of the module, and obtain the parameter optimization plan.

[0043] Compared with the prior art, the advantages and positive effects of the present invention are:

[0044] In the present invention, the self-organizing mapping algorithm is used to identify the changing trends in the data, and the deep learning density analysis module can effectively summarize and predict the changing patterns of density, quickly capture key changing trends from a large amount of dynamic data, and significantly improve the accuracy and efficiency of data analysis. A multi-layer perceptron network is used to analyze deviations from current construction standards and adjust construction parameters to ensure the uniformity and stability of density during the construction process, optimize the monitoring and adjustment process, correct deviations in construction, and improve the stability and life of the road. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a system flow chart of the present invention;

[0046] Figure 2 It is a schematic diagram of the system framework of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0048] Embodiment 1

[0049] See also Figure 1 The present invention provides a technical solution: a real-time monitoring system for density of highway filling construction includes:

[0050] Vibration and pressure data synchronization module: activates the vibration frequency sensor and pressure sensor equipped on the drone, synchronously collects vibration and pressure data at the construction site, performs real-time data calibration, maintains time consistency, and unifies the data format collected by the drone to generate a synchronized data set;

[0051] Real-time monitoring and analysis module: Based on the synchronous data set, it monitors the continuous data stream, identifies and records abnormal fluctuations in the data, and extracts key data suitable for further analysis through data screening logic to obtain the screened data set;

[0052] Deep learning density analysis module: Based on the screened data set, the module converts the data into an analyzable format, uses the self-organizing mapping algorithm to identify the changing trends in the data, summarizes the density change patterns, and converts the results into predictive indicators to generate density analysis results;

[0053] Real-time adjustment module for construction parameters: Based on the density analysis results, a multi-layer perceptron network is used to analyze the deviation from the current construction standard, determine whether the construction parameters need to be adjusted, adjust the compaction intensity and speed through control logic, and generate an adjustment strategy plan;

[0054] Density change monitoring module: Use the data after the adjustment strategy is executed to conduct continuous monitoring, compare the consistency of density data with the preset standard, evaluate the density change trend, and generate a density monitoring report based on the comparison results;

[0055] Density response optimization module: Through the density monitoring report, analyze the changes in density during the monitoring period, adjust the parameters and response strategy of the prediction model, and obtain the parameter optimization plan based on the adjustment results.

[0056] See also Figure 2 The vibration and pressure data synchronization module includes a data acquisition submodule, a data calibration submodule and a data unification submodule, wherein:

[0057] Data acquisition submodule: The vibration frequency sensor and pressure sensor equipped by the drone are used to collect vibration and pressure data at the construction site, and synchronize the data to obtain the original synchronized data set;

[0058] Data calibration submodule: uses the original synchronized data set to calibrate the data to maintain timeline consistency, synchronizes time stamps, and obtains the calibrated data set;

[0059] Data unification submodule: Based on the calibrated data set, the data collected by the drone is processed in a unified format to ensure that all data follow the same format standard and generate a unified format data set;

[0060] In the data acquisition submodule, based on the conditions at the construction site, the vibration frequency sensor and pressure sensor configured on the drone are used, with specific sampling frequency and sensitivity settings to ensure data accuracy, to conduct real-time monitoring of vibration and pressure. The built-in synchronization trigger mechanism of the sensor is used to ensure the time alignment of data acquisition, process the signal to reduce noise interference, and generate the original synchronized data set.

[0061] In the data calibration submodule, the original synchronized data set is used to perform data time calibration, and the timestamp is corrected using a linear interpolation algorithm to ensure the consistency of all data points on the time axis. The time offset is calculated and time correction is performed on each data point. The synchronization of the data time stamp is optimized through the correction algorithm to obtain the calibrated data set.

[0062] In the data unification submodule, based on the calibrated data set, data format unification processing is performed, all data are converted into a unified format using a data encapsulation protocol, the data encoding method and structure are defined to ensure the compatibility of the data in subsequent processing, and the data structure is standardized through format conversion tools to ensure that all data are processed according to the same standard and generate a unified format data set.

[0063] See also Figure 2The real-time monitoring and analysis module includes a data monitoring submodule, an anomaly identification submodule, and a key data extraction submodule, among which:

[0064] Data monitoring submodule: Based on the unified format data set, monitor the continuity and integrity of the data flow, maintain the real-time update of the data flow, and obtain continuous monitoring of the data flow;

[0065] Abnormal identification submodule: It continuously monitors data streams, identifies unexpected fluctuations, records them, and obtains abnormal fluctuation records;

[0066] Key data extraction submodule: Screen out key data points through abnormal fluctuation records, and process and analyze them to obtain key data sets;

[0067] In the data monitoring submodule, based on the unified format data set, the stream data management system is used to monitor the continuity and integrity of the data flow in real time. The data is segmented through the time window technology to verify the continuity of the data in each time period. At the same time, the fault detection algorithm is used to continuously evaluate the data integrity, ensure the real-time update of the data flow, and generate a continuous monitoring data flow.

[0068] In the anomaly identification submodule, the data stream is continuously monitored and an anomaly detection algorithm is adopted, specifically a statistical threshold determination method, which sets a dynamically adjusted threshold to identify unexpected fluctuations in the data, analyzes the deviation of each data point from the predetermined threshold, records all anomalies that exceed the threshold, and timestamps them to obtain abnormal fluctuation records;

[0069] In the key data extraction submodule, data screening technology is used through abnormal fluctuation records, specifically a decision tree-based classification method, to evaluate the impact of each abnormal point according to the preset data importance standard, screen out key data points, and perform further analysis and processing, including cluster analysis of data points to determine data patterns, and obtain key data sets.

[0070] See also Figure 2 ,The deep learning density analysis module includes a data format conversion submodule, a trend identification submodule and a prediction indicator generation submodule, among which:

[0071] Data format conversion submodule: Based on the filtered data set, it eliminates noise and outliers through data cleaning, unifies the data scale and format, structures the data, improves the data usability, and generates a formatted data set;

[0072] Trend identification submodule: Based on the formatted data set, the self-organizing mapping algorithm is used to draw trend graphs and perform regression analysis to extract the main change trends and periodic patterns in the data and obtain trend analysis reports;

[0073] Prediction index generation submodule: Based on the trend analysis report, the prediction model is constructed through modeling tools, the model is trained and verified, prediction criteria and indicators are established, and density prediction indicators are generated;

[0074] In the data format conversion submodule, based on the filtered data set, data cleaning technology is used to first perform noise and outlier elimination operations, detect and delete outliers in the data through automated scripts, including setting anomaly detection thresholds to identify and remove numerical anomalies, and then unify the data scale and format, use data conversion tools to normalize various data to ensure that all data values ​​are on the same scale, structure the data, and use SQL database commands including INSERT and UPDATE statements to reconstruct the data to generate a formatted data set;

[0075] In the trend identification submodule, based on the formatted data set, the self-organizing map algorithm is used. The algorithm constructs a self-organizing map network, sets the learning rate and number of iterations of the network, performs network training, draws trend graphs and performs regression analysis, including using regression analysis tools to fit trend data, evaluate the relationship strength between variables, and extract the main change trends and periodic patterns in the data. The period detection function including fast Fourier transform is used to analyze the data frequency to obtain a trend analysis report;

[0076] In the prediction indicator generation submodule, a prediction model is built based on the trend analysis report through modeling tools. Specifically, a time series prediction model is used to set the time window and step parameters of the model, and model training and verification are carried out. Cross-validation technology is used to evaluate the accuracy and stability of the model, and prediction criteria and indicators are established. The training algorithms used in the prediction model include gradient descent and adaptive learning rate adjustment to improve the accuracy of the prediction. Statistical analysis software is used to analyze the results and generate density prediction indicators.

[0077] Self-organizing map algorithm, according to the formula:

[0078] W(v+1)=W(v)+θ(u,v)·α·β·(D(t)-W(v))+γ·grad(E(W))

[0079] Where: W(v) is the weight vector of the iteration step, W′(v+1) is the weight vector of the next iteration step, θ(u, v) is the neighborhood function, α is the learning rate, β is the dynamic adjustment factor, D(t) is the input vector at time point t, γ is the historical influence coefficient, and H(t, T) is the weighted average of historical data based on the time window τ;

[0080] Execution process: First, the weight vector W(v) is initialized to reflect the model's response to various environmental and operating parameters. Then, key data D(t) is collected in each measurement cycle, including soil density and vibration frequency. Then, the neighborhood function θ(u, v) is applied, and the weight adjustment range is calculated through the learning rate α and the dynamic adjustment factor β. At the same time, the weighted average H(t, T) of historical data is introduced and the historical influence coefficient γ is adjusted to ensure that the model can comprehensively consider new data and historical trends. Finally, the weight vector W′(v+1) is updated to accurately adapt to changes on the construction site and ensure construction quality and road safety.

[0081] See also Figure 2 The construction parameter real-time adjustment module includes a deviation analysis submodule, a parameter decision submodule and a strategy implementation submodule, among which:

[0082] Deviation analysis submodule: Based on the density prediction index, the deviation between the current construction parameters and the standard parameters is evaluated through comparative analysis tools, key deviation points are identified, and deviation analysis results are generated;

[0083] Parameter decision submodule: using the deviation analysis results, the decision support system examines the impact of the deviation, formulates a reasonable parameter adjustment plan, establishes a new construction parameter standard, and generates an optimized adjustment plan;

[0084] Strategy implementation submodule: According to the optimization adjustment plan, a multi-layer perceptron network is used to adjust the compaction strength and running speed of the compactor through the control system, implement construction parameter adjustments, and obtain the adjustment strategy plan;

[0085] In the deviation analysis submodule, based on the density prediction index, a comparative analysis tool is used to calculate the difference between the current construction parameters and the standard parameters by setting a statistical deviation algorithm. The key deviation points are identified by calculating the standard deviation and mean square error. The deviation of each parameter is recorded and marked through a deviation database that is updated in real time, and displayed through a data visualization tool to generate deviation analysis results.

[0086] In the parameter decision submodule, the deviation analysis results are used through the decision support system. The specific implementation includes loading the decision logic model, analyzing the severity of the deviation's impact on construction based on the rule engine and weight scoring system, sorting the importance of parameters according to the preset decision tree, formulating a reasonable parameter adjustment plan, updating the construction parameter database through automated scripts, establishing new construction parameter standards, and generating an optimized adjustment plan;

[0087] In the strategy implementation submodule, a multilayer perceptron network is used according to the optimization adjustment plan. The specific implementation includes configuring the number of network layers and neurons, setting the learning rate and activation function, receiving the adjustment parameter input, and adjusting the compaction strength and running speed of the compactor through the trained weights. At the same time, the operation log is automatically recorded, and the effect is monitored to obtain the adjustment strategy plan.

[0088] Multilayer perceptron network, according to the formula:

[0089] output=f(W′·[h, v, s, t, c] + b′)

[0090] Where: f is the activation function, W′ is the weight matrix, b′ is the bias term, h represents the ground hardness, v is the current speed of the compactor, s is the expected compaction standard, t is the ambient temperature, and c is the weight of the compactor;

[0091] Execution process: First, key parameters are collected in real time through sensors, including ground hardness h, current speed v of the compactor, expected compaction standard s, ambient temperature t, and weight c of the compactor, and transmitted to the multi-layer perceptron network as the input vector [h, v, s, t, c]. The weight matrix W′ and bias term b′ in the network have been adjusted to the best through the gradient descent method during the training phase to accurately map the input to the expected output. The input data is weighted, biased, and passed through the activation function f to output the adjustment strategy of the compactor, including the required compaction strength and speed, which is sent to the control system of the compactor to adjust the operating parameters of the machine in real time, continue to monitor the adjustment effect and collect data, form a closed-loop feedback, and continuously optimize the operation to ensure that the compaction degree meets the preset standards and adapts to changing environmental conditions.

[0092] See also Figure 2 The density change monitoring module includes a data comparison submodule, a trend evaluation submodule and a report generation submodule, among which:

[0093] Data comparison submodule: Based on the density data obtained after executing the adjustment strategy, continuous comparison is performed through time series analysis to identify deviations and perform data consistency verification to generate data consistency results;

[0094] Trend assessment submodule: using data consistency results, draw trend lines through statistical analysis tools, analyze trend fluctuations, and obtain trend change analysis results;

[0095] Report generation submodule: Through trend change analysis results, integrate analysis data, compile structured monitoring reports, and obtain density monitoring reports;

[0096] In the data comparison submodule, based on the density data obtained after executing the adjustment strategy, the time series analysis method is adopted. The specific implementation includes deploying the autoregressive integrated moving average model for continuous comparison, setting model parameters such as difference level, autoregressive term and moving average term, implementing the deviation calculation of continuous data points through automated scripts, performing time series prediction for each data point and identifying the deviation therein, judging the consistency of data through threshold logic, and thus performing data consistency verification to generate data consistency results;

[0097] In the trend assessment submodule, the data consistency results are used to draw trend lines through statistical analysis tools. The specific implementation includes using regression analysis methods, using least squares estimation to determine the best fit of the trend line, analyzing the deviation between the time points in the data set and the trend line, determining the main trend fluctuations of the data, and using periodic analysis tools such as periodograms and seasonal decomposition to reveal the periodic patterns in the data and obtain trend change analysis results;

[0098] In the report generation submodule, the trend change analysis results are used to integrate the analysis data. The specific implementation includes using document management systems and data visualization tools, configuring report templates and chart styles, importing analysis data and trend charts, compiling data points, analysis interpretations and charts into structured monitoring reports through automated processes, setting output formats and publishing parameters, ensuring the logic and format consistency of the report content, and obtaining a density monitoring report.

[0099] See also Figure 2 The density response optimization module includes a parameter adjustment submodule, a strategy optimization submodule and a configuration generation submodule, among which:

[0100] Parameter adjustment submodule: Based on the density monitoring report, statistical methods are used to analyze the impact of response parameters, perform parameter fine-tuning, and generate parameter optimization results;

[0101] Strategy optimization submodule: Use parameter optimization results to simulate strategies through experimental design methods, calibrate strategies to improve model efficiency, and obtain strategy optimization results;

[0102] Configuration generation submodule: Through the strategy optimization results and comprehensive optimization data, the final parameter configuration file is constructed, which is completely consistent with the final result of the module, and the parameter optimization plan is obtained;

[0103] In the parameter adjustment submodule, based on the compactness monitoring report, statistical analysis methods are used, including the deployment of variance analysis to analyze the impact of response parameters on compactness, and the most significant parameters are determined through sensitivity analysis. Fine-tuning is also performed at the same time. The adjustment operation includes setting a new parameter range and step size, iteratively testing new parameter configurations through automated scripts, evaluating the impact of parameter adjustments on model output, optimizing overall construction efficiency, and generating parameter optimization results.

[0104] In the strategy optimization submodule, the parameter optimization results are used to set different levels of control factors and noise factors through experimental design methods. Multiple rounds of simulation experiments are run through statistical software to analyze the impact of each factor on the strategy results. The strategy parameters are adjusted according to the experimental results to improve the prediction efficiency and accuracy of the model and obtain the strategy optimization results.

[0105] In the configuration generation submodule, the strategy optimization results are used to comprehensively optimize the data. The specific implementation includes using the configuration management tool to build a parameter configuration file, loading the optimized parameter data, and automatically generating the configuration file through a script to ensure that the parameters in the configuration file are completely consistent with the strategy optimization results, verify the data consistency and integrity of the configuration file, ensure that the configuration meets the requirements of the construction module, and obtain a parameter optimization plan.

[0106] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A real-time monitoring system for density of highway filling construction, characterized by: The system comprises: Vibration and pressure data synchronization module: activates the vibration frequency sensor and pressure sensor equipped on the drone, synchronously collects vibration and pressure data at the construction site, performs real-time data calibration, maintains time consistency, and unifies the data format collected by the drone to generate a synchronized data set; Real-time monitoring and analysis module: Based on the synchronized data set, monitor the continuous data stream, identify and record abnormal fluctuations in the data, extract key data suitable for further analysis through data screening logic, and obtain the screened data set; Deep learning density analysis module: based on the screened data set, convert the data into an analyzable format, use a self-organizing map algorithm to identify the change trend in the data, summarize the density change pattern, and convert the result into a prediction indicator to generate a density analysis result; Construction parameter real-time adjustment module: Based on the density analysis results, a multi-layer perceptron network is used to analyze the deviation from the current construction standard, determine whether the construction parameters need to be adjusted, adjust the compaction strength and speed through control logic, and generate an adjustment strategy plan; Density change monitoring module: using the data after the execution of the adjustment strategy scheme to conduct continuous monitoring, compare the consistency of the density data with the preset standard, evaluate the density change trend, and generate a density monitoring report based on the comparison results; Density response optimization module: through the density monitoring report, analyze the changes in density during the monitoring period, adjust the parameters and response strategy of the prediction model, and obtain a parameter optimization plan based on the adjustment results.

2. The real-time monitoring system for compaction of highway filling construction according to claim 1 is characterized in that: The vibration and pressure data synchronization module includes a data acquisition submodule, a data calibration submodule and a data unification submodule, wherein: Data acquisition submodule: The vibration frequency sensor and pressure sensor equipped by the drone are used to collect vibration and pressure data at the construction site, and synchronize the data to obtain the original synchronized data set; Data calibration submodule: through the original synchronized data set, correct the data to maintain the consistency of the time line, synchronize the time mark, and obtain the calibrated data set; Data unification submodule: Based on the calibrated data set, the data collected by the drone is processed in a unified format to ensure that all data follow the same format standard and generate a unified format data set.

3. The real-time monitoring system for compaction of highway filling construction according to claim 1 is characterized in that: The real-time monitoring and analysis module includes a data monitoring submodule, an abnormality identification submodule and a key data extraction submodule, wherein: Data monitoring submodule: based on the unified format data set, monitor the continuity and integrity of the data stream, maintain the real-time update of the data stream, and obtain a continuously monitored data stream; Abnormal identification submodule: using the continuous monitoring data stream to identify unexpected fluctuations, record them, and obtain abnormal fluctuation records; Key data extraction submodule: through the abnormal fluctuation records, key data points are screened out, and processed and analyzed to obtain key data sets.

4. The real-time monitoring system for compaction of highway filling construction according to claim 1 is characterized in that: The deep learning density analysis module includes a data format conversion submodule, a trend identification submodule and a prediction index generation submodule, wherein: Data format conversion submodule: based on the screened data set, eliminate noise and outliers through data cleaning, unify the data scale and format, perform structured processing on the data, improve the data availability, and generate a formatted data set; Trend identification submodule: based on the formatted data set, using the self-organizing map algorithm, drawing a trend graph and performing regression analysis, extracting the main change trends and periodic patterns in the data, and obtaining a trend analysis report; Prediction index generation submodule: Based on the trend analysis report, a prediction model is constructed through a modeling tool, model training and verification are performed, prediction criteria and indicators are established, and density prediction indicators are generated.

5. The real-time monitoring system for compaction of highway filling construction according to claim 1 is characterized in that: The self-organizing map algorithm is based on the formula: W(v+1)=W(v)+θ(u,v)·α·β·(D(t)-W(v))+γ·grad(E(W)) Where: W(v) is the weight vector of the iteration step, W′(v+1) is the weight vector of the next iteration step, θ(u, v) is the neighborhood function, α is the learning rate, β is the dynamic adjustment factor, D(t) is the input vector at time point t, γ is the historical influence coefficient, and H(t, τ) is the weighted average of historical data based on the time window τ.

6. The real-time monitoring system for compaction of highway filling construction according to claim 1 is characterized in that: The construction parameter real-time adjustment module includes a deviation analysis submodule, a parameter decision submodule and a strategy implementation submodule, wherein: Deviation analysis submodule: Based on the density prediction index, the deviation between the current construction parameters and the standard parameters is evaluated through comparative analysis tools, key deviation points are identified, and deviation analysis results are generated; Parameter decision submodule: using the deviation analysis results, examining the impact of the deviation through the decision support system, formulating a reasonable parameter adjustment plan, establishing a new construction parameter standard, and generating an optimized adjustment plan; Strategy implementation submodule: According to the optimization adjustment plan, a multi-layer perceptron network is used to adjust the compaction strength and running speed of the compactor through the control system, implement construction parameter adjustment, and obtain an adjustment strategy plan.

7. The real-time monitoring system for compaction of highway filling construction according to claim 1 is characterized in that: The multi-layer perceptron network is based on the formula: o utp ut=f(W′·[h, v, s, t, c] + b′) Where: f is the activation function, W′ is the weight matrix, b′ is the bias term, h represents the ground hardness, v is the current speed of the compactor, s is the expected compaction standard, t is the ambient temperature, and c is the weight of the compactor.

8. The real-time monitoring system for compaction of highway filling construction according to claim 1 is characterized in that: The density change monitoring module includes a data comparison submodule, a trend evaluation submodule and a report generation submodule, wherein: Data comparison submodule: Based on the density data obtained after executing the adjustment strategy, continuous comparison is performed through time series analysis to identify deviations and perform data consistency verification to generate data consistency results; Trend evaluation submodule: using the data consistency results, draw trend lines through statistical analysis tools, analyze trend fluctuations, and obtain trend change analysis results; Report generation submodule: through the trend change analysis results, integrate the analysis data, compile a structured monitoring report, and obtain a density monitoring report.

9. The real-time monitoring system for compaction of highway filling construction according to claim 1, characterized in that: The density response optimization module includes a parameter adjustment submodule, a strategy optimization submodule and a configuration generation submodule, wherein: Parameter adjustment submodule: Based on the density monitoring report, a statistical method is used to analyze the influence of response parameters, fine-tune parameters, and generate parameter optimization results; Strategy optimization submodule: using the parameter optimization results, performing strategy simulation through experimental design methods, calibrating the strategy to improve model efficiency, and obtaining strategy optimization results; Configuration generation submodule: Through the optimization results of the strategy, the optimization data is comprehensively optimized to build the final parameter configuration file, which is completely consistent with the final result of the module, and obtain the parameter optimization plan.

Images