A tamping car network control method and system based on multi-source data fusion

Abstract

The present application belongs to the technical field of tamping car control, and particularly relates to a tamping car network control method and system based on multi-source data fusion. The method acquires multi-source data in the tamping car operation process, classifies the multi-source data, and constructs corresponding regulation and control data sets. The corresponding relationship between adjacent regulation and control data sets is identified, the verification object and the correction object are extracted, the matching degree is calculated, and the consistency score is output. According to the consistency score, the multi-source data is classified into a first type of regulation and control task or a second type of regulation and control task. When the regulation and control task is the first type of regulation and control task, the regulation and control benchmark for regulation and control requirements is determined, and the rationality of the regulation and control benchmark is evaluated to select the best regulation and control scheme. Control instructions are generated based on the best regulation and control scheme to guide the tamping car to perform the operation. The present application improves the operation accuracy of the tamping car and the adaptability to complex track working conditions.

Description

Technical Field

[0001] This invention belongs to the field of tamping machine control technology, specifically relating to a tamping machine network control method and system based on multi-source data fusion. Background Technology

[0002] As an efficient and high-capacity mode of transportation, railway transportation ensures safe and stable operation. In the daily maintenance of railway lines, tamping machines are the core equipment in railway line maintenance operations. Their performance and the accuracy of the control system are directly related to the smooth operation of the line and the overall transportation safety. Therefore, improving the intelligent automation level of tamping machines is a goal that the railway maintenance field is constantly pursuing.

[0003] Existing tamping machine operation monitoring systems rely on a single data source, such as collecting only local parameters like vibration, pressure, or displacement. This data acquisition mode struggles to comprehensively and in real-time acquire all key parameters involved in the tamping operation, such as the overall geometric state (rail elevation, track alignment, gauge, superelevation, torsion, etc.) and operational parameters (tamping depth, tamping torque, vibration frequency, etc.). When facing complex and variable track environments or unexpected situations, such as curves, slopes, and turnout areas, the lack of multi-dimensional and timely data support makes it impossible to accurately assess and predict the operational status, leading to delayed responses and even misjudgments or omissions, thus affecting maintenance operations. Furthermore, when generating and executing control schemes, the existing system only considers the parameters of the current work point in isolation, ignoring the data correlation, consistency of work content, and mutual influence between adjacent control work points. This may lead to uncoordinated parameter adjustments between different work points during continuous operation, resulting in varying degrees of deviation of some key parameters in actual execution; for example, uneven tamping depth or track lifting amount, which in turn affects the overall smoothness and stability of the line, reducing the reliability of the control system and the final quality of maintenance work.

[0004] In view of this, the present invention proposes a network control method and system for tamping machines based on multi-source data fusion. Summary of the Invention

[0005] The purpose of this invention is to provide a network control method for tamping machines based on multi-source data fusion, so as to solve the problems of inconsistent control operations and execution deviations in the existing technology.

[0006] This invention adopts the following technical solution: a network control method for tamping machines based on multi-source data fusion, wherein when multi-source data during the tamping machine operation is acquired, and the processed multi-source data indicates a need for control, the following is executed:

[0007] Determine the control benchmark for controlling demand; evaluate the rationality of the control benchmark to select the optimal control scheme; and generate control instructions based on the optimal control scheme to guide the tamping machine in performing operations.

[0008] The process of processing multi-source data to indicate the need for regulation includes: acquiring multi-source data, which includes track state parameters before operation and regulation parameters during the operation of the tamping machine; classifying the multi-source data, determining the data categories, and constructing the corresponding regulation dataset.

[0009] Track condition parameters include irregularity data.

[0010] Preferably, classifying multi-source data, determining data categories, and constructing corresponding regulatory datasets includes: identifying the correspondence between adjacent regulatory datasets based on multi-source data, and extracting verification objects and correction objects;

[0011] Calculate the matching degree between the verification object and the correction object, and output the consistency score; and classify the multi-source data into the first type of control task or the second type of control task based on the consistency score.

[0012] Preferably, constructing the corresponding control dataset includes: using the uneven data obtained before the operation as the comparison data and the remaining data as the reference data; outputting the comparison results through a preset simulation calculation method, and then performing a ratio calculation with the data source where the comparison data is located to obtain the control deviation;

[0013] After correcting the control parameters and control deviations during operation, the control correction value is obtained, and the control correction value and the parameters before correction are respectively merged into the same control dataset.

[0014] Preferably, identifying the correspondence between adjacent control datasets based on multi-source data includes: setting the data to be compared in two adjacent control datasets as a verification group and a correction group, respectively, and setting the verification value and correction direction to generate a correction object;

[0015] And the calibration object and the verification object are used as training inputs to be processed.

[0016] Preferably, generating the correction object includes: calculating the difference between the correction object and the verification object, and using the result as the correction ratio; and determining whether the verification object is higher or lower than the correction object based on the sign of the ratio, thereby determining the correction direction.

[0017] Preferably, calculating the matching degree between the verification object and the correction object and outputting the consistency score includes: forming the content to be evaluated by adjusting the values ​​of the correction object and the verification object based on the proportional difference;

[0018] The evaluation content is simulated to obtain a conversion value; the conversion value is then calculated with the preset evaluation criteria to obtain a score result, and the score result is output as a consistency score.

[0019] Preferably, classifying multi-source data into a first type of control task or a second type of control task based on consistency scores includes: determining whether the consistency score meets a preset similarity threshold; if it does, classifying the multi-source data into a first type of control task; if it does not, classifying the multi-source data into a second type of control task.

[0020] Preferably, determining the control benchmark for control demand includes, when the control task is a first type of control task: obtaining the orbital state information of the current operation to be performed, and extracting the effective fluctuation area therein;

[0021] Within the effective fluctuation area, the sampling path is determined, and the geometric relationship between each sampling position is measured to calculate the effective parameter value of each point; based on the effective parameter value, the reference point in the current track whose change amplitude exceeds the preset threshold is extracted;

[0022] The reference point is used as a search condition to find relevant parameters in the list of control parameters associated with the first type of control task, and the results are used as the initial control benchmark for the current operation.

[0023] Preferably, assessing the rationality of the control benchmark to select the best control scheme includes: based on the preliminary control benchmark, combining the unevenness score results in the historical operation records to perform a comprehensive calculation to generate the objects to be inspected;

[0024] The control simulation is re-performed based on the object to be inspected to obtain optimized control parameters; and the consistency evaluation results of the optimized control parameters are calculated, and the evaluation result with the best value and its corresponding control parameters are selected as the final parameters.

[0025] A tamping machine network control system based on multi-source data fusion includes: a data acquisition and classification module, used to acquire multi-source data during the tamping machine operation, classify the multi-source data, and construct a control dataset;

[0026] The regulation task identification module is used to respond to the output of the data acquisition and classification module, identify the correspondence between adjacent regulation datasets, calculate the matching degree to output the consistency score, and classify the multi-source data into the first type of regulation task or the second type of regulation task based on the consistency score.

[0027] The control benchmark determination module is used to respond to the classification results of the control task identification module. When the classification result is a first-class control task, it determines the preliminary control benchmark for the first-class control task.

[0028] The scheme optimization and evaluation module is used to respond to the initial control benchmark of the control benchmark determination module, combine it with historical operation records for comprehensive calculation, obtain the optimized control parameters through control simulation, and calculate the consistency evaluation results of the optimized control parameters in order to select the best control scheme.

[0029] The instruction generation and output module is used to respond to the optimal control scheme of the scheme optimization and evaluation module, and generate control instructions to guide the tamping machine to perform operations.

[0030] Beneficial effects

[0031] 1. This invention acquires multi-source data during the tamping machine operation, classifies the multi-source data using the consistency score between adjacent control datasets, and determines the first type or second type of control task; based on the classification results, the system selects the best control scheme to generate control commands; and then adaptively adjusts the tamping machine control strategy according to the reliability and similarity of the data, thereby improving the accuracy of tamping machine operation and its adaptability to complex track conditions.

[0032] 2. When determining the control benchmark, this invention acquires track state information, extracts the effective fluctuation area and determines the reference point. In the first type of control task, relevant parameters are obtained as the control benchmark, making the determination of the control benchmark targeted and accurate, focusing on the effective fluctuation area, realizing the refined and localized control of the tamping machine on the uneven data, and improving the quality and efficiency of track operation.

[0033] 3. This invention is based on the control benchmark and combines the irregularity score results in the historical operation record for comprehensive calculation. The control simulation is re-performed to obtain optimized control parameters, and the consistency evaluation results are calculated. The comprehensive evaluation and optimization process ensures that the control parameters are reasonable and reliable, thereby reducing the operation risk caused by improper parameters and extending the service life of the track. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the control and classification method of the present invention;

[0035] Figure 2 This is a schematic diagram of the method for determining and controlling tamping commands according to the present invention;

[0036] Figure 3 This is a system structure diagram of the present invention. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] Example 1

[0039] like Figure 1 and 2 As shown, this embodiment provides a network control method for tamping machines based on multi-source data fusion. By performing refined processing and logical comparison of multi-source data during the operation, adaptive optimization of tamping operation parameters is achieved. The specific implementation steps are as follows:

[0040] S1. Acquire data and classify it for processing:

[0041] Multi-source data is collected during the operation of the tamping machine by on-board sensors, including track state parameters before operation, i.e. deviation data of track geometric parameters from design values, such as elevation, level, and track orientation deviations; and control parameters during the operation of the tamping machine, such as tamping depth, clamping pressure, and tamping frequency.

[0042] According to preset rules, multi-source data are classified to determine data categories and construct corresponding control datasets, which include a set of original control parameters, track state parameters, and control parameters after correction, to support adaptive optimization of tamping operation parameters; in the process of constructing control datasets, the irregularity data in the track state obtained before the operation is used as the comparison data, and the remaining data is used as reference data.

[0043] The comparison results are output through a preset numerical calculation program, and then the ratio is calculated with the data source where the comparison data is located to obtain the control deviation, which is used to quantify the degree to which the track state parameters deviate from the design value. After the control parameters and control deviation during operation are corrected, the control correction value is obtained, and the control correction value and the parameters before correction are respectively merged into the same control dataset, thereby realizing the fine preprocessing of the original data.

[0044] S2. Extract training parameters and output consistency score:

[0045] Using the classified data, the correspondence between adjacent control datasets is identified, and verification objects and their associated correction objects are extracted. Verification objects are reference or benchmark numerical sequences or datasets, while correction objects are numerical sequences or datasets that need adjustment or evaluation after setting verification groups, correction groups, verification values, and correction directions. Verification and correction objects are input as numerical sequences to be processed into a pre-defined logic processing unit for processing. The matching degree is calculated and used as a similarity index between adjacent control datasets, outputting a consistency score to measure the similarity between adjacent control datasets. During the identification of adjacent control datasets, the data to be compared in two adjacent control datasets are set as verification and correction groups, respectively, and verification values ​​and correction directions are set to generate correction objects.

[0046] Then, the calibration object and the verification object are used as numerical inputs to be processed. The difference between the calibration object and the verification object is calculated, and the result is used as the calibration ratio to determine the relative height relationship between the verification object and the calibration object. The sign of the ratio determines whether the verification object is higher or lower than the calibration object, and then the calibration direction is determined.

[0047] In determining the consistency score, the values ​​of the correction object and the verification object are adjusted based on the ratio difference to form the content to be evaluated. The content to be evaluated contains the information to be input in the current operation state. The input content is simulated to obtain the conversion value. The conversion value is then numerically calculated with the preset evaluation criteria to obtain the scoring result, and the scoring result is output as the consistency score.

[0048] S3. Analyze consistency characteristics and determine the classification results:

[0049] Determine whether the consistency score meets the preset similarity threshold, and compare the consistency score with the preset similarity threshold. If they match, the input data is considered as the result of the first type of regulation task, indicating that the regulation has high reference value. If it is below the similarity threshold, it is considered as the result of the second type of regulation task, indicating that the regulation data has large differences and needs to be archived.

[0050] S4. Determine appropriate parameters for regulation:

[0051] Based on the first type of control task, the control benchmark is determined from the list of associated control parameters. After obtaining the track status information of the current operation to be performed from the third-party equipment, the effective fluctuation area is extracted, that is, the range in which the track geometric parameters deviate from the design value. Multiple sampling positions are determined within the effective fluctuation area, and the sampling positions are aggregated into a sampling path to measure the geometric relationship between each sampling position.

[0052] The effective parameter values ​​of each point are calculated using a preset calculation formula, reflecting the parameter values ​​of the actual track state. Reference points in the current track whose changes exceed a preset threshold are extracted based on the effective parameter values. Then, the reference points are used as search conditions to find relevant parameters in the first type of control records, and the search results are used as the initial control benchmark for the current operation.

[0053] S5. Calculate the evaluation indicators and select the best solution:

[0054] Evaluation indicators are calculated to measure the rationality of the control benchmark. The calculation comprehensively considers the preliminary control benchmark and historical track quality index. Based on the screened preliminary control benchmark, combined with the track irregularity score results (TQI) from historical operation records, a comprehensive calculation is performed to generate the object to be inspected, which is used for subsequent control simulation and evaluation. Then, a new control simulation is performed based on the object to be inspected to obtain optimized control parameters. The consistency evaluation result is then calculated based on this and considered as the final comprehensive evaluation output.

[0055] S6, Output control commands:

[0056] The system selects the optimal evaluation result and its corresponding control parameters from multiple evaluation results as the final parameters, and generates control commands based on the final parameters to guide the tamping machine in its operation. The system automatically matches and optimizes the control parameters, and outputs precise control commands based on the judgment that the scores are not lower than the historical average. The commands cover key information such as execution pressure, clamping time, and tamping path offset. The control commands are sent to the tamping equipment in real time through the field control platform, enabling the equipment to complete efficient control based on the optimal reference parameters.

[0057] The method of the present invention also includes an additional confirmation process, namely, before the instruction is issued, the preset track structure mechanical response calculation program will virtually reconstruct the target track structure according to the selected control parameters and simulate its mechanical behavior and response under tamping operation to confirm whether the tamping operation can achieve the expected recovery effect; after confirming that the tamping can achieve the expected recovery effect, the final issuance and execution will be completed.

[0058] Example 2

[0059] like Figure 3 This embodiment provides a tamping machine network control system based on multi-source data fusion for intelligent regulation and optimization of the tamping machine's operation. This system can be deployed in the tamping machine's onboard control unit or a remote control center. The modules exchange data and work collaboratively through an internal data bus or network interface. Specifically, it includes the following modules:

[0060] The data acquisition and classification module is used to acquire multi-source data during the tamping machine operation, classify the multi-source data, and construct a control dataset;

[0061] Specifically, the module collects multi-source data during the operation of the tamping machine through on-board sensors. The multi-source data includes track state parameters before operation, namely deviation data of track geometric parameters from design values, such as elevation, level, and track orientation deviations; as well as control parameters during the operation of the tamping machine, such as tamping depth, clamping pressure, and tamping frequency.

[0062] This module classifies multi-source data according to preset rules, determines data categories, and constructs corresponding control datasets. When constructing the control dataset, the irregular data obtained before the operation is used as the comparison data, and the remaining data is used as the reference data. The comparison results are output through a preset numerical calculation program, and then the ratio is calculated with the data source where the comparison data is located to obtain the control deviation. For example, assuming that the irregular data obtained before the operation is 2.0mm and the reference data is 50.0mm, the ratio calculation is: 2.0 / 50.0=0.04, that is, the control deviation is 0.04.

[0063] Then, the control parameters and control deviations during operation are corrected to obtain the control correction value. The control correction value and the parameters before correction are then merged into the same control dataset. For example, assuming the current control parameter, such as the tamping depth, is 100.0 mm, the control correction value = 100.0 * (1 + 0.04) = 104.0 mm. This module then stores 104.0 mm together with the original 100.0 mm into the control dataset.

[0064] The regulation task identification module responds to the output of the data acquisition and classification module, identifies the correspondence between adjacent regulation datasets, calculates the matching degree to output a consistency score, and classifies multi-source data into the first type of regulation task or the second type of regulation task based on the consistency score.

[0065] Specifically, this module uses the classified data to identify the correspondence between adjacent control datasets, extracts the verification object and its associated correction object, inputs the verification object and correction object as a numerical sequence to be processed into a preset logic processing unit for processing, calculates the matching degree, and uses the matching degree as a similarity index between adjacent controls to output a consistency score.

[0066] When identifying the correspondence, the data to be compared in two adjacent control datasets are set as the verification group and the correction group, respectively. The difference between the correction object and the verification object is calculated, and the result is used as the correction ratio difference. For example, assuming that in two adjacent control datasets, the verification group data is {10,20} and the correction group data is {12,22}, then the correction ratio difference is calculated as follows: Difference 1 is 12-10=+2; Difference 2 is 22-20=+2.

[0067] The sign of the difference indicates whether the verification object is higher or lower than the calibration object, thus determining the calibration direction. For example, since the difference is positive, it means that the calibration group value is higher than the verification group, and the calibration direction is determined to be "upward adjustment".

[0068] When calculating the consistency score, the values ​​of the correction object and the verification object are adjusted based on the ratio difference to form the content to be evaluated; for example, assuming the correction object is 12 and the verification object is 10, after adjusting the values, the content to be evaluated can be set to 1.2.

[0069] The module then simulates the content to be evaluated to obtain a conversion value; for example, the conversion value obtained after simulation is 0.85. Finally, the module performs numerical calculations with the conversion value and a preset evaluation standard, assumed to be 1.0, to obtain a score result, which is used as the consistency score; for example, consistency score = 0.85 / 1.0 = 0.85.

[0070] This module further determines whether the consistency score meets the preset similarity threshold. For example, assuming the similarity threshold is set to 0.80, if the consistency score, such as 0.85, is greater than or equal to the similarity threshold, the multi-source data is classified as the first type of control task; if the consistency score, such as 0.75, is lower than the similarity threshold, the multi-source data is classified as the second type of control task.

[0071] The control benchmark determination module is used to respond to the classification results of the control task identification module. When the classification result is a first-class control task, the control benchmark used for the first-class control task is determined.

[0072] Specifically, the module obtains the track status information of the current task to be performed from a third-party device, extracts the effective fluctuation area, that is, the range in which the track geometric parameters deviate from the design value, determines multiple sampling positions within the effective fluctuation area, and aggregates the sampling positions into a sampling path, measures the geometric relationship between each sampling position, and calculates the effective parameter value of each point using a preset calculation formula.

[0073] For example, assuming the effective fluctuation range is The module selects sampling locations of 105m and 115m within the area, calculates effective parameter values ​​of 0.5 and 0.6 respectively, and extracts reference points in the current track whose change amplitude exceeds the preset change amplitude threshold based on the effective parameter values; for example, if the change amplitude threshold is set to 0.55, then 0.6 at 115m exceeds the threshold and is determined as a reference point.

[0074] Finally, the module uses the reference point as the search condition to find relevant parameters in the list of control parameters associated with the first type of control task, and uses the found result as the control benchmark for the current operation; for example, if the module searches for parameters related to 115m in the list of control parameters and finds the parameter "clamping pressure 15MPa", then the control benchmark is set to 15MPa.

[0075] The scheme optimization and evaluation module is used to respond to the control benchmark of the control benchmark determination module, perform comprehensive calculations in conjunction with historical operation records, obtain optimized control parameters through control simulation, and calculate the consistency evaluation results of the optimized control parameters in order to select the best control scheme.

[0076] Specifically, this module performs comprehensive calculations based on the screened control benchmark and the track irregularity score results (TQI) of the railway track in historical operation records to generate the object to be inspected. For example, assuming the control benchmark is 15MPa and the historical TQI is 3.5, the module comprehensively calculates and generates the object to be inspected, which includes the benchmark of 15MPa and TQI of 3.5. Based on the object to be inspected, the control simulation is performed again to obtain the optimized control parameters.

[0077] For example, if the optimized control parameter obtained from the simulation is 14.8 MPa, this module calculates the consistency evaluation result of the optimized control parameter and selects the evaluation result with the best value and its corresponding control parameter as the final parameter; for example, if the calculated consistency evaluation result is 0.92, and there is another set of parameters with an evaluation result of 0.95 corresponding to 14.9 MPa, then 14.9 MPa is selected as the final parameter.

[0078] The instruction generation and output module is used to respond to the optimal control scheme of the scheme optimization and evaluation module, and generate control instructions to guide the tamping machine to perform operations;

[0079] Specifically, this module selects the optimal evaluation result and its corresponding control parameters from multiple evaluation results as the final parameters, and generates control instructions based on the final parameters to guide the tamping machine in its operation. The module automatically matches the optimal control parameters and outputs precise control instructions based on the judgment that the scores are not lower than the historical average. The control instructions cover key aspects such as execution pressure, clamping time, and tamping path offset. The control instructions are sent to the tamping equipment in real time through the on-site control platform, enabling the equipment to perform efficient control based on the final parameters.

[0080] In addition, before issuing control commands, the module calls a preset track structure mechanical response calculation program to virtually reconstruct the target track structure based on the selected control parameters. After confirming that the tamping can achieve the expected recovery effect, the final issuance and execution are completed.

[0081] Through the coordinated operation of the above modules, this system can achieve intelligent and adaptive control of the tamping machine operation process, effectively improving the accuracy and efficiency of tamping operations.

Claims

1. A network control method for tamping machines based on multi-source data fusion, characterized in that, When multi-source data from the tamping machine operation is acquired, and the processed data indicates a need for control, execute: Determine the control benchmark for regulating demand; assess the rationality of the control benchmark in order to select the best control scheme; And generate control commands based on the optimal control scheme to guide the tamping machine to perform operations; The process of processing multi-source data to indicate the need for regulation includes: acquiring multi-source data, which includes track state parameters before operation and regulation parameters during the operation of the tamping machine; classifying the multi-source data, determining the data categories, and constructing the corresponding regulation dataset. Track condition parameters include irregularity data.

2. The tamping machine network control method based on multi-source data fusion according to claim 1, characterized in that, Classifying multi-source data, determining data categories, and constructing corresponding regulatory datasets include: Based on multi-source data, the correspondence between adjacent control datasets is identified, and the verification objects and correction objects are extracted; Calculate the matching degree between the verification object and the correction object, and output the consistency score; and classify the multi-source data into the first type of control task or the second type of control task based on the consistency score.

3. The tamping machine network control method based on multi-source data fusion according to claim 1, characterized in that, Constructing the corresponding control dataset includes: using the irregular data obtained before the operation as the comparison data and the remaining data as the reference data; The comparison results are output through a preset simulation calculation method, and then the ratio is calculated with the data source where the comparison data is located to obtain the control deviation; after correcting the control parameters and control deviation during operation, the control correction value is obtained, and the control correction value and the parameters before correction are respectively merged into the same control dataset.

4. The tamping machine network control method based on multi-source data fusion according to claim 2, characterized in that, Based on multi-source data, identifying the correspondence between adjacent regulatory datasets includes: The data to be compared in two adjacent sets of control datasets are set as the verification group and the correction group, respectively. The verification value and correction direction are set to generate the correction object; and the correction object and the verification object are used as the training input to be processed.

5. The tamping machine network control method based on multi-source data fusion according to claim 4, characterized in that, Generating a correction object includes: calculating the difference between the correction object and the verification object, and using the result as the correction ratio; Furthermore, the sign of the difference determines whether the object being checked is higher or lower than the object being corrected, thereby determining the direction of correction.

6. The tamping machine network control method based on multi-source data fusion according to claim 2, characterized in that, Calculate the matching degree between the verification object and the correction object, and output the consistency score, including: The content to be evaluated is formed by adjusting the values ​​of the correction object and the verification object based on the ratio difference. The evaluation content is simulated to obtain a conversion value; the conversion value is then calculated with the preset evaluation criteria to obtain a score result, and the score result is output as a consistency score.

7. The tamping machine network control method based on multi-source data fusion according to claim 2, characterized in that, Based on consistency scores, multi-source data can be categorized into either Category I or Category II control tasks, including: Determine whether the consistency score meets the preset similarity threshold. If it does, classify the multi-source data as the first type of control task; otherwise, classify the multi-source data as the second type of control task.

8. The tamping machine network control method based on multi-source data fusion according to claim 2, characterized in that, Determine the control benchmark for controlling demand, which includes the following when the control task is a type I control task: Obtain the trajectory status information of the current task to be executed and extract the effective fluctuation area; determine the sampling path within the effective fluctuation area and measure the geometric relationship between each sampling position to calculate the effective parameter value of each point; Based on the effective parameter values, extract the reference points in the current orbit where the change exceeds the preset threshold; and use the reference points as search conditions to find relevant parameters in the list of control parameters associated with the first type of control task, and use the search results as the preliminary control benchmark for the current operation.

9. The tamping machine network control method based on multi-source data fusion according to claim 1, characterized in that, Assessing the rationality of the regulatory benchmark in order to select the optimal regulatory approach includes: Based on the initial control benchmark, and combined with the unevenness score results in the historical operation record, a comprehensive calculation is performed to generate the object to be inspected; and a new control simulation is performed based on the object to be inspected to obtain the optimized control parameters. Furthermore, the consistency evaluation results of the optimized control parameters are calculated, and the evaluation result with the best numerical value and its corresponding control parameter are selected as the final parameters.

10. A network control system for a tamping machine based on multi-source data fusion, characterized in that, include: The data acquisition and classification module is used to acquire multi-source data during the tamping machine operation, classify the multi-source data, and construct a control dataset; The regulation task identification module is used to respond to the output of the data acquisition and classification module, identify the correspondence between adjacent regulation datasets, calculate the matching degree to output the consistency score, and classify the multi-source data into the first type of regulation task or the second type of regulation task based on the consistency score. The control benchmark determination module is used to respond to the classification results of the control task identification module. When the classification result is a first-class control task, it determines the preliminary control benchmark for the first-class control task. The scheme optimization and evaluation module is used to respond to the initial control benchmark of the control benchmark determination module, combine it with historical operation records for comprehensive calculation, obtain the optimized control parameters through control simulation, and calculate the consistency evaluation results of the optimized control parameters in order to select the best control scheme. The instruction generation and output module is used to respond to the optimal control scheme of the scheme optimization and evaluation module, and generate control instructions to guide the tamping machine to perform operations.

Images