A data processing method and device for off-road container fork truck maintenance training

Abstract

The application discloses a kind of data processing method and device for off-road container fork truck maintenance training, the method comprises: obtaining fork truck maintenance historical data set;The fork truck maintenance historical data set is handled, and the fork truck maintenance standardization process data set is obtained;Real-time operation data of user is obtained;The real-time operation data of user and the fork truck maintenance standardization process data set are handled, and the training evaluation result of user is obtained.It can be seen that the application can accurately record and evaluate the operation process of user, evaluate the training result of user, effectively manage and utilize training data, so as to improve the quality and efficiency of maintenance training.

Description

Technical Field

[0001] This invention relates to the field of human-computer interaction technology, and in particular to a data processing method and apparatus for off-road container forklift maintenance training. Background Technology

[0002] With the continuous development of industrial technology, equipment is becoming increasingly complex, placing higher demands on the skills of maintenance personnel. Traditional maintenance training methods either rely on the master-apprentice model, which is inefficient, costly, and unsuitable for large-scale training; or on static or semi-static teaching materials such as text, images, and videos. While these materials provide rich information, they have limitations in practical training. For example, trainees cannot intuitively simulate actual maintenance operations and cannot receive real-time feedback and guidance. To overcome these limitations, the application of human-computer interaction (HCI) technology in maintenance training systems has gained increasing attention in recent years. Among these, gesture recognition technology, as a natural, intuitive, and equipment-free HCI interaction method, has significant advantages. Gestures can simulate various actions in actual maintenance operations, such as rotation, pressing, and stretching, thus providing trainees with a more realistic and immersive training experience. Therefore, this paper provides a data processing method and device for maintenance training systems to improve gesture recognition accuracy, thereby improving the accuracy and efficiency of interactive control in maintenance training systems and enhancing the user's HCI interaction experience. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a data processing method and apparatus for off-road container forklift maintenance training, which can accurately record and evaluate the user's operation process, provide personalized training suggestions, and effectively manage and utilize training data.

[0004] To address the aforementioned technical problems, a first aspect of the present invention discloses a data processing method for off-road container forklift maintenance training, the method comprising: S1, Obtain the forklift maintenance history dataset; S2, process the forklift maintenance history dataset to obtain a standardized forklift maintenance process dataset; S3, obtains real-time user operation data; S4. Process the user's real-time operation data and the forklift maintenance standardized process dataset to obtain the user's training evaluation results.

[0005] As an optional implementation, in the first aspect of the present invention, processing the forklift maintenance historical dataset to obtain a standardized forklift maintenance process dataset includes: S21, Filter the forklift maintenance history dataset to obtain a filtered dataset; S22, Label the filtered dataset to obtain a labeled text dataset; S23, perform a completeness check on the labeled text dataset to obtain a completeness check result; if the completeness check result is yes, execute S24; if the completeness check result is no, supplement the filtered dataset to obtain a supplemented filtered dataset, and execute S23. S24, extract key information of maintenance operation from the labeled text dataset; S25, process the key information of the maintenance operation to obtain time sequence information and spatial information; S26, The time sequence information and spatial information are processed to obtain a standardized state diagram for forklift maintenance; the standardized state diagram for forklift maintenance includes a pre-maintenance preparation state diagram, a repair maintenance state diagram, a preventive maintenance state diagram, and a predictive maintenance state diagram. S27, Integrate the forklift maintenance standardized state diagram to obtain a forklift maintenance standardized process dataset.

[0006] As an optional implementation, in the first aspect of the present invention, the extraction of key maintenance operation information from the labeled text dataset includes: S241, process any data in the labeled text dataset to obtain a document embedding vector; S242, Perform dimensionality reduction processing on the document embedding vector to obtain a dimensionality-reduced vector; S243, perform clustering on the reduced-dimensional vector to obtain M category vectors for each topic, where M is a positive integer; S244: Represent the category vector of each topic to obtain key information for maintenance operations.

[0007] As an optional implementation, in the first aspect of the present invention, processing any data in the annotated text dataset to obtain a document embedding vector includes: S2411, Process any data in the labeled text dataset to obtain the inserted label data sequence; S2412, the inserted label data sequence is mapped to a preset Chinese character vector dictionary to obtain character embedding information; S2413, perform position embedding on the inserted label data sequence to obtain position embedding information; S2414, The character embedding information and the position embedding information are processed to obtain the document embedding vector.

[0008] As an optional implementation, in the first aspect of the present invention, the dimensionality reduction processing of the document embedding vector to obtain a dimensionality-reduced vector includes: S2421, for the document embedding vector After processing, the connection structure diagram is obtained: For vector dimensions, for The corresponding connection structure diagram, express The Euclidean distance between them express The Euclidean distance between them; , for The k-nearest neighbor average threshold, i.e., the local density. , The maximum local density within the neighborhood; S2422, Calculate two points Connection probability between : For point The corresponding scale parameters are preset parameters. ; S2423, regarding the connection probability The objective function is obtained through processing. The objective function expression is: In the formula, The learning rate controls the step size for each update. Assume the low-dimensional vector is... , For dimension, ,in These are hyperparameters, with values ​​ranging from 0 to 1. , The regularization coefficient is . , for The mean; S2424, The objective function and the preset random initialization dimensionality reduction vector are processed to obtain the dimensionality reduction vector.

[0009] As an optional implementation, in the first aspect of the present invention, the clustering process of the dimensionality-reduced vector to obtain M category vectors of topics includes: S2431, for the dimensionality reduction vector The local density of each point is obtained through processing. , i =1,2,..., m , m The number of data; in, The preset density kernel; Calculate the high-density distance for each point : Select The top k largest points are used as the initial cluster centers; S2432, Based on the initial cluster center, calculate the cluster affiliation of each point and construct a cluster set. The formula for calculating distance is: in, , This is the separation balance coefficient, ranging from 0.1 to 0.3, used to adjust intra-cluster compactness and inter-cluster discriminability. For clusters The distance value, For the first Clusters, ; S2433, cluster The process is performed to obtain the offset vector. : The offset vector The expression is: in, For clusters The set of points within, As the cluster centroid, The principal direction vector within the cluster is obtained through the first principal component of PCA. This indicates a direction consistency indicator function, retaining those with an angle less than the principal direction. The vector, The directional consistency threshold is set to a value between 0.7 and 0.85. S2434, regarding the offset vector The process yields category vectors for M topics.

[0010] As an optional implementation, in the first aspect of the present invention, the step of representing the category vector of each topic to obtain key information for maintenance operations includes: S2441, calculate the importance of each topic's category vector to obtain an importance score for each topic; The expression for the importance score is as follows: in, For thematic clusters Importance score Indicates a candidate word. Represents the set of all topic clusters. Indicates the inclusion word The number of other topic clusters, Indicates the number of topic clusters; S2442, Based on the importance score of each topic, select the top N words as the key information for maintenance operations of each topic cluster.

[0011] A second aspect of this invention discloses a data processing device for off-road container forklift maintenance training, the device comprising: The historical data acquisition module is used to acquire historical datasets of forklift maintenance. The data processing module is used to process the forklift maintenance historical dataset to obtain a standardized forklift maintenance process dataset. The real-time operation data acquisition module is used to acquire the user's real-time operation data; The training evaluation module is used to process the user's real-time operation data and the forklift maintenance standardized process dataset to obtain the user's training evaluation results.

[0012] As an optional implementation, in the second aspect of the present invention, processing the forklift maintenance historical dataset to obtain a standardized forklift maintenance process dataset includes: S21, Filter the forklift maintenance history dataset to obtain a filtered dataset; S22, Label the filtered dataset to obtain a labeled text dataset; S23, perform a completeness check on the labeled text dataset to obtain a completeness check result; if the completeness check result is yes, execute S24; if the completeness check result is no, supplement the filtered dataset to obtain a supplemented filtered dataset, and execute S23. S24, extract key information of maintenance operation from the labeled text dataset; S25, process the key information of the maintenance operation to obtain time sequence information and spatial information; S26, The time sequence information and spatial information are processed to obtain a standardized state diagram for forklift maintenance; the standardized state diagram for forklift maintenance includes a pre-maintenance preparation state diagram, a repair maintenance state diagram, a preventive maintenance state diagram, and a predictive maintenance state diagram. S27, Integrate the forklift maintenance standardized state diagram to obtain a forklift maintenance standardized process dataset.

[0013] As an optional implementation, in the second aspect of the present invention, the extraction of key maintenance operation information from the labeled text dataset includes: S241, process any data in the labeled text dataset to obtain a document embedding vector; S242, Perform dimensionality reduction processing on the document embedding vector to obtain a dimensionality-reduced vector; S243, perform clustering on the reduced-dimensional vector to obtain M category vectors for each topic, where M is a positive integer; S244: Represent the category vector of each topic to obtain key information for maintenance operations.

[0014] As an optional implementation, in the second aspect of the present invention, processing any data in the annotated text dataset to obtain a document embedding vector includes: S2411, Process any data in the labeled text dataset to obtain the inserted label data sequence; S2412, the inserted label data sequence is mapped to a preset Chinese character vector dictionary to obtain character embedding information; S2413, perform position embedding on the inserted label data sequence to obtain position embedding information; S2414, The character embedding information and the position embedding information are processed to obtain the document embedding vector.

[0015] As an optional implementation, in the second aspect of the present invention, the dimensionality reduction processing of the document embedding vector to obtain a dimensionality-reduced vector includes: S2421, for the document embedding vector After processing, the connection structure diagram is obtained: For vector dimensions, for The corresponding connection structure diagram, express The Euclidean distance between them express The Euclidean distance between them; , for The k-nearest neighbor average threshold, i.e., the local density. , The maximum local density within the neighborhood; S2422, Calculate two points Connection probability between : For point The corresponding scale parameters are preset parameters. ; S2423, regarding the connection probability The objective function is obtained through processing. The objective function expression is: In the formula, The learning rate controls the step size for each update. Assume the low-dimensional vector is... , For dimension, ,in These are hyperparameters, with values ​​ranging from 0 to 1. , The regularization coefficient is . , for The mean; S2424, The objective function and the preset random initialization dimensionality reduction vector are processed to obtain the dimensionality reduction vector.

[0016] As an optional implementation, in the second aspect of the present invention, the clustering process of the dimensionality-reduced vector to obtain M category vectors for topics includes: S2431, for the dimensionality reduction vector The local density of each point is obtained through processing. , i =1,2,..., m , m The number of data; in, The preset density kernel; Calculate the high-density distance for each point : Select The top k largest points are used as the initial cluster centers; S2432, Based on the initial cluster center, calculate the cluster affiliation of each point and construct a cluster set. The formula for calculating distance is: in, , This is the separation balance coefficient, ranging from 0.1 to 0.3, used to adjust intra-cluster compactness and inter-cluster discriminability. For clusters The distance value, For the first Clusters, ; S2433, cluster The process is performed to obtain the offset vector. : The offset vector The expression is: in, For clusters The set of points within, As the cluster centroid, The principal direction vector within the cluster is obtained through the first principal component of PCA. This indicates a direction consistency indicator function, retaining those with an angle less than the principal direction. The vector, The directional consistency threshold is set to a value between 0.7 and 0.85. S2434, regarding the offset vector The process yields category vectors for M topics.

[0017] As an optional implementation, in the second aspect of the present invention, the step of representing the category vector of each topic to obtain key information for maintenance operations includes: S2441, calculate the importance of each topic's category vector to obtain an importance score for each topic; The expression for the importance score is as follows: in, For thematic clusters Importance score Indicates a candidate word. Represents the set of all topic clusters. Indicates the inclusion word The number of other topic clusters, Indicates the number of topic clusters; S2442, Based on the importance score of each topic, select the top N words as the key information for maintenance operations of each topic cluster.

[0018] A third aspect of the present invention discloses another data processing apparatus for off-road container forklift maintenance training, the apparatus comprising: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute some or all of the steps in the data processing method for off-road container forklift maintenance training disclosed in the first aspect of the present invention.

[0019] The fourth aspect of the present invention discloses a computer-storable medium storing computer instructions, which, when invoked, are used to execute some or all of the steps in the data processing method for off-road container forklift maintenance training disclosed in the first aspect of the present invention.

[0020] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: This invention discloses a data processing method for off-road container forklift maintenance training. It can accurately record and evaluate the user's operation process, evaluate the user's training results, accurately record and evaluate the trainee's operation process, provide personalized training suggestions, and effectively manage and utilize training data, thereby improving the quality and efficiency of maintenance training and reducing training costs. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating a data processing method for off-road container forklift maintenance training disclosed in an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of a data processing device for off-road container forklift maintenance training disclosed in an embodiment of the present invention; Figure 3 This is a schematic diagram of another data processing device for off-road container forklift maintenance training disclosed in an embodiment of the present invention. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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.

[0024] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0025] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0026] In all embodiments of the present invention, the variables involved in all computational expressions or mathematical functions are dimensionless before calculation. The values ​​of the independent variables in all computational expressions or mathematical functions in these embodiments conform to the reasonable requirements of the input range of the computational expression or mathematical function, ensuring that the computational expression or mathematical function can be calculated smoothly without violating physical laws or mathematical rules.

[0027] This invention discloses a data processing method and apparatus for training in the maintenance of off-road container forklifts. The method includes: acquiring a historical forklift maintenance dataset; processing the historical forklift maintenance dataset to obtain a standardized forklift maintenance process dataset; acquiring real-time user operation data; and processing the user's real-time operation data and the standardized forklift maintenance process dataset to obtain the user's training evaluation result. Therefore, this invention can accurately record and evaluate the user's operation process, assess the user's training results, and effectively manage and utilize training data, thereby improving the quality and efficiency of maintenance training. Detailed descriptions follow.

[0028] Example 1 Please see Figure 1, Figure 1 This is a flowchart illustrating a data processing method for off-road container forklift maintenance training disclosed in an embodiment of the present invention. Figure 1 The data processing method described for off-road container forklift maintenance training is applied in the field of human-computer interaction technology, and the embodiments of this invention are not limited thereto. Figure 1 As shown, the data processing method for off-road container forklift maintenance training may include the following operations: S1, Obtain the forklift maintenance history dataset; This includes tool usage data, parts disassembly data, installation sequence data, operating force data, operating time data, operating position data, and operating angle data; S2, process the forklift maintenance history dataset to obtain a standardized forklift maintenance process dataset; S3, obtains real-time user operation data; In this embodiment, sensor devices, including wearable sensors (such as smart gloves and smart glasses), tool sensors (such as torque sensors and pressure sensors), environmental sensors (such as position sensors and angle sensors), and vision sensors (such as cameras), are used to acquire real-time user operation data. Operation data includes, but is not limited to: tool usage data (such as the type of tool used and tool operation time), parts disassembly / installation sequence data, operation force data, operation time data, operation position data, and operation angle data.

[0029] S4. Process the user's real-time operation data and the forklift maintenance standardized process dataset to obtain the user's training evaluation results.

[0030] The user's real-time operation data and the forklift maintenance standardized process dataset are processed to obtain the operation step sequence model and the standard operation step sequence model, the operation time sequence relationship model and the standard operation time sequence relationship model, and the operation parameter association model and the standard operation parameter association model. Optionally, training evaluation results may include the correctness of operating procedures, the rationality of operating sequence, and the accuracy of operating parameters.

[0031] Calculate the similarity between the operation step sequence model and the standard operation step sequence model; S_seq=LCS(Seq_user,Seq_std) / max(len(Seq_user),len(Seq_std))×100 points In the formula: LCS is the longest common subsequence algorithm, which matches compliant and valid process nodes; Seq_user is the student operation step sequence model; Seq_std is the standard operation step sequence model; the full score is 100 points. len(Seq_user) is the total number of steps in the student operation step sequence model, and len(Seq_std) is the total number of processes in the standard operation step sequence model; Analyze the differences between the operation timing relationship model and the standard operation timing relationship model; In the formula: _user represents the actual time spent by the student in step i; _std represents the standard compliance time for step i; λ1 is the timeout penalty coefficient, commonly ranging from 1.2 to 1.5; the larger the cumulative timing deviation, the more points are deducted.

[0032] Compare the degree of matching between the operational parameter association model and the standard operational parameter association model.

[0033] point In the formula: n1 is the total number of core monitoring parameters (torque, pressure, angle, position, etc.); _user is the real-time parameter for the k-th practical operation; _std is the standard center value of the k-th term; Let the safe allowable floating range be the k-th parameter.

[0034] In the formula: , , These are compliance weighting coefficients, such as S_seq weight 0.4, S_time weight 0.15, S_param weight 0.45, etc., which can be set by the experiment.

[0035] Optionally, the process of processing the forklift maintenance history dataset to obtain a standardized forklift maintenance process dataset includes: S21, Filter the forklift maintenance history dataset to obtain a filtered dataset; The purpose of filtering is to remove data that is irrelevant to forklift maintenance, perform data noise reduction, data standardization, and data format conversion to improve data quality and consistency.

[0036] S22, Label the filtered dataset to obtain a labeled text dataset; Initial labels are generated through rule matching and weakly supervised model pre-labeling, and then multiple rounds of cross-validation and correction are performed manually to obtain an annotated text dataset, ensuring label consistency and accuracy.

[0037] S23, perform a completeness check on the labeled text dataset to obtain a completeness check result; if the completeness check result is yes, execute S24; if the completeness check result is no, supplement the filtered dataset to obtain a supplemented filtered dataset, and execute S23. Completeness verification includes: field-level verification, checking whether all required fields of text samples (such as text content, tag type, annotator ID, annotation time) are complete, without empty values ​​or format errors; sample-level verification, statistically analyzing the sample distribution, coverage, and boundary sample completeness of each tag category to verify that no category is omitted or no sample is missing; and rule consistency verification, comparing the annotation results with the preset annotation specifications to verify the consistency of the tag definition and the compliance of the annotation logic.

[0038] S24, extract key information of maintenance operation from the labeled text dataset; The labeled text dataset is processed to obtain a schema of entities and relationships in the maintenance domain, and key information types are defined (such as fault phenomena, part numbers, operation steps, tool models, parameter thresholds, etc.). Entity recognition and relationship extraction are performed using a domain pre-trained model, and structured maintenance operation keys are obtained through rule verification and manual review.

[0039] S25, process the key information of the maintenance operation to obtain time sequence information and spatial information; The steps for calculating time-series information are: performing functional time encoding to find a mapping from the time domain to a d-dimensional vector space. Any two timestamps Learn the relationship pattern between these two time intervals, representing it as the inner product of their corresponding time codes, i.e. The aforementioned time series pattern is further formalized into a kernel function with translation invariance. And its mapping function is .

[0040] To learn about the aforementioned timing kernel function and its mapping function, inspired by Mercer's theory, the mapping function can be defined as follows: by Mapping function for frequency It can be further formalized as Using truncated d-dimensional mapping functions To approximate the original infinite-dimensional mapping. One cycle (i.e.) The mapping functions are concatenated to form the final time code. Given a specific node v, the time encoding of that node can be further defined as: , in, It is a series of node-related mapping functions used to calculate Fourier coefficients. It is implemented using a multilayer perceptron. And force the output of the last layer of the perceptron to be a positive number, so as to satisfy the inherent characteristics of Mercer's theory.

[0041] The time intervals associated with each node in the time series graph are temporally encoded to model the temporal patterns of the nodes over continuous time. A Graph Attention Network (C2GAT) is then used to adaptively learn the number of temporal information points on the graph.

[0042] The steps for calculating spatial information: Spatial information refers to location information in the modeling space, achieved through hierarchical geocoding. The entire map is divided into four equal parts, labeled 0, 1, 2, and 3. The tile labeled "0" is then further divided into four smaller tiles, labeled 00, 01, 02, and 03. These smaller tiles share a common prefix: "0," which is the number of the tile above them. The same division process is applied to the tiles labeled "1," "2," and "3." This process is recursively repeated until the minimum required size of the tile is met.

[0043] Each operational step in the critical information of maintenance operations includes timing information and spatial information; S26, The time sequence information and spatial information are processed to obtain a standardized state diagram for forklift maintenance; the standardized state diagram for forklift maintenance includes a pre-maintenance preparation state diagram, a repair maintenance state diagram, a preventive maintenance state diagram, and a predictive maintenance state diagram. In the standardized state diagram of forklift maintenance, the nodes are the key information of maintenance operations, including two attributes: temporal information and spatial information. Each edge represents the relationship between the key information of various maintenance operations, and the direction of the arrows represents the information path. S27, Integrate the forklift maintenance standardized state diagram to obtain a forklift maintenance standardized process dataset.

[0044] The collected operational data, generated operational process models, operational evaluation results, and training suggestions are stored in the database, and data indexes and relationships are established.

[0045] Sequence mining algorithms are used to analyze the sequential relationships between operation steps and to construct an operation step sequence model. Time series analysis methods are used to analyze the relationship between operation time and operation sequence, and an operation time series relationship model is constructed. The association rule mining algorithm is used to analyze the association relationships between operation parameters and to construct an operation parameter association model.

[0046] Optionally, the step of extracting key information from the labeled text dataset to obtain maintenance operation information includes: S241, process any data in the labeled text dataset to obtain a document embedding vector; S242, Perform dimensionality reduction processing on the document embedding vector to obtain a dimensionality-reduced vector; S243, perform clustering on the reduced-dimensional vector to obtain M category vectors for each topic, where M is a positive integer; S244: Represent the category vector of each topic to obtain key information for maintenance operations.

[0047] Optionally, processing any data in the labeled text dataset to obtain a document embedding vector includes: S2411, Process any data in the labeled text dataset to obtain the inserted label data sequence; For any sentence in the labeled text data Processing, in Previously, the first tag was added, in Then, a second label is added to obtain the text sequence. ,in For sentences The first in One character; Determine the text sequence The length, if The length is less than 256, in Insert a third tag between the second tag if the text sequence If the length of the text sequence is greater than 256, delete characters sequentially before the second label until the length of the text sequence equals 256, resulting in the interpolated text sequence. .

[0048] S2412, the inserted label data sequence is mapped to a preset Chinese character vector dictionary to obtain character embedding information; S2413, perform position embedding on the inserted label data sequence to obtain position embedding information; S2414, The character embedding information and the position embedding information are processed to obtain the document embedding vector.

[0049] Optionally, the step of reducing the dimensionality of the document embedding vector to obtain a reduced-dimensional vector includes: S2421, for the document embedding vector After processing, the connection structure diagram is obtained: For vector dimensions, for The corresponding connection structure diagram, express The Euclidean distance between them express The Euclidean distance between them; , for The k-nearest neighbor average threshold, i.e., the local density. , The maximum local density within the neighborhood; S2422, Calculate two points Connection probability between : For point The corresponding scale parameters are preset parameters. ; S2423, regarding the connection probability The objective function is obtained through processing. The objective function expression is: In the formula, The learning rate controls the step size for each update. Assume the low-dimensional vector is... , For dimension, ,in These are hyperparameters, with values ​​ranging from 0 to 1. , The regularization coefficient is . , for The mean; S2424, The objective function and the preset random initialization dimensionality reduction vector are processed to obtain the dimensionality reduction vector.

[0050] Specifically, the method involves randomly initializing the dimensionality reduction vector and continuously updating and iterating it to minimize the objective function. In this way, the structure in the high-dimensional space is preserved in the low-dimensional space, resulting in the final dimensionality reduction result (the dimensionality reduction vector).

[0051] Optionally, the clustering process performed on the reduced-dimensional vector to obtain M category vectors for different topics includes: S2431, for the dimensionality reduction vector The local density of each point is obtained through processing. , i =1,2,..., m , m The number of data; in, The preset density kernel; Calculate the high-density distance for each point : Select The top k largest points are used as the initial cluster centers; S2432, Based on the initial cluster center, calculate the cluster affiliation of each point and construct a cluster set. The formula for calculating distance is: in, , This is the separation balance coefficient, ranging from 0.1 to 0.3, used to adjust intra-cluster compactness and inter-cluster discriminability. For clusters The distance value, For the first Clusters, ; S2433, cluster The process is performed to obtain the offset vector. : The offset vector The expression is: in, For clusters The set of points within, As the cluster centroid, The principal direction vector within the cluster is obtained through the first principal component of PCA. This indicates a direction consistency indicator function, retaining those with an angle less than the principal direction. The vector, The directional consistency threshold is set to a value between 0.7 and 0.85. S2434, regarding the offset vector The process yields category vectors for M topics.

[0052] Construct an offset vector matrix W, where each row of W corresponds to the offset vector of a cluster; construct a joint optimization objective function with orthogonal redundancy prevention constraints: In the formula, U is the adaptive orthogonal projection transformation matrix; T is the standard topic vector matrix to be solved; I This is the identity matrix of the corresponding order; The orthogonal constraint strength coefficients are preset; after the optimal solution converges, the initial topic vector is extracted row by row, and L2 global normalization is performed: The final output consists of category vectors for M topics. .

[0053] Optionally, the step of representing the category vector of each topic to obtain key information for maintenance operations includes: S2441, calculate the importance of each topic's category vector to obtain an importance score for each topic; The expression for the importance score is as follows: in, For thematic clusters Importance score Indicates a candidate word. Represents the set of all topic clusters. Indicates the inclusion word The number of other topic clusters, Indicates the number of topic clusters; S2442, Based on the importance score of each topic, select the top N words as the key information for maintenance operations of each topic cluster.

[0054] As can be seen, the present invention discloses a data processing method for off-road container forklift maintenance training, which can accurately record and evaluate the user's operation process, evaluate the user training results, accurately record and evaluate the trainee's operation process, provide personalized training suggestions, and effectively manage and utilize training data, thereby improving the quality and efficiency of maintenance training and reducing training costs.

[0055] Example 2 Please see Figure 2 , Figure 2 This is a schematic diagram of the structure of a data processing device for off-road container forklift maintenance training disclosed in an embodiment of the present invention. Figure 2 The data processing device described for off-road container forklift maintenance training is applied in the field of human-computer interaction technology, and the embodiments of this invention are not limited thereto. Figure 2As shown, the data processing device for off-road container forklift maintenance training may include the following operations: S301, Historical Data Acquisition Module, is used to acquire historical datasets of forklift maintenance. S302, Data processing module, used to process the forklift maintenance historical dataset to obtain a forklift maintenance standardized process dataset; S303, Real-time Operation Data Acquisition Module, is used to acquire the user's real-time operation data; S304, Training Evaluation Module, is used to process the user's real-time operation data and the forklift maintenance standardized process dataset to obtain the user's training evaluation results.

[0056] Example 3 Please see Figure 3 , Figure 3 This is a schematic diagram of another data processing device for off-road container forklift maintenance training disclosed in an embodiment of the present invention. Figure 3 The data processing described for off-road container forklift maintenance training is applied in the field of human-computer interaction technology, and the embodiments of this invention are not limited thereto. Figure 3 As shown, the data processing used for off-road container forklift maintenance training may include the following operations: Memory 401 storing executable program code; Processor 402 coupled to memory 401; The processor 402 calls the executable program code stored in the memory 401 to execute the steps in the data processing method for off-road container forklift maintenance training described in Embodiment 1.

[0057] Example 4 This invention discloses a computer-readable storage medium storing a computer program for electronic data interchange, wherein the computer program enables a computer to perform the steps in the data processing method for off-road container forklift maintenance training described in Embodiment 1.

[0058] The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0059] Through the detailed description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0060] Finally, it should be noted that the data processing method and apparatus for off-road container forklift maintenance training disclosed in the embodiments of the present invention are merely preferred embodiments of the present invention and are only used to illustrate the technical solutions of the present invention, not to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A data processing method for training the maintenance of off-road container forklifts, characterized in that, The method includes: S1, Obtain the forklift maintenance history dataset; S2, Process the forklift maintenance history dataset to obtain a standardized forklift maintenance process dataset; S3, obtains real-time user operation data; S4. Process the user's real-time operation data and the forklift maintenance standardized process dataset to obtain the user's training evaluation results.

2. The data processing method for off-road container forklift maintenance training according to claim 1, characterized in that, The process of processing the forklift maintenance history dataset to obtain a standardized forklift maintenance process dataset includes: S21, Filter the forklift maintenance history dataset to obtain a filtered dataset; S22, Label the filtered dataset to obtain a labeled text dataset; S23, perform a completeness check on the labeled text dataset to obtain a completeness check result; if the completeness check result is yes, execute S24; if the completeness check result is no, supplement the filtered dataset to obtain a supplemented filtered dataset, and execute S23. S24, extract key information of maintenance operation from the labeled text dataset; S25, process the key information of the maintenance operation to obtain time sequence information and spatial information; S26, The time sequence information and spatial information are processed to obtain a standardized state diagram for forklift maintenance; the standardized state diagram for forklift maintenance includes a pre-maintenance preparation state diagram, a repair maintenance state diagram, a preventive maintenance state diagram, and a predictive maintenance state diagram. S27, Integrate the forklift maintenance standardized state diagram to obtain a forklift maintenance standardized process dataset.

3. The data processing method for off-road container forklift maintenance training according to claim 2, characterized in that, The extraction of key maintenance operation information from the labeled text dataset includes: S241, process any data in the labeled text dataset to obtain a document embedding vector; S242, Perform dimensionality reduction processing on the document embedding vector to obtain a dimensionality-reduced vector; S243, perform clustering on the reduced-dimensional vector to obtain M category vectors for each topic, where M is a positive integer; S244: Represent the category vector of each topic to obtain key information for maintenance operations.

4. The data processing method for off-road container forklift maintenance training according to claim 3, characterized in that, The process of processing any data in the annotated text dataset to obtain a document embedding vector includes: S2411, Process any data in the labeled text dataset to obtain the inserted label data sequence; S2412, the inserted label data sequence is mapped to a preset Chinese character vector dictionary to obtain character embedding information; S2413, perform position embedding on the inserted label data sequence to obtain position embedding information; S2414, The character embedding information and the position embedding information are processed to obtain the document embedding vector.

5. The data processing method for off-road container forklift maintenance training according to claim 3, characterized in that, The step of reducing the dimensionality of the document embedding vector to obtain a reduced-dimensional vector includes: S2421, for the document embedding vector After processing, the connection structure diagram is obtained: For vector dimensions, for The corresponding connection structure diagram, express The Euclidean distance between them express The Euclidean distance between them; , for The k-nearest neighbor average threshold, i.e., the local density. , The maximum local density within the neighborhood; S2422, Calculate two points Connection probability between : For point The corresponding scale parameters are preset parameters. ; S2423, regarding the connection probability The objective function is obtained through processing. The objective function expression is: In the formula, The learning rate controls the step size for each update. Assume the low-dimensional vector is... , For dimension, ,in These are hyperparameters, with values ​​ranging from 0 to 1. , The regularization coefficient is . , for The mean; S2424, The objective function and the preset random initialization dimensionality reduction vector are processed to obtain the dimensionality reduction vector.

6. The data processing method for off-road container forklift maintenance training according to claim 3, characterized in that, The clustering process performed on the reduced-dimensional vector yields M category vectors for different topics, including: S2431, for the dimensionality reduction vector The local density of each point is obtained through processing. , i =1,2,..., m , m The number of data; in, The preset density kernel; Calculate the high-density distance for each point : Select The top k largest points are used as the initial cluster centers; S2432, Based on the initial cluster center, calculate the cluster affiliation of each point and construct a cluster set. The formula for calculating distance is: in, , This is the separation balance coefficient, ranging from 0.1 to 0.3, used to adjust intra-cluster compactness and inter-cluster discriminability. For clusters The distance value, For the first Clusters, ; S2433, cluster The process is performed to obtain the offset vector. : The offset vector The expression is: in, For clusters The set of points within, As the cluster centroid, The principal direction vector within the cluster is obtained through the first principal component of PCA. This indicates a direction consistency indicator function, retaining those with an angle less than the principal direction. The vector, The directional consistency threshold is set to a value between 0.7 and 0.

85. S2434, regarding the offset vector The process yields category vectors for M topics.

7. The data processing method for off-road container forklift maintenance training according to claim 3, characterized in that, The process of representing the category vector of each topic to obtain key information for maintenance operations includes: S2441, calculate the importance of each topic's category vector to obtain an importance score for each topic; The expression for the importance score is as follows: in, For thematic clusters Importance score Indicates a candidate word. Represents the set of all topic clusters. Indicates the inclusion word The number of other topic clusters, Indicates the number of topic clusters; S2442, Based on the importance score of each topic, select the top N words as the key information for maintenance operations of each topic cluster.

8. A data processing device for training in the maintenance of off-road container forklifts, characterized in that, The device includes: The historical data acquisition module is used to acquire historical datasets of forklift maintenance. The data processing module is used to process the forklift maintenance historical dataset to obtain a standardized forklift maintenance process dataset. The real-time operation data acquisition module is used to acquire the user's real-time operation data; The training evaluation module is used to process the user's real-time operation data and the forklift maintenance standardized process dataset to obtain the user's training evaluation results.

9. A data processing device for training in the maintenance of off-road container forklifts, characterized in that, The device includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the data processing method for off-road container forklift maintenance training as described in any one of claims 1-7.

10. A computer-storable medium, characterized in that, The computer storage medium stores computer instructions, which, when invoked, are used to execute the data processing method for off-road container forklift maintenance training as described in any one of claims 1-7.

Images