Enterprise intelligent man-machine cooperation method and system

Abstract

The invention relates to the technical field of man-machine cooperation, and aims to solve the problems that the traditional man-machine cooperation process mainly depends on a preset flow and timing rhythm control, and is difficult to dynamically capture personnel misoperation, machine response delay or cooperation rhythm imbalance, so that response strategy lagging and low regulation and control efficiency are caused in the face of complex and variable man-machine interaction situations. According to the enterprise intelligent man-machine cooperation method and system, through comprehensive analysis with behavior response offset, cooperation rhythm consistency and a personnel fatigue state as core indexes, a set of intelligent man-machine cooperation process which is distinct in hierarchy, logic closed-loop and sustainable iterative optimization is constructed; the real-time response capability of man-machine cooperation and the production takt time stability are improved, the cooperation efficiency fluctuation can be dynamically monitored, the precise regulation and control strategy is output in time, quantitative evaluation and actual improvement of man-machine cooperation efficiency are achieved, and efficient production of a production line and the quality level of products are guaranteed.

Description

Technical Field

[0001] This invention relates to the field of human-machine collaboration technology, specifically to an intelligent human-machine collaboration method and system for enterprises. Background Technology

[0002] With the continuous development of next-generation artificial intelligence and the industrial internet, enterprise-level intelligent manufacturing systems are accelerating into a stage of deep integration of human-machine collaboration. Building upon traditional automation, current industrial enterprises are placing greater emphasis on flexible production, personalized customization, and intelligent scheduling capabilities, thereby promoting the widespread application of intelligent human-machine collaboration technologies. This technology not only encompasses the interactive management between human operators and automated equipment but also extends to the comprehensive intelligent processing of operational behavior, real-time response control, operator fatigue perception, and efficiency evaluation. Especially in the fields of assembly manufacturing, precision testing, warehousing and logistics, and electronic packaging, intelligent human-machine collaboration systems possess the ability to adjust execution rhythm in real time, dynamically identify task deviations, and predict abnormal states, providing solid support for improving production safety, optimizing production line configuration, and reducing operating costs.

[0003] However, in existing human-machine collaboration practices, the lack of fine-grained temporal analysis of human-machine operation behavior and fatigue state perception mechanisms often makes it difficult to accurately identify abnormal interactions or rhythm deviations during the collaboration process. Firstly, traditional production lines mainly rely on preset processes and timed cycle control, making it difficult to dynamically capture human error, machine response delays, or collaborative rhythm imbalances, thus hindering the formation of adaptive control strategies. Furthermore, current practices generally rely on statistical averages or discrete time point behavior logs, neglecting the analysis of continuous differences between human operations and machine execution. Consequently, when facing complex and ever-changing human-machine interaction scenarios, response strategies lag behind, control efficiency is low, and production line efficiency and product quality are severely affected. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an intelligent human-machine collaboration method and system for enterprises, solving the problems mentioned in the background.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent human-machine collaboration method for enterprises, comprising the following steps:

[0006] S1. Divide the human-machine collaboration operation process into task units, and extract the timing information and duration information of human-machine collaboration behavior by combining the human-machine collaboration behavior log. By analyzing the degree of behavior response deviation and behavior collaboration deviation of each task unit, issue human-machine collaboration analysis instructions.

[0007] S2. After receiving the human-machine collaboration analysis instruction, analyze the consistency of the human-machine operation rhythm of each task unit during the human-machine collaboration operation based on the timing information of human-machine collaboration behavior. Analyze the degree of human-machine interaction interference of each task unit during the human-machine collaboration operation based on the human-machine collaboration behavior feature information recorded in the human-machine collaboration behavior log. Analyze the degree of personnel fatigue in each task unit during the human-machine collaboration operation through the deployed near-infrared camera and combined with the facial key point recognition algorithm.

[0008] S3. Based on the consistency of human-machine operation rhythm, the degree of human-machine interaction interference, and the degree of human fatigue, after comprehensive comparative analysis, a set of inefficient collaborative task units is selected, and corresponding machine collaboration control strategies are executed on the corresponding inefficient collaborative task units in the set of inefficient collaborative task units.

[0009] S4. After executing the corresponding machine collaboration control strategy, analyze the quality and production efficiency of the finished products produced during the human-machine collaboration operation after the control to determine the effectiveness of the executed machine collaboration control strategy, and execute the corresponding human collaboration control strategy for the corresponding low-efficiency collaboration task unit.

[0010] Preferably, step S1 specifically includes:

[0011] S11. Divide the task interaction units in the human-computer collaboration process. The specific division process includes:

[0012] During the human-machine collaborative operation, the operation process is divided into task units according to the standard task operation sequence, resulting in several task units. Combined with the human-machine collaborative behavior log, the timing information and duration information of the human-machine collaborative behavior are extracted. The timing information of the human-machine collaborative behavior includes the time point of human operation and the time point of machine execution for each task unit, and the duration information of the human-machine collaborative behavior includes the time of human operation and the time of machine execution for each task unit.

[0013] The standard task operation sequence includes handling tasks, sorting tasks, assembly tasks, packaging tasks, and inspection tasks. The human-machine collaboration behavior log is used to record the timestamp information and characteristic information of human-machine collaboration behavior during the human-machine collaboration operation process.

[0014] S12. Perform feature recognition on the timing information of human-machine collaborative behavior, calculate the difference in response time between the extracted human operation time point and machine execution time point of each task unit, and obtain the behavior response offset value of each task unit. Specifically: Behavior response offset value = (human operation time point - machine execution time point). Perform feature recognition on the duration information of human-machine collaborative behavior, calculate the difference in collaboration duration between the extracted human operation duration and machine execution duration of each task unit, and obtain the behavior collaboration offset value of each task unit. Specifically: Behavior collaboration offset value = (human operation duration - machine execution duration).

[0015] Preferably, step S1 further includes:

[0016] S13. Based on the magnitude of the behavioral response offset value and behavioral collaboration offset value of each task unit, determine whether the human-machine collaboration operation process is in an abnormal state, and issue a human-machine collaboration analysis command. The specific process includes:

[0017] If the behavior response offset value of the corresponding task unit is not equal to zero, a behavior response abnormal signal is issued; if the behavior response offset values ​​of the corresponding task units are all equal to zero, a behavior response normal signal is issued.

[0018] If the behavior collaboration offset value of the corresponding task unit is not equal to zero, a behavior collaboration abnormal signal is issued; if the behavior collaboration offset values ​​of the corresponding task units are all equal to zero, a behavior collaboration normal signal is issued.

[0019] If both a normal behavior response signal and a normal behavior collaboration signal are received simultaneously, it indicates that the human-machine collaboration operation process is in a normal state, and a normal human-machine collaboration instruction is issued, which means: continue to execute the subsequent human-machine collaboration operation process; otherwise, it indicates that the human-machine collaboration operation process is in an abnormal state, and a human-machine collaboration analysis instruction is issued.

[0020] Preferably, step S2 specifically includes:

[0021] S21. Upon receiving the human-machine collaboration analysis instruction, feature recognition is performed on the duration information of human-machine collaboration behavior. The extracted human operation duration and machine execution duration of the corresponding task units are correlated. After dimensionless processing, the consistency of human-machine operation rhythm in each task unit during the human-machine collaboration operation is analyzed, and the human-machine collaboration consistency coefficient of each task unit is determined. Specifically: In the formula, Let represent the human-machine collaboration consistency coefficient of the i-th task unit. and These represent the human operation time and machine execution time for the i-th task unit, respectively.

[0022] S22. Perform feature recognition on the human-machine collaboration behavior feature information recorded in the human-machine collaboration behavior log, and correlate the number of human error operations, total number of human-machine interactions, number of machine rollbacks due to errors, and total number of machine commands received in each task unit. After dimensionless processing, analyze the degree of human-machine interaction interference in each task unit during the human-machine collaboration operation, and determine the human-machine interaction interference coefficient of each task unit, specifically as follows: In the formula, This represents the human-computer interaction interference coefficient of the i-th task unit. , , and These represent the number of human error operations, the total number of human-machine interactions, the number of machine rollbacks due to errors, and the total number of machine commands received in the i-th task unit, respectively.

[0023] S23. By deploying near-infrared cameras and combining them with facial key point recognition algorithms, the eyelid opening and closing state and mouth shape state of personnel are monitored in real time. The total duration of eye closure and the number of yawns of personnel in each task unit during the human-machine collaborative operation are obtained. After normalization processing, the fatigue level of personnel in each task unit during the human-machine collaborative operation is analyzed, and the fatigue coefficient of personnel in each task unit is determined.

[0024] Preferably, step S3 specifically includes:

[0025] S31. The sum of the human-computer interaction interference coefficient and the personnel fatigue coefficient of the corresponding task unit is calculated by subtracting the human-computer collaboration consistency coefficient of the corresponding task unit to obtain the comprehensive collaboration efficiency value of each task unit. The average comprehensive collaboration efficiency is obtained by combining the statistical mean calculation algorithm.

[0026] S32. Compare and analyze the overall collaboration efficiency value of each task unit with the average overall collaboration efficiency to screen out the set of inefficient collaboration task units. The specific process includes:

[0027] If the overall collaboration efficiency value of the corresponding task unit exceeds the average overall collaboration efficiency, the corresponding task unit is marked as a high-efficiency collaboration task unit, and a set of high-efficiency collaboration task units is constructed. If the overall collaboration efficiency value of the corresponding task unit does not exceed the average overall collaboration efficiency, the corresponding task unit is marked as a low-efficiency collaboration task unit, and a set of low-efficiency collaboration task units is constructed. At the same time, the behavioral response offset value and behavioral collaboration offset value of each low-efficiency collaboration task unit in the low-efficiency collaboration task unit set are recorded.

[0028] Preferably, step S3 further includes:

[0029] S33. Execute the corresponding machine collaboration control strategy for the corresponding inefficient collaborative task unit. The specific execution process includes:

[0030] Based on the magnitude of the behavioral response offset value and behavioral collaboration offset value of the selected inefficient collaborative task unit, the corresponding machine collaboration control strategy is executed. If both the behavioral response offset value and the behavioral collaboration offset value of the inefficient collaborative task unit are greater than zero, the first machine collaboration control strategy is executed, which involves advancing the machine execution time of the inefficient collaborative task unit and increasing the machine execution duration. If both the behavioral response offset value and the behavioral collaboration offset value of the inefficient collaborative task unit are less than zero, the second machine collaboration control strategy is executed, which involves continuing to maintain the human-machine collaboration operation process of the inefficient collaborative task unit. Otherwise, the third machine collaboration control strategy is executed, which involves advancing the machine execution time of the inefficient collaborative task unit or increasing the machine execution duration.

[0031] Preferably, step S4 specifically includes:

[0032] S41. After executing the corresponding machine collaboration control strategy, the status of the finished products produced during the human-machine collaboration operation under control within a unit cycle is monitored in real time, and the finished product data information after control is obtained. The finished product data information after control includes the finished product production quantity and yield rate within a unit cycle.

[0033] S42. Perform feature identification on the finished product data information after regulation, correlate the extracted finished product production quantity and yield rate within a unit cycle, and after dimensionless processing, analyze the quality and production efficiency level of the finished products produced during the human-machine collaborative operation after regulation within a unit cycle, and determine the optimal production coefficient after regulation, specifically: xyc=a1×nsc+a2×vop; where xyc represents the optimal production coefficient after regulation, nsc and vop represent the finished product production quantity and yield rate within a unit cycle, respectively, and a1 and a2 represent the weight values ​​of the finished product production quantity and yield rate, respectively.

[0034] Preferably, step S4 further includes:

[0035] S43. Before implementing the corresponding machine collaboration control strategy, the quality and production efficiency of the finished products produced during the human-machine collaboration operation before the control are analyzed in the historical unit cycle to determine the pre-control yield coefficient, which is then used as the yield comparison threshold.

[0036] S44. Compare and analyze the adjusted yield coefficient with the yield comparison threshold to determine whether the currently implemented machine collaboration control strategy is effective, and implement the corresponding human collaboration control strategy for the corresponding inefficient collaboration task units. The specific process includes:

[0037] If the yield coefficient after adjustment exceeds the yield comparison threshold, it indicates that the currently implemented machine collaboration adjustment strategy is effective, and the first machine collaboration adjustment strategy is generated, which is: continue to maintain the human-machine collaboration operation process after adjustment.

[0038] If the yield coefficient after adjustment does not exceed the yield comparison threshold, it indicates that the currently implemented machine collaboration control strategy is invalid. A second machine collaboration control strategy is then generated, which is to arrange rest and personnel replacement for the operators in the corresponding inefficient collaboration task units.

[0039] An intelligent human-machine collaboration system for enterprises includes an early warning analysis module, a human-machine collaboration analysis module, a machine control module, and a personnel control module;

[0040] The early warning analysis module is used to divide the human-machine collaboration operation process into task units, and extract the timing information and duration information of human-machine collaboration behavior by combining human-machine collaboration behavior logs. By analyzing the degree of behavioral response deviation and behavioral collaboration deviation of each task unit, it issues human-machine collaboration analysis instructions.

[0041] The human-machine collaboration analysis module is used to analyze the consistency of human-machine operation rhythm of each task unit during human-machine collaboration operation based on the timing information of human-machine collaboration behavior after receiving the human-machine collaboration analysis instruction, analyze the degree of human-machine interaction interference of each task unit during human-machine collaboration operation based on the human-machine collaboration behavior feature information recorded in the human-machine collaboration behavior log, and analyze the degree of human-machine interaction interference of each task unit during human-machine collaboration operation through the deployed near-infrared camera and combined with the facial key point recognition algorithm.

[0042] The machine control module is used to screen out the set of inefficient collaborative task units after comprehensive comparative analysis based on the consistency of human-machine operation rhythm, the degree of human-machine interaction interference, and the degree of human fatigue, and to execute the corresponding machine collaboration control strategy on the corresponding inefficient collaborative task units in the set of inefficient collaborative task units.

[0043] The personnel control module is used to determine the effectiveness of the machine collaboration control strategy by analyzing the quality and production efficiency of the finished products produced during the human-machine collaboration operation after the corresponding machine collaboration control strategy is executed, and to execute the corresponding personnel collaboration control strategy for the corresponding low-efficiency collaboration task unit.

[0044] This invention provides an intelligent human-machine collaboration method and system for enterprises, which has the following beneficial effects:

[0045] (1) The present invention provides an intelligent human-machine collaboration method and system for enterprises. In view of the problems that the existing human-machine collaboration mechanism in the intelligent manufacturing environment is difficult to dynamically identify anomalies, difficult to accurately perceive fatigue state and lack effective control strategies, a comprehensive analysis model with behavioral response deviation, collaboration rhythm consistency and personnel fatigue state as core indicators is established. Combined with standardized task unit division, human-machine behavior log extraction, infrared camera tracking and multi-parameter feature quantification, a set of intelligent collaboration process with clear hierarchy, logical closed loop and sustainable iterative optimization is constructed. Through deployment in the actual production environment, it can not only significantly improve the real-time response capability of human-machine cooperation and improve the stability of production rhythm, but also dynamically monitor the fluctuation of collaboration efficiency and output accurate control suggestions in a timely manner, and finally realize the quantitative evaluation and effective improvement of human-machine collaboration efficiency, and ensure production line safety and finished product stability.

[0046] (2) By introducing a mechanism for extracting behavioral response offset and collaboration offset values ​​during the task unit division stage, the technical shortcomings of traditional production lines that rely solely on timing beats or average data to detect collaboration imbalances in real time are effectively addressed. By analyzing the difference between the time points of personnel operation and the time points of machine execution, a fine perception of operational lag or machine response delay behavior is achieved. At the same time, by comparing and analyzing the time of personnel operation and machine execution in each task unit, the implicit offset trend in continuous collaboration is further revealed. This anomaly identification logic based on fine-grained time series not only improves the observability of the human-machine collaboration process, but also issues anomaly warning commands in the early stage of collaboration imbalance, providing a basis for subsequent multi-dimensional analysis and control mechanisms, and enhancing the enterprise's early response capability to rhythm deviation and misoperation anomalies in high-frequency operation scenarios.

[0047] (3) To address the lack of fatigue monitoring and multi-factor interference identification in traditional methods, a three-dimensional collaborative efficiency index system was designed, covering three key parameters: human-machine collaboration consistency coefficient, human-machine interaction interference coefficient, and personnel fatigue coefficient. By deploying near-infrared cameras and combining them with facial key point recognition algorithms, the duration of eye closure and the number of yawns were collected in real time to quantify the operator's fatigue level. At the same time, based on the human-machine interaction behavior log, the number of misoperations, machine rollback records, and instruction response frequency were statistically analyzed to accurately reflect the degree of interaction interference. Combined with the comparison of the operation rhythm of personnel and equipment, a high-dimensional collaborative data matrix that can be used for horizontal comparison and vertical trend analysis was formed. This multi-dimensional analysis mechanism not only improves the accuracy and robustness of human-machine state identification, but also enables abnormal collaboration diagnosis to leap from a single indicator to a comprehensive analysis stage, providing quantitative support for machine and personnel control strategies.

[0048] (4) To address the problems of lagging and poor strategy adaptability of existing collaborative control mechanisms, a two-way control mechanism of response and prediction is constructed based on the behavioral offset and efficiency evaluation value output by the analysis model. First, according to the comparison between the collaborative efficiency value and the mean, efficient and inefficient task units are accurately divided, and the behavioral response offset and collaborative offset are further associated for inefficient units to achieve refined adaptation of control strategies. For those with obvious offset trends, the first strategy of starting machine operation in advance and adjusting the duration is implemented. For those with insignificant fluctuations, the strategy of maintaining is adopted. For situations with uneven parameter distribution, the third strategy of adjusting machine time points or fine-tuning execution duration is introduced. At the same time, the effectiveness of the control strategy can be judged by combining the changing trends of finished product production quantity and yield rate within a unit cycle. If the optimization range does not meet the standard, the strategy of personnel replacement is automatically switched to achieve comprehensive optimization of the human-machine cooperation structure. This control process has obvious adaptive and closed-loop characteristics, which significantly improves the flexibility of human-machine collaboration configuration and production consistency. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of a human-machine collaboration method for enterprises according to the present invention;

[0050] Figure 2 This is a block diagram of an intelligent human-machine collaboration system for enterprises according to the present invention;

[0051] Figure 3 This is a logic diagram of an intelligent human-machine collaboration method for enterprises according to the present invention. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Example 1

[0054] Please see Figure 1 and Figure 3 This invention provides an intelligent human-machine collaboration method for enterprises, comprising the following steps:

[0055] S1. Divide the human-machine collaboration operation process into task units, and extract the timing information and duration information of human-machine collaboration behavior by combining the human-machine collaboration behavior log. By analyzing the degree of behavior response deviation and behavior collaboration deviation of each task unit, issue human-machine collaboration analysis instructions.

[0056] Specifically, the steps in S1 include:

[0057] S11. Divide the task interaction units in the human-computer collaboration process. The specific division process includes:

[0058] During the human-machine collaborative operation, the operation process is divided into task units according to the standard task operation sequence, resulting in several task units. Combined with the human-machine collaborative behavior log, the timing information and duration information of the human-machine collaborative behavior are extracted. The timing information of the human-machine collaborative behavior includes the time point of human operation and the time point of machine execution for each task unit, and the duration information of the human-machine collaborative behavior includes the time of human operation and the time of machine execution for each task unit.

[0059] It should be noted that the human operation time point and machine execution time point refer to the specific time nodes in each task unit when a person begins to operate a certain task and the corresponding automated machine starts to execute the task. They are usually recorded in the form of timestamps to reflect the sequence and response relationship of human and machine operations. The human operation duration and machine execution duration respectively represent the duration from the start to the end of the human operation and the duration from the machine receiving the instruction to completing the corresponding execution action, reflecting the continuous collaboration cycle in the task execution process. This timing and duration information not only reveals the degree of rhythm matching between humans and machines in the operation process, but also provides a basic data source for the calculation of subsequent behavior response offset values ​​and collaboration duration offset values, thereby helping to identify potential problems such as asynchronous rhythm, collaboration delays, and operational conflicts.

[0060] The standard task operation sequence includes handling, sorting, assembly, packaging, and inspection tasks. The human-machine collaboration behavior log records the timestamps and characteristics of human-machine collaboration behaviors during the operation. Its core function is to provide detailed and traceable human-machine interaction behavior trajectories and operational characteristic data, providing fundamental information support for subsequent human-machine collaboration efficiency analysis, anomaly detection, and control strategy formulation. This log is typically recorded by sensors in an embedded data acquisition unit, specifically including the start time, end time, operation type, and duration of the operation at each task node, as well as key event node information such as the instruction reception time, execution time, and response delay of the corresponding automated equipment. Furthermore, the behavior log can integrate signals from video surveillance, trigger-based sensors, and equipment control systems to automatically mark erroneous operations, collaboration interruptions, and instruction rollbacks. By standardizing and extracting features from this log data, a precise temporal collaboration map can be constructed based on task units, thereby supporting multi-dimensional modeling and evaluation of human-machine behavior rhythm deviations, interaction interference, and collaboration efficiency.

[0061] S12. Perform feature recognition on the timing information of human-machine collaborative behavior, calculate the difference in response time between the extracted human operation time point and machine execution time point of each task unit, and obtain the behavior response offset value of each task unit. Specifically: Behavior response offset value = (human operation time point - machine execution time point). Perform feature recognition on the duration information of human-machine collaborative behavior, calculate the difference in collaboration duration between the extracted human operation duration and machine execution duration of each task unit, and obtain the behavior collaboration offset value of each task unit. Specifically: Behavior collaboration offset value = (human operation duration - machine execution duration).

[0062] Specifically, the S1 steps also include:

[0063] S13. Based on the magnitude of the behavioral response offset value and behavioral collaboration offset value of each task unit, determine whether the human-machine collaboration operation process is in an abnormal state, and issue a human-machine collaboration analysis command. The specific process includes:

[0064] If the behavior response offset value of the corresponding task unit is not equal to zero, a behavior response abnormal signal is issued; if the behavior response offset values ​​of the corresponding task units are all equal to zero, a behavior response normal signal is issued.

[0065] If the behavior collaboration offset value of the corresponding task unit is not equal to zero, a behavior collaboration abnormal signal is issued; if the behavior collaboration offset values ​​of the corresponding task units are all equal to zero, a behavior collaboration normal signal is issued.

[0066] If both a normal behavior response signal and a normal behavior collaboration signal are received simultaneously, it indicates that the human-machine collaboration operation process is in a normal state, and a normal human-machine collaboration instruction is issued, which means: continue to execute the subsequent human-machine collaboration operation process; otherwise, it indicates that the human-machine collaboration operation process is in an abnormal state, and a human-machine collaboration analysis instruction is issued.

[0067] For example, in a certain task unit, the human operation starts at 08:15:32.500, and the machine execution starts at 08:15:33.200. The behavioral response offset value of this task unit is -0.700 seconds, indicating that the human operation precedes the machine response by 0.7 seconds. Furthermore, the human operation duration is 12.6 seconds, and the machine execution duration is 14.1 seconds. The behavioral collaboration offset value of this task unit is -1.5 seconds, reflecting that the machine execution lasts longer than the human operation. Based on the fact that both offset values ​​are not zero, the system automatically identifies the problem of human-machine rhythm difference and execution duration imbalance in this task unit, thereby issuing behavioral response abnormality signals and behavioral collaboration abnormality signals, and subsequently triggering human-machine collaboration analysis instructions.

[0068] S2. After receiving the human-machine collaboration analysis instruction, analyze the consistency of the human-machine operation rhythm of each task unit during the human-machine collaboration operation based on the timing information of human-machine collaboration behavior. Analyze the degree of human-machine interaction interference of each task unit during the human-machine collaboration operation based on the human-machine collaboration behavior feature information recorded in the human-machine collaboration behavior log. Analyze the degree of personnel fatigue in each task unit during the human-machine collaboration operation through the deployed near-infrared camera and combined with the facial key point recognition algorithm.

[0069] Specifically, the steps in S2 include:

[0070] S21. Upon receiving the human-machine collaboration analysis instruction, feature recognition is performed on the duration information of human-machine collaboration behavior. The extracted human operation duration and machine execution duration of the corresponding task units are correlated. After dimensionless processing, the consistency of human-machine operation rhythm in each task unit during the human-machine collaboration operation is analyzed, and the human-machine collaboration consistency coefficient of each task unit is determined. Specifically: In the formula, Let represent the human-machine collaboration consistency coefficient of the i-th task unit. and These represent the human operation time and machine execution time for the i-th task unit, respectively.

[0071] S22. Perform feature recognition on the human-machine collaboration behavior feature information recorded in the human-machine collaboration behavior log, and correlate the number of human error operations, total number of human-machine interactions, number of machine rollbacks due to errors, and total number of machine commands received in each task unit. After dimensionless processing, analyze the degree of human-machine interaction interference in each task unit during the human-machine collaboration operation, and determine the human-machine interaction interference coefficient of each task unit, specifically as follows: In the formula, This represents the human-computer interaction interference coefficient of the i-th task unit. , , and Let represent the number of human error operations, the total number of human-machine interactions, the number of machine rollbacks due to errors, and the total number of machine commands received in the i-th task unit, respectively. This indicates the degree of interference of the human-computer interaction process caused by the human error in the i-th task unit. This indicates the degree of interference of the machine rollback operation caused by human error in the i-th task unit on the human-computer interaction process;

[0072] It should be noted that the number of human error operations refers to the cumulative number of times that personnel fail to perform tasks according to preset procedures or steps during task execution, resulting in task interruption, accidental triggering, and execution failure, reflecting the accuracy and stability of personnel operations; the total number of human-machine interactions refers to the total number of all interactive behaviors between personnel and equipment in a task unit, including instruction input, feedback confirmation, and prompt response events, which is an important indicator for measuring the frequency of human-machine interaction and the complexity of collaboration; the number of machine rollbacks due to human error indicates the number of times the machine enters a rollback process or abnormal reset state due to human error, revealing the negative impact of human interference on the normal operation of the machine; the total number of machine instructions received refers to the total number of all instructions received by the equipment in the task unit from personnel input or system scheduling, which is a basic parameter for evaluating interaction density and instruction processing capabilities.

[0073] S23. By deploying near-infrared cameras and combining them with facial key point recognition algorithms, the eyelid opening and closing status and lip shape status of personnel are monitored in real time. The total duration of eye closure and the number of yawns for personnel in each task unit during the human-machine collaborative operation are obtained. After normalization processing and combining with a linear weighted algorithm, the fatigue level of personnel in each task unit during the human-machine collaborative operation is analyzed, and the fatigue coefficient of personnel in each task unit is determined. Specifically: xpp i =b1×tby i +b2×nhq i In the formula, xpp i tby represents the fatigue coefficient of personnel in the i-th task unit. i and nhq i Let b1 and b2 represent the total duration of eye closure and the number of yawns of the personnel in the i-th task unit, respectively. The specific values ​​are set by the personnel in this technical field.

[0074] S3. Based on the consistency of human-machine operation rhythm, the degree of human-machine interaction interference, and the degree of human fatigue, after comprehensive comparative analysis, a set of inefficient collaborative task units is selected, and corresponding machine collaboration control strategies are executed on the corresponding inefficient collaborative task units in the set of inefficient collaborative task units.

[0075] Specifically, the steps in S3 include:

[0076] S31. The sum of the human-computer interaction interference coefficient and the personnel fatigue coefficient of the corresponding task unit is calculated by subtracting the human-computer collaboration consistency coefficient of the corresponding task unit to obtain the comprehensive collaboration efficiency value of each task unit. The average comprehensive collaboration efficiency is obtained by combining the statistical mean calculation algorithm.

[0077] S32. Compare and analyze the overall collaboration efficiency value of each task unit with the average overall collaboration efficiency to screen out the set of inefficient collaboration task units. The specific process includes:

[0078] If the overall collaboration efficiency value of the corresponding task unit exceeds the average overall collaboration efficiency, the corresponding task unit is marked as a high-efficiency collaboration task unit, and a set of high-efficiency collaboration task units is constructed. If the overall collaboration efficiency value of the corresponding task unit does not exceed the average overall collaboration efficiency, the corresponding task unit is marked as a low-efficiency collaboration task unit, and a set of low-efficiency collaboration task units is constructed. At the same time, the behavioral response offset value and behavioral collaboration offset value of each low-efficiency collaboration task unit in the low-efficiency collaboration task unit set are recorded.

[0079] Specifically, the S3 steps also include:

[0080] S33. Execute the corresponding machine collaboration control strategy for the corresponding inefficient collaborative task unit. The specific execution process includes:

[0081] Based on the magnitude of the behavioral response offset value and behavioral collaboration offset value of the selected inefficient collaborative task unit, the corresponding machine collaboration control strategy is executed. If both the behavioral response offset value and the behavioral collaboration offset value of the inefficient collaborative task unit are greater than zero, the first machine collaboration control strategy is executed, which involves advancing the machine execution time of the inefficient collaborative task unit and increasing the machine execution duration. If both the behavioral response offset value and the behavioral collaboration offset value of the inefficient collaborative task unit are less than zero, the second machine collaboration control strategy is executed, which involves continuing to maintain the human-machine collaborative operation process of the inefficient collaborative task unit. Conversely, if the behavioral response offset value of the inefficient collaborative task unit is greater than zero and the behavioral collaboration offset value is less than zero, or if the behavioral response offset value of the inefficient collaborative task unit is less than zero and the behavioral collaboration offset value is greater than zero, the third machine collaboration control strategy is executed, which involves advancing the machine execution time of the inefficient collaborative task unit or increasing the machine execution duration.

[0082] S4. After executing the corresponding machine collaboration control strategy, analyze the quality and production efficiency of the finished products produced during the human-machine collaboration operation after the control to determine the effectiveness of the executed machine collaboration control strategy, and execute the corresponding human collaboration control strategy for the corresponding low-efficiency collaboration task unit.

[0083] Specifically, the steps in S4 include:

[0084] S41. After executing the corresponding machine collaboration control strategy, the status of the finished products produced during the human-machine collaboration operation under control within a unit cycle is monitored in real time, and the finished product data information after control is obtained. The finished product data information after control includes the finished product production quantity and yield rate within a unit cycle.

[0085] It should be noted that the finished product production quantity refers to the total number of products completed in a unit cycle through human-machine collaborative processes after adjustment, reflecting the operating rhythm and output capacity of the production line; the yield rate is the percentage of qualified products that meet quality standards and do not require rework or rejection among all finished products in the unit cycle, usually expressed as a percentage.

[0086] S42. Feature recognition is performed on the finished product data information after regulation. The extracted finished product production quantity and yield rate within a unit cycle are correlated. After dimensionless processing, the quality and production efficiency level of the finished products produced during the human-machine collaborative operation after regulation within a unit cycle are analyzed to determine the optimal production coefficient after regulation. Specifically: xyc=a1×nsc+a2×vop; where xyc represents the optimal production coefficient after regulation, nsc and vop represent the finished product production quantity and yield rate within a unit cycle, respectively, and a1 and a2 represent the weight values ​​of the finished product production quantity and yield rate, respectively. The specific values ​​are set by the personnel in this technical field.

[0087] Specifically, the S4 steps also include:

[0088] S43. Before implementing the corresponding machine collaboration control strategy, the quality and production efficiency of the finished products produced during the human-machine collaboration operation before the control are analyzed in the historical unit cycle to determine the pre-control yield coefficient, which is then used as the yield comparison threshold.

[0089] It should be noted that the pre-control yield coefficient is obtained by analyzing the historical finished product production quantity and historical yield rate collected in the historical unit cycle, and combined with the post-control yield coefficient calculation process. It represents the basic output capacity and quality stability performance of the human-machine collaboration system under the pre-control state.

[0090] S44. Compare and analyze the adjusted yield coefficient with the yield comparison threshold to determine whether the currently implemented machine collaboration control strategy is effective, and implement the corresponding human collaboration control strategy for the corresponding inefficient collaboration task units. The specific process includes:

[0091] If the yield coefficient after adjustment exceeds the yield comparison threshold, it indicates that the currently implemented machine collaboration adjustment strategy is effective, and the first machine collaboration adjustment strategy is generated, which is: continue to maintain the human-machine collaboration operation process after adjustment.

[0092] If the yield coefficient after adjustment does not exceed the yield comparison threshold, it indicates that the currently implemented machine collaboration control strategy is invalid. A second machine collaboration control strategy is then generated, which is to arrange rest and personnel replacement for the operators in the corresponding inefficient collaboration task units.

[0093] In this embodiment, the operation process of human-machine collaboration is divided into task units, and human-machine collaboration behavior logs are combined to extract the timing information and duration information of human-machine collaboration behavior. By analyzing the degree of behavioral response deviation and behavioral collaboration deviation of each task unit, human-machine collaboration analysis instructions are issued, which can realize early automatic identification and analysis instruction triggering of abnormal collaboration status, and improve the sensitivity and response speed of abnormality perception.

[0094] After receiving the human-machine collaboration analysis command, the system analyzes the consistency of human-machine operation rhythm in each task unit during the human-machine collaboration operation based on the timing information of human-machine collaboration behavior. Based on the human-machine collaboration behavior feature information recorded in the human-machine collaboration behavior log, the system analyzes the degree of human-machine interaction interference in each task unit during the human-machine collaboration operation. Through the deployment of near-infrared cameras and the combination of facial key point recognition algorithms, the system analyzes the degree of human fatigue in each task unit during the human-machine collaboration operation. Through multi-dimensional in-depth analysis of rhythm consistency, interaction interference, and human fatigue, the system achieves accurate modeling and comprehensive diagnosis of human-machine status, providing a basis for subsequent control strategies.

[0095] Based on the consistency of human-machine operation rhythm, the degree of human-machine interaction interference, and the degree of human fatigue, after comprehensive comparative analysis, a set of inefficient collaborative task units is selected. Corresponding machine collaboration control strategies are then implemented for the corresponding inefficient collaborative task units in the set. By comprehensively judging various indicators of human-machine collaboration efficiency, inefficient tasks are accurately selected and targeted machine control strategies are implemented to effectively optimize the machine execution rhythm and task matching degree, thereby improving overall collaboration efficiency.

[0096] After implementing the corresponding machine collaboration control strategy, the effectiveness of the implemented machine collaboration control strategy is judged by analyzing the quality and production efficiency of the finished products produced during the human-machine collaboration operation after the control. The corresponding human collaboration control strategy is then implemented for the corresponding low-efficiency collaboration task units. This process involves feedback analysis of the output data after the control to judge the effect of the strategy and guide personnel to be switched or rested, so as to ensure continuous optimization of the collaboration process and stable and controllable production quality.

[0097] Example 2

[0098] Please refer to Figure 1 and Figure 2 Specifically: an intelligent human-machine collaboration system for enterprises, including an early warning analysis module, a human-machine collaboration analysis module, a machine control module, and a personnel control module;

[0099] The early warning analysis module is used to divide the human-machine collaboration operation process into task units, and extract the timing information and duration information of human-machine collaboration behavior by combining human-machine collaboration behavior logs. By analyzing the degree of behavioral response deviation and behavioral collaboration deviation of each task unit, it issues human-machine collaboration analysis instructions.

[0100] The human-machine collaboration analysis module is used to analyze the consistency of human-machine operation rhythm of each task unit during human-machine collaboration operation based on the timing information of human-machine collaboration behavior after receiving the human-machine collaboration analysis instruction, analyze the degree of human-machine interaction interference of each task unit during human-machine collaboration operation based on the human-machine collaboration behavior feature information recorded in the human-machine collaboration behavior log, and analyze the degree of human-machine interaction interference of each task unit during human-machine collaboration operation through the deployed near-infrared camera and combined with the facial key point recognition algorithm.

[0101] The machine control module is used to screen out the set of inefficient collaborative task units after comprehensive comparative analysis based on the consistency of human-machine operation rhythm, the degree of human-machine interaction interference, and the degree of human fatigue, and to execute the corresponding machine collaboration control strategy on the corresponding inefficient collaborative task units in the set of inefficient collaborative task units.

[0102] The personnel control module is used to determine the effectiveness of the machine collaboration control strategy by analyzing the quality and production efficiency of the finished products produced during the human-machine collaboration operation after the corresponding machine collaboration control strategy is executed, and to execute the corresponding personnel collaboration control strategy for the corresponding low-efficiency collaboration task unit.

[0103] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent human-machine collaboration in enterprises, characterized in that: Includes the following steps: S1. Divide the human-machine collaboration operation process into task units, and extract the timing information and duration information of human-machine collaboration behavior by combining the human-machine collaboration behavior log. By analyzing the degree of behavior response deviation and behavior collaboration deviation of each task unit, issue human-machine collaboration analysis instructions. S2. After receiving the human-machine collaboration analysis instruction, analyze the consistency of the human-machine operation rhythm of each task unit during the human-machine collaboration operation based on the timing information of human-machine collaboration behavior. Analyze the degree of human-machine interaction interference of each task unit during the human-machine collaboration operation based on the human-machine collaboration behavior feature information recorded in the human-machine collaboration behavior log. Analyze the degree of personnel fatigue in each task unit during the human-machine collaboration operation through the deployed near-infrared camera and combined with the facial key point recognition algorithm. S3. Based on the consistency of human-machine operation rhythm, the degree of human-machine interaction interference, and the degree of human fatigue, after comprehensive comparative analysis, a set of inefficient collaborative task units is selected, and corresponding machine collaboration control strategies are executed on the corresponding inefficient collaborative task units in the set of inefficient collaborative task units. S4. After executing the corresponding machine collaboration control strategy, analyze the quality and production efficiency of the finished products produced during the human-machine collaboration operation after the control to determine the effectiveness of the executed machine collaboration control strategy, and execute the corresponding human collaboration control strategy for the corresponding low-efficiency collaboration task unit.

2. The intelligent human-machine collaboration method for enterprises according to claim 1, characterized in that: The specific steps in S1 include: S11. Divide the task interaction units in the human-computer collaboration process. The specific division process includes: During the human-machine collaborative operation, the operation process is divided into task units according to the standard task operation sequence, resulting in several task units. Combined with the human-machine collaborative behavior log, the timing information and duration information of the human-machine collaborative behavior are extracted. The timing information of the human-machine collaborative behavior includes the time point of human operation and the time point of machine execution for each task unit, and the duration information of the human-machine collaborative behavior includes the time of human operation and the time of machine execution for each task unit. The standard task operation sequence includes handling tasks, sorting tasks, assembly tasks, packaging tasks, and inspection tasks. The human-machine collaboration behavior log is used to record the timestamp information and characteristic information of human-machine collaboration behavior during the human-machine collaboration operation process. S12. Perform feature recognition on the timing information of human-machine collaborative behavior, calculate the difference in response time between the extracted human operation time point and machine execution time point of each task unit, and obtain the behavior response offset value of each task unit. Specifically: Behavior response offset value = (human operation time point - machine execution time point). Perform feature recognition on the duration information of human-machine collaborative behavior, calculate the difference in collaboration duration between the extracted human operation duration and machine execution duration of each task unit, and obtain the behavior collaboration offset value of each task unit. Specifically: Behavior collaboration offset value = (human operation duration - machine execution duration).

3. The intelligent human-machine collaboration method for enterprises according to claim 2, characterized in that: The specific steps in S1 also include: S13. Based on the magnitude of the behavioral response offset value and behavioral collaboration offset value of each task unit, determine whether the human-machine collaboration operation process is in an abnormal state, and issue a human-machine collaboration analysis command. The specific process includes: If the behavior response offset value of the corresponding task unit is not equal to zero, a behavior response abnormal signal is issued; if the behavior response offset values ​​of the corresponding task units are all equal to zero, a behavior response normal signal is issued. If the behavior collaboration offset value of the corresponding task unit is not equal to zero, a behavior collaboration abnormal signal is issued; if the behavior collaboration offset values ​​of the corresponding task units are all equal to zero, a behavior collaboration normal signal is issued. If both a normal behavior response signal and a normal behavior collaboration signal are received simultaneously, it indicates that the human-machine collaboration operation process is in a normal state, and a normal human-machine collaboration instruction is issued, which means: continue to execute the subsequent human-machine collaboration operation process; otherwise, it indicates that the human-machine collaboration operation process is in an abnormal state, and a human-machine collaboration analysis instruction is issued.

4. The intelligent human-machine collaboration method for enterprises according to claim 3, characterized in that: The specific steps in S2 include: S21. Upon receiving the human-machine collaboration analysis instruction, feature recognition is performed on the duration information of human-machine collaboration behavior. The extracted human operation duration and machine execution duration of the corresponding task units are correlated. After dimensionless processing, the consistency of human-machine operation rhythm in each task unit during the human-machine collaboration operation is analyzed, and the human-machine collaboration consistency coefficient of each task unit is determined. Specifically: In the formula, Let represent the human-machine collaboration consistency coefficient of the i-th task unit. and These represent the human operation time and machine execution time for the i-th task unit, respectively. S22. Perform feature recognition on the human-machine collaboration behavior feature information recorded in the human-machine collaboration behavior log, and correlate the number of human error operations, total number of human-machine interactions, number of machine rollbacks due to errors, and total number of machine commands received in each task unit. After dimensionless processing, analyze the degree of human-machine interaction interference in each task unit during the human-machine collaboration operation, and determine the human-machine interaction interference coefficient of each task unit, specifically as follows: In the formula, This represents the human-computer interaction interference coefficient of the i-th task unit. , , and These represent the number of human error operations, the total number of human-machine interactions, the number of machine rollbacks due to errors, and the total number of machine instructions received in the i-th task unit, respectively. S23. By deploying near-infrared cameras and combining them with facial key point recognition algorithms, the eyelid opening and closing state and mouth shape state of personnel are monitored in real time. The total duration of eye closure and the number of yawns of personnel in each task unit during the human-machine collaborative operation are obtained. After normalization processing, the fatigue level of personnel in each task unit during the human-machine collaborative operation is analyzed, and the fatigue coefficient of personnel in each task unit is determined.

5. The intelligent human-machine collaboration method for enterprises according to claim 4, characterized in that: The specific steps of S3 include: S31. The sum of the human-computer interaction interference coefficient and the personnel fatigue coefficient of the corresponding task unit is calculated by subtracting the human-computer collaboration consistency coefficient of the corresponding task unit to obtain the comprehensive collaboration efficiency value of each task unit. The average comprehensive collaboration efficiency is obtained by combining the statistical mean calculation algorithm. S32. Compare and analyze the overall collaboration efficiency value of each task unit with the average overall collaboration efficiency to screen out the set of inefficient collaboration task units. The specific process includes: If the overall collaboration efficiency value of the corresponding task unit exceeds the average overall collaboration efficiency, the corresponding task unit is marked as a high-efficiency collaboration task unit, and a set of high-efficiency collaboration task units is constructed. If the overall collaboration efficiency value of the corresponding task unit does not exceed the average overall collaboration efficiency, the corresponding task unit is marked as a low-efficiency collaboration task unit, and a set of low-efficiency collaboration task units is constructed. At the same time, the behavioral response offset value and behavioral collaboration offset value of each low-efficiency collaboration task unit in the low-efficiency collaboration task unit set are recorded.

6. The intelligent human-machine collaboration method for enterprises according to claim 5, characterized in that: The specific steps in S3 also include: S33. Execute the corresponding machine collaboration control strategy for the corresponding inefficient collaborative task unit. The specific execution process includes: Based on the magnitude of the behavioral response offset value and behavioral collaboration offset value of the selected inefficient collaborative task unit, the corresponding machine collaboration control strategy is executed. If both the behavioral response offset value and the behavioral collaboration offset value of the inefficient collaborative task unit are greater than zero, the first machine collaboration control strategy is executed, which involves advancing the machine execution time of the inefficient collaborative task unit and increasing the machine execution duration. If both the behavioral response offset value and the behavioral collaboration offset value of the inefficient collaborative task unit are less than zero, the second machine collaboration control strategy is executed, which involves continuing to maintain the human-machine collaboration operation process of the inefficient collaborative task unit. Otherwise, the third machine collaboration control strategy is executed, which involves advancing the machine execution time of the inefficient collaborative task unit or increasing the machine execution duration.

7. The intelligent human-machine collaboration method for enterprises according to claim 6, characterized in that: The specific steps of S4 include: S41. After executing the corresponding machine collaboration control strategy, the status of the finished products produced during the human-machine collaboration operation under control within a unit cycle is monitored in real time, and the finished product data information after control is obtained. The finished product data information after control includes the finished product production quantity and yield rate within a unit cycle. S42. Perform feature identification on the finished product data information after regulation, correlate the extracted finished product production quantity and yield rate within a unit cycle, and after dimensionless processing, analyze the quality and production efficiency level of the finished products produced during the human-machine collaborative operation after regulation within a unit cycle, and determine the optimal production coefficient after regulation.

8. The intelligent human-machine collaboration method for enterprises according to claim 7, characterized in that: The specific steps in S4 also include: S43. Before implementing the corresponding machine collaboration control strategy, the quality and production efficiency of the finished products produced during the human-machine collaboration operation before the control are analyzed in the historical unit cycle to determine the pre-control yield coefficient, which is then used as the yield comparison threshold. S44. Compare and analyze the adjusted yield coefficient with the yield comparison threshold to determine whether the currently implemented machine collaboration control strategy is effective, and implement the corresponding human collaboration control strategy for the corresponding inefficient collaboration task units. The specific process includes: If the yield coefficient after adjustment exceeds the yield comparison threshold, it indicates that the currently implemented machine collaboration adjustment strategy is effective, and the first machine collaboration adjustment strategy is generated, which is: continue to maintain the human-machine collaboration operation process after adjustment. If the yield coefficient after adjustment does not exceed the yield comparison threshold, it indicates that the currently implemented machine collaboration control strategy is invalid. A second machine collaboration control strategy is then generated, which is to arrange rest and personnel replacement for the operators in the corresponding inefficient collaboration task units.

9. An intelligent human-machine collaboration system for enterprises, used to implement the intelligent human-machine collaboration method for enterprises as described in any one of claims 1 to 8, characterized in that: It includes an early warning analysis module, a human-machine collaboration analysis module, a machine control module, and a personnel control module; The early warning analysis module is used to divide the human-machine collaboration operation process into task units, and extract the timing information and duration information of human-machine collaboration behavior by combining human-machine collaboration behavior logs. By analyzing the degree of behavioral response deviation and behavioral collaboration deviation of each task unit, it issues human-machine collaboration analysis instructions. The human-machine collaboration analysis module is used to analyze the consistency of human-machine operation rhythm of each task unit during human-machine collaboration operation based on the timing information of human-machine collaboration behavior after receiving the human-machine collaboration analysis instruction, analyze the degree of human-machine interaction interference of each task unit during human-machine collaboration operation based on the human-machine collaboration behavior feature information recorded in the human-machine collaboration behavior log, and analyze the degree of human-machine interaction interference of each task unit during human-machine collaboration operation through the deployed near-infrared camera and combined with the facial key point recognition algorithm. The machine control module is used to screen out the set of inefficient collaborative task units after comprehensive comparative analysis based on the consistency of human-machine operation rhythm, the degree of human-machine interaction interference, and the degree of human fatigue, and to execute the corresponding machine collaboration control strategy on the corresponding inefficient collaborative task units in the set of inefficient collaborative task units. The personnel control module is used to determine the effectiveness of the machine collaboration control strategy by analyzing the quality and production efficiency of the finished products produced during the human-machine collaboration operation after the corresponding machine collaboration control strategy is executed, and to execute the corresponding personnel collaboration control strategy for the corresponding low-efficiency collaboration task unit.

Images