Production personnel management method, system, device and medium based on industrial internet of things

Abstract

This application discloses a method, system, equipment, and medium for production personnel management based on the Industrial Internet of Things (IIoT), relating to the technical field of IIoT. The method includes: determining association feature values ​​between a single inspection personnel and a single production personnel based on the inspection data and the work posture data; constructing association vectors corresponding to the single inspection personnel and each production personnel based on the association feature values, and constructing association matrices corresponding to each inspection personnel and each production personnel based on the association feature values; performing matrix decomposition on the association matrix to extract the main feature directions; for each association vector, determining directional similarity and projection length based on the association vector and the main feature directions, and determining inspection quality indicators based on the directional similarity and projection length; and assigning regions to each inspection personnel based on the inspection quality indicators. This application improves the efficiency of production personnel management.

Description

Technical Field

[0001] This application relates to the technical field of industrial Internet of Things (IIoT), and in particular to production personnel management methods, systems, equipment, and media based on IIoT. Background Technology

[0002] In production management practices within an Industrial Internet of Things (IIoT) environment, the supervision and assurance of production personnel's working status primarily rely on inspection personnel. Existing management methods generally depend on pre-set fixed inspection routes or area allocation based on historical experience. Their core focus is on ensuring the physical coverage of inspection activities. Although this approach can utilize devices such as position sensors to collect basic inspection trajectories and production personnel behavior data, it merely achieves data aggregation, and the value of the data is not fully realized.

[0003] The fundamental bottleneck currently facing technology lies in the disconnect between inspection data and production personnel's work data at the analytical level. While the system records a large amount of raw data, it cannot effectively reveal the intrinsic link between inspection activities and production results. Due to the lack of scientific methods for quantifying the quality of inspection work, management decisions struggle to move beyond simply recording "whether to inspect" and cannot assess "how effective the inspection is." This directly leads to a lack of dynamic optimization basis for the allocation of inspection resources, making it difficult to achieve the management goal of precise adjustments based on actual effectiveness. Summary of the Invention

[0004] To improve the efficiency of production personnel management, this application provides a production personnel management method, system, equipment, and medium based on the Industrial Internet of Things.

[0005] Firstly, this application provides a production personnel management method based on the Industrial Internet of Things, employing the following technical solution: A production personnel management method based on the Industrial Internet of Things (IIoT) is applied to an IIoT system, which includes a management platform, a sensor network platform, and an object platform connected in sequence. The method is executed by the management platform and includes: The inspection data of each inspection personnel and the work posture data of each production personnel are obtained, and the correlation feature value between a single inspection personnel and a single production personnel is determined based on the inspection data and the work posture data. The inspection data includes position data and action posture data, and the correlation feature value is used to characterize the influence effectiveness of the single inspection personnel on the work status of the single production personnel. Based on the associated feature values, construct an association vector between a single inspection personnel and each production personnel, and construct an association matrix between each inspection personnel and each production personnel based on the associated feature values. Perform matrix decomposition on the correlation matrix to extract the main feature directions; For each associated vector, the directional similarity and projection length are determined based on the associated vector and the main feature direction, and the inspection quality index of each inspection personnel is determined based on the directional similarity and the projection length. Based on the aforementioned inspection quality indicators, each inspection personnel is assigned to a specific area to enable them to conduct inspections of production personnel within that area.

[0006] By adopting the above technical solution, the inspection data of each inspector and the work posture data of each production worker are first acquired. Based on the inspection data and work posture data, the correlation feature values ​​between individual inspectors and individual production workers are determined. The inspection data includes position data and action posture data. The correlation feature values ​​characterize the effectiveness of an individual inspector's influence on the work status of an individual production worker. Then, a correlation vector is constructed between each individual inspector and each production worker based on the correlation feature values. A correlation matrix is ​​then constructed between each individual inspector and each production worker based on the correlation feature values. The correlation matrix is ​​then decomposed to extract the main feature directions. For each correlation vector, the directional similarity and projection length are determined based on the correlation vector and the main feature directions. Finally, the inspection quality indicators for each inspector are determined based on the directional similarity and projection length. Based on inspection quality indicators, each inspection personnel is assigned to a specific area to ensure inspection coverage of production personnel within that area. This method changes the traditional production personnel inspection management model that relies on fixed routes and static experience, achieving a paradigm shift from "physical space coverage" to "behavioral efficiency optimization." By deeply integrating and causally analyzing fragmented inspection data collected by the Industrial Internet of Things with production personnel work data, a scientific and quantitative assessment of the work quality of inspection personnel is achieved, quantifying the real impact of inspection behavior on production efficiency. Through the data-driven discovery-based optimal inspection model of this invention, inspection resources are dynamically and accurately allocated, placing high-performing inspection personnel in the most needed areas. This forms a complete closed loop from data perception and intelligent analysis to decision optimization, thereby improving inspection efficiency and enhancing the management efficiency of production personnel.

[0007] Optionally, the step of determining the correlation feature value between a single inspection personnel and a single production personnel based on the inspection data and the work posture data includes: For each inspector, the corresponding inspection data is identified to obtain the corresponding inspection state change sequence. The inspection state change sequence is used to characterize the changes in the action posture and spatial position of the individual inspector over time. The inspection state change sequence includes a position change sequence and an action posture change sequence. For each production worker, the corresponding work posture data is identified to obtain a corresponding work posture change sequence, wherein the work posture change sequence is used to characterize the change of the work posture of the individual production worker over time. Based on the intervention behavior recognition model, the sequence of action posture changes is classified in a fine-grained manner to obtain the sequence of intervention behavior types, wherein the sequence of intervention behavior types includes the intervention behavior of the inspection personnel and the time of occurrence of the behavior; Based on the work posture change sequence, abnormal work status events of each production worker are extracted. The abnormal work status events are used to characterize the change of the production worker's work status from an abnormal state to a normal state. The abnormal state includes non-standard operation state and leisure state. For each production worker, the workstation position of the production worker is obtained, and based on the workstation position and the position change sequence, a spatial proximity event between the inspection personnel and the production worker is determined, wherein the spatial proximity event is used to characterize the inspection personnel entering the preset proximity range of the production worker; Based on the temporal correlation or overlap between the abnormal work status event and the spatial proximity event, a first intervention indicator for the single inspection personnel on the single production personnel is determined, and based on the causal relationship between the sequence of intervention behavior types and the abnormal work status event, a second intervention indicator for the single inspection personnel on the single production personnel is determined. The first intervention indicator includes a first number of interventions and a first intervention success rate, and the second intervention indicator includes a second number of interventions and a second intervention success rate. Based on the first intervention indicator and the second intervention indicator, the correlation characteristic value between the individual inspection personnel and the individual production personnel is determined.

[0008] By adopting the above technical solution, in order to determine the correlation feature values ​​between a single inspector and a single production worker, for each inspector, the corresponding inspection data is identified to obtain a corresponding inspection state change sequence. This inspection state change sequence characterizes the changes in the individual inspector's posture and spatial position over time, and includes a position change sequence and a posture change sequence. Then, for each production worker, the corresponding work posture data is identified to obtain a corresponding work posture change sequence. This work posture change sequence characterizes the changes in the individual production worker's work posture over time. Then, based on an intervention behavior recognition model, the posture change sequences are fine-grainedly classified to obtain an intervention behavior type sequence. This intervention behavior type sequence includes the inspector's intervention behavior and the time of occurrence. Finally, based on the work posture change sequence, abnormal work status events for each production worker are extracted. These abnormal work status events are used for... The process characterizes the transition of production workers' work status from an abnormal state to a normal state. Abnormal states include non-standard operation states and idle states. Then, for each production worker, the worker's workstation location is obtained, and based on the workstation location and the sequence of location changes, spatial proximity events between the inspector and the production worker are determined. These spatial proximity events characterize the inspector entering the production worker's preset proximity range. Then, based on the temporal correlation or overlap between the abnormal work status events and the spatial proximity events, a first intervention indicator for a single inspector on a single production worker is determined. Based on the causal relationship between the sequence of intervention behavior types and the abnormal work status events, a second intervention indicator for a single inspector on a single production worker is determined. The first intervention indicator includes the number of first interventions and the success rate of the first intervention, while the second intervention indicator includes the number of second interventions and the success rate of the second intervention. Finally, based on the first and second intervention indicators, the correlation characteristic values ​​between a single inspector and a single production worker are determined.

[0009] Optionally, the step of determining the correlation characteristic value between the individual inspection personnel and the individual production personnel based on the first intervention indicator and the second intervention indicator includes: The total number of interventions is determined based on the first number of interventions and the second number of interventions, and the overall intervention success rate is determined based on the first number of interventions, the first intervention success rate, the second number of interventions, and the second intervention success rate. The total number of interventions and the overall intervention success rate are weighted and fused to obtain the correlation feature value between the individual inspection personnel and the individual production personnel.

[0010] By adopting the above technical solution, in order to determine the correlation characteristic value between a single inspection personnel and a single production personnel, the total number of interventions is determined based on the first number of interventions and the second number of interventions, and the comprehensive intervention success rate is determined based on the first number of interventions, the first intervention success rate, the second number of interventions and the second intervention success rate. Then, the total number of interventions and the comprehensive intervention success rate are weighted and fused to obtain the correlation characteristic value between a single inspection personnel and a single production personnel.

[0011] Optionally, the step of performing matrix decomposition on the correlation matrix to extract the main feature directions includes: The correlation matrix is ​​centered to obtain the corresponding centered matrix; Calculate the covariance matrix of the centered matrix; The covariance matrix is ​​decomposed into eigenvalues ​​to obtain eigenvalues ​​and corresponding eigenvectors. The eigenvector corresponding to the maximum value among the eigenvalues ​​is taken as the main feature direction.

[0012] By adopting the above technical solution, in order to extract the main feature direction, the correlation matrix is ​​centered to obtain the corresponding centered matrix. Then, the covariance matrix of the centered matrix is ​​calculated, and the covariance matrix is ​​decomposed into eigenvalues ​​to obtain eigenvalues ​​and corresponding eigenvectors. Then, the eigenvector corresponding to the maximum value among the eigenvalues ​​is taken as the main feature direction.

[0013] Optionally, the step of performing matrix decomposition on the correlation matrix to extract the main feature directions includes: Perform singular value decomposition on the correlation matrix to obtain a singular value matrix and a right singular matrix; Extract the maximum singular value from the singular value matrix, and extract the corresponding target right singular feature vector in the right singular matrix based on the maximum singular value; The target right singular feature vector is used as the main feature direction.

[0014] By adopting the above technical solution, in order to extract the main feature direction, the correlation matrix is ​​decomposed into singular value matrix and right singular matrix. Then, the maximum singular value is extracted from the singular value matrix, and the target right singular feature vector corresponding to the maximum singular value is extracted from the right singular matrix. Then, the target right singular feature vector is used as the main feature direction.

[0015] Optionally, the step of determining the directional similarity and projection length based on the association vector and the main feature direction includes: The main feature directions are normalized to obtain the unit vectors corresponding to the main feature directions; Calculate the cosine value between the association vector and the unit vector, and perform mapping processing on the cosine value to obtain the corresponding directional similarity, wherein the directional similarity is used to characterize the directional consistency between the behavior pattern of the individual inspector and the optimal inspection pattern; Perform a dot product operation on the associated vector and the unit vector to obtain the projection length, wherein the projection length is used to represent the contribution intensity of a single inspector in the direction of the optimal inspection mode.

[0016] By adopting the above technical solution, in order to determine the directional similarity and projection length, the main feature directions are normalized to obtain the unit vectors corresponding to the main feature directions. Then, the cosine value between the associated vector and the unit vector is calculated, and the cosine value is mapped to obtain the corresponding directional similarity. The directional similarity is used to characterize the directional consistency between the behavior pattern of a single inspector and the optimal inspection pattern. Then, the dot product operation is performed on the associated vector and the unit vector to obtain the projection length. The projection length is used to represent the contribution intensity of a single inspector in the direction of the optimal inspection pattern.

[0017] Optionally, the step of determining the inspection quality indicators of each inspection personnel based on the directional similarity and the projection length includes: The directional similarity and the projection length are multiplied, and the product result is used as the comprehensive efficiency index of the inspection personnel. The comprehensive performance index of each inspection personnel is normalized, and the normalization result is mapped to a unified scale range to obtain inspection quality indicators that can be directly used for comparison.

[0018] By adopting the above technical solution, in order to determine the inspection quality indicators of each inspection personnel, the directional similarity and projection length are multiplied, and the product result is used as the comprehensive performance index of the inspection personnel. Then, the comprehensive performance index of each inspection personnel is normalized, and the normalization result is mapped to a unified scale range to obtain inspection quality indicators that can be directly used for comparison.

[0019] Secondly, this application also provides a production personnel management system based on the Industrial Internet of Things, which adopts the following technical solution: A production personnel management system based on the Industrial Internet of Things (IIoT) includes a management platform, a sensor network platform, and an object platform that are sequentially connected via communication. The management platform is configured with: The correlation feature value generation module is used to acquire the inspection data of each inspection personnel and the work posture data of each production personnel, and determine the correlation feature value between a single inspection personnel and a single production personnel based on the inspection data and the work posture data. The inspection data includes position data and action posture data, and the correlation feature value is used to characterize the influence effectiveness of the single inspection personnel on the work status of the single production personnel. The feature construction module is used to construct an association vector between a single inspection personnel and each production personnel based on the associated feature values, and to construct an association matrix between each inspection personnel and each production personnel based on the associated feature values. The feature decomposition module is used to perform matrix decomposition on the correlation matrix and extract the main feature directions; The inspection quality index generation module is used to determine the directional similarity and projection length of each associated vector based on the associated vector and the main feature direction, and to determine the inspection quality index of each inspection personnel based on the directional similarity and the projection length. The inspection personnel allocation module is used to allocate each inspection personnel to a region based on the inspection quality indicators, so as to realize the inspection of production personnel in each region.

[0020] Thirdly, this application also provides a computer device, which adopts the following technical solution: A computer device includes a memory and a processor, the memory storing a computer program executable on the processor, the processor executing the computer program to implement the method described in the first aspect.

[0021] Fourthly, this application also provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium storing a computer program capable of being loaded by a processor and executing the method described in the first aspect.

[0022] In summary, this application includes at least the following beneficial technical effects: First, it acquires the inspection data of each inspector and the work posture data of each production worker, and determines the correlation feature values ​​between a single inspector and a single production worker based on the inspection data and work posture data. The inspection data includes position data and action posture data, and the correlation feature values ​​characterize the effectiveness of a single inspector's influence on the work status of a single production worker. Then, it constructs correlation vectors between a single inspector and each production worker based on the correlation feature values, and constructs correlation matrices between each inspector and each production worker based on the correlation feature values. Next, it performs matrix decomposition on the correlation matrices to extract the main feature directions. Then, for each correlation vector, it determines the direction similarity and projection length based on the correlation vector and the main feature directions, and determines the inspection quality of each inspector based on the direction similarity and projection length. The method involves identifying and assigning inspection personnel to specific areas based on these indicators, enabling them to effectively inspect production workers within each area. This approach departs from the traditional production worker inspection management model, which relies on fixed routes and static experience. It represents a paradigm shift from "physical space coverage" to "behavioral efficiency optimization." By deeply integrating fragmented inspection data collected through the Industrial Internet of Things (IIoT) with production worker work data and conducting causal analysis, the method achieves a scientific and quantitative assessment of inspection personnel's work quality, quantifying the real impact of inspection behavior on production efficiency. Through this data-driven optimal inspection model, the method dynamically and accurately allocates inspection resources, placing high-performing inspection personnel in the most needed areas. This forms a complete closed loop from data perception and intelligent analysis to decision optimization, thereby improving inspection efficiency and enhancing the management efficiency of production workers. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the overall process of an embodiment of this application.

[0024] Figure 2 This is a structural diagram of one application scenario of the system in this application embodiment.

[0025] Figure 3 This is a structural diagram of another application scenario of the system according to an embodiment of this application.

[0026] Figure 4 This is a structural block diagram of the computer device described in this application. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0028] This application discloses a production personnel management method based on the Industrial Internet of Things.

[0029] Reference Figure 1 A production personnel management method based on the Industrial Internet of Things (IIoT) is applied to an IIoT system, which includes a management platform, a sensor network platform, and an object platform connected in sequence. The method is executed by the management platform and includes: Step S11: Obtain the inspection data of each inspection personnel and the work posture data of each production personnel, and determine the correlation feature value between a single inspection personnel and a single production personnel based on the inspection data and work posture data.

[0030] The inspection data includes location data and motion posture data, and the associated feature values ​​are used to characterize the effectiveness of a single inspection personnel in influencing the working status of a single production worker.

[0031] It should be noted that in step S11, the raw data (location, action, work posture) of inspection personnel and production personnel are obtained through industrial IoT devices, and the correlation feature value of a single inspection personnel to a single production personnel is calculated using event sequence analysis and causal inference models. This breaks through the bottleneck of the two types of data being separated in the existing technology, and transforms the raw and isolated data into indicators that can directly quantify the actual impact effectiveness of inspection behavior (i.e., how effective the inspection is).

[0032] Step S12: Construct the association vector between a single inspection personnel and each production personnel based on the association feature value, and construct the association matrix between each inspection personnel and each production personnel based on the association feature value.

[0033] It should be noted that in step S12, using all the associated feature values ​​calculated in step S11, an associated vector is constructed for each inspector (describing the comprehensive impact pattern on all production personnel), and the vectors of all inspectors are combined into an associated matrix. This systematically and digitally models the complex, grid-like "person-to-person" interaction relationship, transforming a management problem into a data object that can be mathematically analyzed and optimized.

[0034] Step S13: Perform matrix decomposition on the correlation matrix and extract the main feature directions.

[0035] It should be noted that in step S13, matrix decomposition (such as PCA or SVD) is performed on the correlation matrix to extract the main feature directions that can represent the main change patterns in the global data. This achieves dimensionality reduction and feature extraction of high-dimensional complex data, and the inherent laws of the "optimal inspection mode" are discovered in a data-driven manner. This replaces the traditional subjective setting that relies on fixed routes and historical experience, and provides an objective and optimal benchmark for evaluating inspection quality.

[0036] Step S14: For each associated vector, determine the directional similarity and projection length based on the associated vector and the main feature direction, and determine the inspection quality index of each inspector based on the directional similarity and projection length.

[0037] It should be noted that in step S14, by calculating the directional similarity (whether the behavior pattern is correct) and projection length (how strong the contribution) between the association vector of each inspector and the main feature direction, and combining the two to generate a comprehensive inspection quality index, a multi-dimensional and refined scientific quantification of the quality of inspection work is achieved. This not only focuses on the "quantity" of inspection, but also on its "quality," thereby solving the fundamental bottleneck of "inability to evaluate the inspection effect" and providing an accurate basis for management decisions.

[0038] Step S15: Assign each inspection personnel to a specific area based on the inspection quality indicators to enable inspection of production personnel in each area.

[0039] In the above implementation, firstly, the inspection data of each inspector and the work posture data of each production worker are acquired. Then, based on the inspection data and work posture data, the correlation feature values ​​between a single inspector and a single production worker are determined. The inspection data includes position data and action posture data. The correlation feature values ​​characterize the effectiveness of a single inspector's influence on the work status of a single production worker. Next, a correlation vector is constructed between a single inspector and each production worker based on the correlation feature values. Then, a correlation matrix is ​​constructed between each inspector and each production worker based on the correlation feature values. The correlation matrix is ​​then decomposed to extract the main feature directions. For each correlation vector, the direction similarity and projection length are determined based on the correlation vector and the main feature directions. Finally, the inspection quality indicators for each inspector are determined based on the direction similarity and projection length. Based on inspection quality indicators, each inspection personnel is assigned to a specific area to ensure inspections of production personnel within that area. This method changes the traditional production personnel inspection management model that relies on fixed routes and static experience, achieving a paradigm shift from "physical space coverage" to "behavioral efficiency optimization." By deeply integrating and causally analyzing fragmented inspection data collected by the Industrial Internet of Things with production personnel work data, a scientific and quantitative assessment of the work quality of inspection personnel is achieved, quantifying the real impact of inspection behavior on production efficiency. Through the data-driven discovery-based optimal inspection model of this invention, inspection resources are dynamically and accurately allocated, placing high-performing inspection personnel in the most needed areas. This forms a complete closed loop from data perception and intelligent analysis to decision optimization, thereby improving inspection efficiency and enhancing the management efficiency of production personnel.

[0040] As a further implementation of the method, the step of determining the correlation characteristic values ​​between a single inspection worker and a single production worker based on inspection data and work posture data includes: Step S21: For each inspector, identify the corresponding inspection data to obtain the corresponding inspection state change sequence. The inspection state change sequence is used to characterize the changes in the action posture and spatial position of a single inspector over time. The inspection state change sequence includes a position change sequence and an action posture change sequence.

[0041] Step S22: For each production worker, the corresponding work posture data is identified to obtain the corresponding work posture change sequence, wherein the work posture change sequence is used to characterize the change of a single production worker's work posture over time.

[0042] Step S23: Based on the intervention behavior recognition model, perform fine-grained classification of the action posture change sequence to obtain the intervention behavior type sequence, wherein the intervention behavior type sequence includes the intervention behavior of the inspection personnel and the time of occurrence of the behavior.

[0043] Step S24: Based on the work posture change sequence, extract the abnormal work status events of each production worker. The abnormal work status events are used to characterize the change of the production worker's work status from an abnormal state to a normal state. Abnormal states include non-standard operation state and leisure state.

[0044] Step S25: For each production worker, obtain the workstation location of the production worker, and determine the spatial proximity event between the inspection personnel and the production worker based on the workstation location and the location change sequence. The spatial proximity event is used to characterize the inspection personnel entering the preset proximity range of the production worker.

[0045] Step S26: Based on the temporal correlation or overlap between abnormal work status events and spatial proximity events, determine the first intervention indicator for a single inspector to a single production worker, and based on the causal relationship between the sequence of intervention behavior types and abnormal work status events, determine the second intervention indicator for a single inspector to a single production worker. The first intervention indicator includes the number of first interventions and the success rate of the first intervention, and the second intervention indicator includes the number of second interventions and the success rate of the second intervention.

[0046] It should be noted that in step S26, a dual verification mechanism is used to quantify the effectiveness of the inspection. First, it analyzes the temporal correlation between spatial proximity events (inspection personnel entering the vicinity of production personnel) and abnormal work status events (production personnel recovering from abnormal operation to normal operation). If the recovery of the production personnel's status closely follows (or overlaps with) the approach of the inspection personnel, it is recorded as an effective first intervention, and the number of first interventions and the success rate of the first intervention are calculated accordingly. This indicator reflects the potential deterrence and indirect influence of the inspection personnel's presence. Second, to provide more direct evidence, the method further analyzes whether the sequence of intervention behavior types (such as specific actions such as "gesture guidance" and "verbal communication" identified by the system) constitute a causal relationship with the abnormal work status events. When a clear intervention behavior is detected that directly leads to an immediate improvement in the production personnel's status, it is recorded as a more reliable second intervention, and the number of second interventions and the success rate of the second intervention are generated. This indicator directly captures the proactive and effective actions of the inspection personnel. By combining indirect indicators based on spatial correlation and direct causal indicators based on behavioral recognition, step S26 constructs a robust and multi-dimensional evaluation system, thereby achieving a scientific quantification of the actual impact effectiveness on inspection personnel, rather than just their physical existence.

[0047] Step S27: Based on the first intervention index and the second intervention index, determine the correlation characteristic value between a single inspection personnel and a single production personnel.

[0048] In the above implementation, to determine the correlation feature values ​​between a single inspector and a single production worker, for each inspector, the corresponding inspection data is identified to obtain a corresponding inspection state change sequence. This inspection state change sequence characterizes the changes in the individual inspector's posture and spatial position over time, and includes a position change sequence and a posture change sequence. Then, for each production worker, the corresponding work posture data is identified to obtain a corresponding work posture change sequence. This work posture change sequence characterizes the changes in the individual production worker's work posture over time. Then, based on an intervention behavior recognition model, the posture change sequences are fine-grainedly classified to obtain an intervention behavior type sequence. This intervention behavior type sequence includes the inspector's intervention behavior and the time of occurrence. Finally, based on the work posture change sequence, abnormal work status events for each production worker are extracted. These abnormal work status events are used to represent… The process involves identifying when a production worker's work status changes from an abnormal state to a normal state. Abnormal states include non-standard operating states and idle states. For each production worker, their workstation location is obtained. Based on the workstation location and the sequence of location changes, spatial proximity events between the inspector and the production worker are determined. These spatial proximity events characterize the inspector entering the production worker's pre-defined proximity range. Then, based on the temporal correlation or overlap between the abnormal work status events and the spatial proximity events, a first intervention indicator for a single inspector on a single production worker is determined. Based on the causal relationship between the sequence of intervention behavior types and the abnormal work status events, a second intervention indicator for a single inspector on a single production worker is determined. The first intervention indicator includes the number of first interventions and the success rate of the first intervention. The second intervention indicator includes the number of second interventions and the success rate of the second intervention. Finally, based on the first and second intervention indicators, the correlation characteristic values ​​between a single inspector and a single production worker are determined.

[0049] As a further implementation of the method, the step of determining the correlation characteristic values ​​between a single inspection personnel and a single production personnel based on a first intervention index and a second intervention index includes: Step S31: Determine the total number of interventions based on the first number of interventions and the second number of interventions, and determine the comprehensive intervention success rate based on the first number of interventions, the first intervention success rate, the second number of interventions, and the second intervention success rate.

[0050] Step S32: Weighted fusion of the total number of interventions and the overall intervention success rate to obtain the correlation feature value between a single inspection personnel and a single production personnel.

[0051] In the above implementation, in order to determine the correlation characteristic value between a single inspection personnel and a single production personnel, the total number of interventions is determined based on the first number of interventions and the second number of interventions, and the comprehensive intervention success rate is determined based on the first number of interventions, the first intervention success rate, the second number of interventions and the second intervention success rate. Then, the total number of interventions and the comprehensive intervention success rate are weighted and fused to obtain the correlation characteristic value between a single inspection personnel and a single production personnel.

[0052] As a further implementation of the method, the step of performing matrix decomposition on the correlation matrix and extracting the main feature directions includes: Step S41: The correlation matrix is ​​centered to obtain the corresponding centered matrix.

[0053] Step S42: Calculate the covariance matrix of the centered matrix.

[0054] Step S43: Perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors.

[0055] Step S44: Take the eigenvector corresponding to the maximum value among the eigenvalues ​​as the main feature direction.

[0056] It should be noted that, through steps S41 to S44, the optimal pattern that best distinguishes between high and low inspection efficiency is discovered from the behavioral data of all inspection personnel using a data-driven approach. The specific process is as follows: First, the correlation matrix is ​​centered by subtracting the average value of each dimension to eliminate the bias caused by differences in the basic status of different production personnel, so that subsequent analysis can focus on the changes brought about by inspection behavior rather than absolute values. Next, the covariance matrix of the centered matrix is ​​calculated. This matrix quantitatively describes the synergistic change relationship of different inspection personnel in their influence patterns on various production personnel, that is, whether their behavioral patterns are similar or opposite. Then, the covariance matrix is ​​decomposed into eigenvalues ​​to decouple multiple independent "feature directions" and their corresponding "eigenvalues" hidden in the data. Each feature direction represents a potential inspection pattern, and its eigenvalue quantifies the importance of the data variability that the pattern can explain. Finally, the eigenvector corresponding to the largest eigenvalue is selected as the main feature direction because this direction carries the most significant change information in the original data, and is thus mathematically defined as the "optimal benchmark" that best distinguishes between high-efficiency and low-efficiency inspections. Through this series of transformations, complex, high-dimensional interpersonal interaction data was successfully extracted into a core vector with clear statistical significance, representing the optimal inspection strategy.

[0057] In the above implementation, in order to extract the main feature direction, the correlation matrix is ​​centered to obtain the corresponding centered matrix. Then, the covariance matrix of the centered matrix is ​​calculated, and the covariance matrix is ​​decomposed into eigenvalues ​​to obtain eigenvalues ​​and corresponding eigenvectors. Then, the eigenvector corresponding to the maximum value among the eigenvalues ​​is taken as the main feature direction.

[0058] As a further implementation of the method, the step of performing matrix decomposition on the correlation matrix and extracting the main feature directions includes: Step S51: Perform singular value decomposition on the correlation matrix to obtain the singular value matrix and the right singular matrix.

[0059] Step S52: Extract the maximum singular value from the singular value matrix, and extract the corresponding target right singular eigenvector in the right singular matrix based on the maximum singular value.

[0060] Step S53: Take the target right singular feature vector as the main feature direction.

[0061] It should be noted that, through steps S51 to S54, singular value decomposition is directly performed on the constructed association matrix. This is a mathematical process that decomposes the original matrix into the product of three specific matrices, yielding a left singular matrix, a singular value matrix (a diagonal matrix whose diagonal elements are singular values), and a right singular matrix. The singular values ​​themselves have clear physical meaning: they are arranged in descending order, and each singular value quantifies the proportion of "energy" or "importance" of its corresponding "pattern" in the original data. The algorithm then extracts the largest singular value from the singular value matrix, as its corresponding pattern direction carries the most important structural information and degree of variation in the data. Next, based on the index of this largest singular value, the corresponding column is extracted from the right singular matrix, i.e., the target right singular eigenvector. This specific vector defines a new dimension (i.e., a latent factor) on which the projection distribution of all inspectors (i.e., the row vectors of the association matrix) differs most significantly, thus best distinguishing the behavioral patterns of high-efficiency and low-efficiency inspectors. Then, the right singular eigenvector of this target is established as the main feature direction. Using the SVD method, the system does not need to calculate the covariance matrix first and can directly process the original correlation matrix. It is not only computationally efficient and numerically stable, but more importantly, it successfully extracts a benchmark direction with clear mathematical meaning and business interpretability from complex human-to-human interaction data, representing the global optimal inspection strategy.

[0062] In the above implementation, in order to extract the main feature direction, the correlation matrix is ​​decomposed into singular value matrix and right singular matrix. Then, the maximum singular value is extracted from the singular value matrix, and the target right singular feature vector corresponding to the maximum singular value is extracted from the right singular matrix. Then, the target right singular feature vector is used as the main feature direction.

[0063] As a further implementation of the method, the step of determining the directional similarity and projection length based on the association vector and the main feature direction includes: Step S61: Normalize the main feature directions to obtain the unit vectors corresponding to the main feature directions.

[0064] Step S62: Calculate the cosine value between the associated vector and the unit vector, and perform mapping processing on the cosine value to obtain the corresponding directional similarity. The directional similarity is used to characterize the directional consistency between the behavior pattern of a single inspector and the optimal inspection pattern.

[0065] Step S63: Perform a dot product operation on the associated vector and the unit vector to obtain the projection length, where the projection length is used to represent the contribution intensity of a single inspector in the direction of the optimal inspection mode.

[0066] In the above implementation, in order to determine the directional similarity and projection length, the main feature directions are normalized to obtain the unit vectors corresponding to the main feature directions. Then, the cosine value between the associated vector and the unit vector is calculated, and the cosine value is mapped to obtain the corresponding directional similarity. The directional similarity is used to characterize the directional consistency between the behavior pattern of a single inspector and the optimal inspection pattern. Then, the dot product operation is performed on the associated vector and the unit vector to obtain the projection length. The projection length is used to represent the contribution intensity of a single inspector in the direction of the optimal inspection pattern.

[0067] As a further implementation of the method, the step of determining the inspection quality indicators of each inspector based on directional similarity and projection length includes: Step S71: Perform a product operation on the directional similarity and projection length, and use the product result as the comprehensive performance index of the inspection personnel.

[0068] Step S72: Normalize the comprehensive performance index of each inspection personnel and map the normalization result to a unified scale range to obtain inspection quality indicators that can be directly used for comparison.

[0069] In the above implementation, in order to determine the inspection quality index of each inspection personnel, the directional similarity and projection length are multiplied, and the product result is used as the comprehensive performance index of the inspection personnel. Then, the comprehensive performance index of each inspection personnel is normalized, and the normalization result is mapped to a unified scale range to obtain inspection quality indexes that can be directly used for comparison.

[0070] This application also discloses a production personnel management system based on the Industrial Internet of Things.

[0071] refer to Figure 2 A production personnel management system based on the Industrial Internet of Things (IIoT) includes a management platform, a sensor network platform, and an object platform that are sequentially connected in communication. The management platform is configured with: The correlation feature value generation module is used to acquire the inspection data of each inspection personnel and the work posture data of each production personnel, and to determine the correlation feature value between a single inspection personnel and a single production personnel based on the inspection data and work posture data. The inspection data includes position data and action posture data, and the correlation feature value is used to characterize the influence effectiveness of a single inspection personnel on the work status of a single production personnel. The feature construction module is used to construct the association vector between a single inspection personnel and each production personnel based on the associated feature values, and to construct the association matrix between each inspection personnel and each production personnel based on the associated feature values. The feature decomposition module is used to perform matrix decomposition on the correlation matrix and extract the main feature directions; The inspection quality index generation module is used to determine the directional similarity and projection length for each associated vector based on the associated vector and the main feature direction, and to determine the inspection quality index of each inspector based on the directional similarity and projection length. The inspection personnel allocation module is used to assign each inspection personnel to a specific area based on inspection quality indicators, so as to enable the inspection of production personnel in each area.

[0072] The overall framework of another application scenario of the production personnel management system based on the Industrial Internet of Things of this invention is as follows: Figure 3As shown, the system can include a user platform, a service platform, a management platform, a sensor network platform, and an object platform that interact sequentially, forming a five-platform architecture based on the Industrial Internet of Things (IIoT). The service platform consists of a main service database, multiple service sub-platforms, and multiple service sub-databases. The management platform includes a feature value generation module, a feature construction module, a feature decomposition module, an inspection quality index generation module, and an inspection personnel allocation module. The management platform can interact with the sensor network platform and the service platform. The sensor network platform can include a main sensor database, multiple sensor network sub-platforms, and multiple sensor sub-databases. In this embodiment, there are n sensor network sub-platforms and n sensor sub-databases. Each sensor network sub-platform has a corresponding sensor sub-database. The sensor network platform can interact with the object platform.

[0073] By leveraging the interaction between the various functional platforms of the industrial IoT-based production personnel management system, which is based on the aforementioned three or five platforms, a complete closed-loop information operation logic is established, ensuring the orderly operation of perceived and control information and realizing intelligent equipment management.

[0074] Specifically, the production personnel management system based on the Industrial Internet of Things in this embodiment includes a management platform. The management platform is configured to: acquire inspection data from each inspection personnel and work posture data from each production worker; determine the association feature values ​​between individual inspection personnel and individual production workers based on the inspection data and work posture data; wherein the inspection data includes location data and action posture data, and the association feature values ​​characterize the influence effectiveness of individual inspection personnel on the work status of individual production workers; construct association vectors corresponding to individual inspection personnel and each production worker based on the association feature values; construct association matrices corresponding to each inspection personnel and each production worker based on the association feature values; perform matrix decomposition on the association matrix to extract the main feature directions; for each association vector, determine the directional similarity and projection length based on the association vector and the main feature directions, and determine the inspection quality indicators for each inspection personnel based on the directional similarity and projection length; and allocate regions to each inspection personnel based on the inspection quality indicators to achieve inspection of production workers within each region.

[0075] The production personnel management system based on the Industrial Internet of Things of the present invention can implement any of the methods of production personnel management based on the Industrial Internet of Things, and the specific working process of the production personnel management system based on the Industrial Internet of Things of the present invention can refer to the corresponding process in the above-mentioned production personnel management methods based on the Industrial Internet of Things.

[0076] This application also discloses a computer device.

[0077] refer to Figure 3A computer device includes a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement any of the above-described methods for production personnel management based on the Industrial Internet of Things.

[0078] This application also discloses a computer-readable storage medium.

[0079] A computer-readable storage medium storing a computer program that can be loaded by a processor and executed by any of the above-described methods for production personnel management based on the Industrial Internet of Things.

[0080] The computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device; the program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0081] The above are all preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. A production personnel management method based on the Industrial Internet of Things, characterized in that, Applied to an industrial Internet of Things (IIoT) system, the IIoT system includes a management platform, a sensor network platform, and an object platform that are sequentially connected in communication. The method is executed by the management platform and includes: The inspection data of each inspection personnel and the work posture data of each production personnel are obtained, and the correlation feature value between a single inspection personnel and a single production personnel is determined based on the inspection data and the work posture data. The inspection data includes position data and action posture data, and the correlation feature value is used to characterize the influence effectiveness of the single inspection personnel on the work status of the single production personnel. Based on the associated feature values, construct an association vector between a single inspection personnel and each production personnel, and construct an association matrix between each inspection personnel and each production personnel based on the associated feature values. Perform matrix decomposition on the correlation matrix to extract the main feature directions; For each associated vector, the directional similarity and projection length are determined based on the associated vector and the main feature direction, and the inspection quality index of each inspection personnel is determined based on the directional similarity and the projection length. Based on the aforementioned inspection quality indicators, each inspection personnel is assigned to a specific area to enable them to conduct inspections of production personnel within that area.

2. The production personnel management method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of determining the correlation feature value between a single inspection worker and a single production worker based on the inspection data and the work posture data includes: For each inspector, the corresponding inspection data is identified to obtain the corresponding inspection state change sequence. The inspection state change sequence is used to characterize the changes in the action posture and spatial position of the individual inspector over time. The inspection state change sequence includes a position change sequence and an action posture change sequence. For each production worker, the corresponding work posture data is identified to obtain a corresponding work posture change sequence, wherein the work posture change sequence is used to characterize the change of the work posture of the individual production worker over time. Based on the intervention behavior recognition model, the sequence of action posture changes is classified in a fine-grained manner to obtain the sequence of intervention behavior types, wherein the sequence of intervention behavior types includes the intervention behavior of the inspection personnel and the time of occurrence of the behavior; Based on the work posture change sequence, abnormal work status events of each production worker are extracted. The abnormal work status events are used to characterize the change of the production worker's work status from an abnormal state to a normal state. The abnormal state includes non-standard operation state and leisure state. For each production worker, the workstation position of the production worker is obtained, and based on the workstation position and the position change sequence, a spatial proximity event between the inspection personnel and the production worker is determined, wherein the spatial proximity event is used to characterize the inspection personnel entering the preset proximity range of the production worker; Based on the temporal correlation or overlap between the abnormal work status event and the spatial proximity event, a first intervention indicator for the single inspection personnel on the single production personnel is determined, and based on the causal relationship between the sequence of intervention behavior types and the abnormal work status event, a second intervention indicator for the single inspection personnel on the single production personnel is determined. The first intervention indicator includes a first number of interventions and a first intervention success rate, and the second intervention indicator includes a second number of interventions and a second intervention success rate. Based on the first intervention indicator and the second intervention indicator, the correlation characteristic value between the individual inspection personnel and the individual production personnel is determined.

3. The production personnel management method based on the Industrial Internet of Things according to claim 2, characterized in that, The step of determining the correlation characteristic value between the individual inspection personnel and the individual production personnel based on the first intervention indicator and the second intervention indicator includes: The total number of interventions is determined based on the first number of interventions and the second number of interventions, and the overall intervention success rate is determined based on the first number of interventions, the first intervention success rate, the second number of interventions, and the second intervention success rate. The total number of interventions and the overall intervention success rate are weighted and fused to obtain the correlation feature value between the individual inspection personnel and the individual production personnel.

4. The production personnel management method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of performing matrix decomposition on the correlation matrix and extracting the main feature directions includes: The correlation matrix is ​​centered to obtain the corresponding centered matrix; Calculate the covariance matrix of the centered matrix; The covariance matrix is ​​decomposed into eigenvalues ​​to obtain eigenvalues ​​and corresponding eigenvectors. The eigenvector corresponding to the maximum value among the eigenvalues ​​is taken as the main feature direction.

5. The production personnel management method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of performing matrix decomposition on the correlation matrix and extracting the main feature directions includes: Perform singular value decomposition on the correlation matrix to obtain a singular value matrix and a right singular matrix; Extract the maximum singular value from the singular value matrix, and extract the corresponding target right singular feature vector in the right singular matrix based on the maximum singular value; The target right singular feature vector is used as the main feature direction.

6. The production personnel management method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of determining the directional similarity and projection length based on the association vector and the main feature direction includes: The main feature directions are normalized to obtain the unit vectors corresponding to the main feature directions; Calculate the cosine value between the association vector and the unit vector, and perform mapping processing on the cosine value to obtain the corresponding directional similarity, wherein the directional similarity is used to characterize the directional consistency between the behavior pattern of the individual inspector and the optimal inspection pattern; Perform a dot product operation on the associated vector and the unit vector to obtain the projection length, wherein the projection length is used to represent the contribution intensity of a single inspector in the direction of the optimal inspection mode.

7. The production personnel management method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of determining the inspection quality indicators of each inspection personnel based on the directional similarity and the projection length includes: The directional similarity and the projection length are multiplied, and the product result is used as the comprehensive efficiency index of the inspection personnel. The comprehensive performance index of each inspection personnel is normalized, and the normalization result is mapped to a unified scale range to obtain inspection quality indicators that can be directly used for comparison.

8. A production personnel management system based on the Industrial Internet of Things, characterized in that, It includes a management platform, a sensor network platform, and an object platform that are connected in sequence. The management platform is configured with: The correlation feature value generation module is used to acquire the inspection data of each inspection personnel and the work posture data of each production personnel, and determine the correlation feature value between a single inspection personnel and a single production personnel based on the inspection data and the work posture data. The inspection data includes position data and action posture data, and the correlation feature value is used to characterize the influence effectiveness of the single inspection personnel on the work status of the single production personnel. The feature construction module is used to construct an association vector between a single inspection personnel and each production personnel based on the associated feature values, and to construct an association matrix between each inspection personnel and each production personnel based on the associated feature values. The feature decomposition module is used to perform matrix decomposition on the correlation matrix and extract the main feature directions; The inspection quality index generation module is used to determine the directional similarity and projection length of each associated vector based on the associated vector and the main feature direction, and to determine the inspection quality index of each inspection personnel based on the directional similarity and the projection length. The inspection personnel allocation module is used to allocate each inspection personnel to a region based on the inspection quality indicators, so as to realize the inspection of production personnel in each region.

9. A computer device, characterized in that, The method includes a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer program is stored that can be loaded by a processor and execute the method of any one of claims 1 to 7.

Images