Human-machine collaborative work task intelligent matching method and system

Abstract

The present application relates to the technical field of man-machine cooperation, in particular to a work task intelligent matching method and system for man-machine cooperation, comprising: a task splitting module, a proxy right evaluation module, a task static allocation module and a task dynamic optimization module; the task splitting module splits a task into a series of independent atomic tasks; the proxy right evaluation module evaluates the proportion of the proxy right of artificial and collaborative robots in the atomic task; the task static allocation module matches man and machine based on the proxy right and completes task allocation; the task dynamic optimization module performs dynamic optimization based on work details; the system calculates the proxy right of the atomic task through the proxy right evaluation module, thereby improving the accuracy and efficiency of task allocation and ensuring work efficiency.

Description

Technical Field

[0001] This invention relates to the field of human-machine collaboration technology, specifically to a method and system for intelligent matching of work tasks in human-machine collaboration. Background Technology

[0002] Human-machine collaboration technology refers to the deep integration of human and machine intelligence technologies. In the process of jointly completing tasks, through effective interaction and cooperation, work efficiency and task quality are improved. It is widely used in manufacturing, healthcare, services and military fields.

[0003] Human-machine collaboration technology allocates tasks to humans and collaborative robots based on the complexity of the tasks, the capabilities of humans and machines, and production needs. This leverages the advantages of humans in complex decision-making, creative thinking, and situational understanding, while utilizing the capabilities of collaborative robots in data processing, pattern recognition, and execution efficiency to jointly solve complex problems.

[0004] However, existing task allocation methods mainly focus on the cost and time of task completion, often assigning tasks to entities with lower costs or shorter time requirements, without matching tasks with the entity's situation to ensure the safety and efficiency of the entity in completing the task.

[0005] In view of this, we propose an intelligent matching method and system for human-machine collaborative work tasks. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent matching method and system for human-machine collaborative work tasks, so as to solve the problem mentioned in the background art of not matching and analyzing tasks with subjects when allocating human-machine collaborative tasks.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A human-machine collaborative work task intelligent matching system includes: a task splitting module, a delegation authority evaluation module, a static task allocation module, and a dynamic task optimization module. The task splitting module breaks down tasks into a series of independent atomic tasks. It obtains the actual content of the work tasks, identifies and splits the dependencies between the work content, then draws an acyclic graph to clarify the dependencies, and finally verifies the split tasks to determine their executableness and generate atomic tasks. The delegation authority evaluation module evaluates the proportion of human and collaborative robot delegation authority in the atomic tasks. It calculates the delegation authority of the task to determine whether the atomic task is more suitable for human processing. The system still utilizes collaborative robots to improve the accuracy of task allocation and the efficiency of task processing. The static task allocation module matches humans and machines based on delegation rights and completes task allocation. The static task allocation unit assigns atomic tasks to humans and collaborative robots based on the delegation rights of atomic tasks, improving the matching degree between humans, collaborative robots, and atomic tasks, thereby improving task completion efficiency. The dynamic task optimization module performs dynamic optimization based on work details. During the human-machine collaborative task execution process, dynamic task optimization collects work data and dynamically adjusts the task allocation scheme based on the existing work data to ensure the rationality of task allocation.

[0008] Preferably, the delegation evaluation module includes a task feature unit, a task risk unit, a task security unit, a hierarchical analysis unit, and a delegation calculation unit. The task feature unit performs feature analysis and extraction on atomic tasks, judging their flexibility, complexity, technicality, repetitiveness, urgency, and other characteristics based on the content of the atomic tasks. The task risk unit evaluates the risks of manual execution, as humans are affected by physiological and psychological fatigue during atomic task execution, leading to health risks and impacting task completion over extended periods. The task security unit evaluates the security of both manual and collaborative robot task results, determining the safer delegation method for executing the atomic task. The hierarchical analysis unit calculates delegation weights based on task features, task risk, and task security, constructing a hierarchical analysis model and calculating delegation weights through hierarchical analysis. The delegation calculation unit calculates delegation rights based on these weights, determining the appropriate delegation method for the atomic task—whether manual or collaborative robot delegation is more suitable.

[0009] Preferably, the hierarchical analysis unit includes a hierarchical analysis model component, a judgment matrix component, a verification component, and a global weight calculation component. The hierarchical analysis model component constructs a model based on task characteristics, task risks, and task safety. It decomposes the research objective layer by layer into multiple constituent factors, constructing a top-down ordered hierarchical structure model. The judgment matrix component generates a matrix based on historical matching data for each level of the hierarchical analysis model. The historical matching data provides judgments on the relative importance of each factor at each level of the indicator system, and these judgments are numerically represented using an exponential scale to form the judgment matrix. The verification component calculates local weights based on the judgment matrix and verifies the consistency of the judgment matrix. It uses a sum-product method to calculate the local weights of each indicator and calculates the consistency ratio, thereby verifying whether the judgment matrix meets the consistency requirements. The global weight calculation component synthesizes global weights based on local weights and obtains the global weights of each secondary indicator through weight synthesis.

[0010] Preferably, the static task allocation module includes a proxy right matching unit, a priority matching unit, a workload matching unit, and a task allocation unit. The proxy right matching unit matches tasks to humans and collaborative robots based on proxy rights. It obtains the proxy right calculation results for each atomic task, thereby classifying the atomic tasks into human tasks and collaborative robot tasks. The priority matching unit matches tasks based on the task priorities of humans and collaborative robots. It judges the priorities of each atomic task within the human and collaborative robot tasks to determine the order of task execution. The workload matching unit matches tasks based on the human's physical condition and workload. It obtains the human's age and physical examination data to evaluate the human's health, and then matches tasks based on the workload of the corresponding atomic tasks to ensure the human's physical and mental health. The task allocation unit allocates tasks based on the multi-dimensional matching results. It combines proxy right matching, priority matching, and workload matching to determine the atomic task allocation results, thereby completing the task allocation of atomic tasks and proxies.

[0011] Preferably, the dynamic task optimization includes a task detection unit, an optimization calculation unit, and a feedback unit. The task detection unit detects and collects work data from both human and collaborative robot agents, including data such as work time, efficiency, results, and effectiveness, to determine the matching degree between atomic tasks and agents. The optimization calculation unit optimizes the weights of similar tasks based on the work data. The optimization calculation unit connects to a hierarchical analysis model (AHP) and adjusts the AHP based on the work data to calculate the agency rights. The feedback unit sends the optimization calculation results to a static task matching module to match and optimize unmatched tasks. The feedback unit obtains the recalculated agency rights and sends them to the static task matching module, which then reassigns unassigned tasks to ensure efficient task completion.

[0012] A method for intelligent matching of work tasks for human-machine collaboration includes the following steps: Step 1: The task splitting module breaks down the task into a series of sub-tasks and further refines them into atomic tasks; The task decomposition module breaks down tasks into a series of tasks through workflow. One workflow step corresponds to one series, and then the tasks within the series are further refined and broken down into atomic tasks. Step 2: The proxy evaluation module evaluates and judges the human-machine proxy rights of the atomic task; The delegation evaluation module obtains atomic tasks, analyzes and evaluates them, and calculates the delegation rights of the atomic tasks. Step 3: The static task allocation module matches and allocates tasks based on the proxy rights and the actual content of the atomic tasks. The static task allocation module first completes the allocation of manual agent tasks and collaborative robot agent tasks based on the proxy rights. Then, it matches the corresponding agent according to the actual content of the atomic tasks of the corresponding category to complete the task allocation. Step 4: The task dynamic optimization module collects work data in real time and optimizes the allocation of unassigned tasks based on the work data; The task dynamic optimization module adjusts the delegation rights based on actual work data, and then reallocates unassigned tasks according to the new delegation rights.

[0013] Furthermore, step 2 above also includes the following steps: Step 2.1: The task feature unit performs feature extraction and analysis based on the task content; The task feature unit determines the basic characteristics of the task based on the content of the atomic task, such as: flexibility, complexity, technicality, repetitiveness, urgency, etc. Step 2.2: The task risk unit evaluates the risks to human personnel during task execution; The task risk unit evaluates the psychological and physiological risks of manual proxy execution based on the complexity, repetition and other characteristics of the task; Step 2.3: The task security unit evaluates the security of the task execution process and results; The task safety unit evaluates the safety of the atomic task during execution by both manual and collaborative robots. Step 2.4: Based on the above analysis and evaluation, the hierarchical analysis unit constructs a hierarchical analysis model to calculate the proxy weights; The hierarchical analysis unit constructs a hierarchical analysis model based on task characteristics, task risks, and task security, and calculates the agent weights through hierarchical analysis; Step 2.5: The delegation calculation unit calculates the delegation rights of each atomic task based on the delegation weight; The delegation calculation unit determines the delegation rights based on the delegation weight, thereby judging whether the atomic task is more suitable for human delegation or collaborative robot delegation.

[0014] Furthermore, step 2.4 above also includes the following steps: Step 2.4.1: Construct the hierarchical analysis model using the hierarchical analysis model components; The hierarchical analysis model component decomposes task characteristics, task risks, and task security into multiple components layer by layer, constructing a top-down ordered hierarchical analysis model. Step 2.4.2: The judgment matrix component generates a judgment model based on historical data and in conjunction with each layer of the hierarchical analysis model; The judgment matrix component, based on historical task allocation, task process, task results and other data, compares the importance of each factor in the same level of the hierarchical analysis model with the factors in the next level to form a judgment matrix. Step 2.4.3: The verification component calculates the local weights of the judgment matrix by performing a sum-product operation and verifies consistency. The verification component uses the sum-product method to calculate the local weights of each indicator and calculate the consistency ratio, thereby verifying whether the judgment matrix meets the consistency requirements. Step 2.4.4: If the consistency check fails, make manual adjustments based on historical data until the check passes. If the consistency check fails, it is adjusted manually. The judgment matrix is ​​adjusted manually based on historical data to ensure that the consistency check passes, thereby improving the reliability of the decision. Step 2.4.5: After the verification is passed, the global weight calculation component will synthesize the local weights to calculate the global matrix; After the test is passed, the global weight calculation component obtains the global weight through weight synthesis.

[0015] Furthermore, step 3 above also includes the following steps: Step 3.1: The proxy matching unit obtains the proxy calculation results and classifies the atomic tasks accordingly; The proxy matching unit obtains the proxy calculation results of each atomic task, and then divides the atomic tasks into manual tasks and collaborative robot tasks; Step 3.2: The priority matching unit performs priority judgment on the classified atomic tasks to determine the task allocation order; The priority matching unit determines the order of task execution based on the priority of each atomic task of the manual task and the collaborative robot task. Step 3.3: The workload matching unit acquires human data and matches it with the workload of each task to ensure the physiological and mental health of each human. The workload matching unit acquires the age and physical examination data of the artificial being and then evaluates the physical health of the artificial being. It then matches the workload of the atomic tasks within the corresponding category to ensure the physical and mental health of the artificial being. Step 3.4: The task allocation unit obtains multivariate matching data and completes task allocation; The task allocation unit combines proxy matching, priority matching, and workload matching to determine the atomic task allocation result, thereby completing the task allocation of atomic tasks and proxies.

[0016] Furthermore, step 4 above also includes the following steps: Step 4.1: After the task begins, the task detection unit collects and detects the work data of each human and collaborative robot. The task detection unit collects work data from both human agents and collaborative robot agents to determine whether the current type of atomic task needs to have its delegation authority adjusted. Step 4.2: The optimization calculation unit acquires working data and connects it to the hierarchical analysis unit to adjust and calculate the model; The computing unit is optimized to access the hierarchical analysis model. The hierarchical analysis model is optimized using working data, and then the agency rights are recalculated. Step 4.3: The feedback component sends the proxy rights to the proxy rights matching unit; The feedback unit sends the new agency rights to the agency rights matching unit to classify human agents and collaborative robot agents that have not been assigned atomic tasks; Step 4.4: The task allocation unit reassigns and adjusts the unassigned tasks; The task allocation unit, in conjunction with the new agency matching results, reallocates unassigned tasks, thus completing the optimization and adjustment.

[0017] Compared with the prior art, the beneficial effects of the present invention are: 1. A method and system for intelligent matching of work tasks for human-machine collaboration. This system calculates the delegation rights of atomic tasks through a delegation rights evaluation module, thereby improving the accuracy and efficiency of task allocation and ensuring work efficiency.

[0018] 2. A method and system for intelligent matching of work tasks for human-machine collaboration. This system achieves multi-dimensional data matching through a static task allocation module, thereby improving the fit between agent and atomic tasks and ensuring task efficiency and completion.

[0019] 3. A method and system for intelligent matching of work tasks for human-machine collaboration. This system optimizes and adjusts task allocation based on actual work data through a task dynamic optimization module, thereby avoiding and reducing the deviation between tasks and agents and improving task matching degree. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the main framework of the intelligent task matching system of the present invention; Figure 2 This is a detailed framework diagram of the intelligent task matching system of the present invention; Figure 3 This is a flowchart of the main body of the intelligent task matching method of the present invention; Figure 4 This is a detailed flowchart of the intelligent task matching method of the present invention.

[0021] In the picture: 1. Task splitting module; 2. Agency Rights Evaluation Module; 21. Task Feature Unit; 22. Task Risk Unit; 23. Task Security Unit; 24. Hierarchical Analysis Unit; 241. Hierarchical Analysis Model Component; 242. Judgment Matrix Component; 243. Verification Component; 244. Global Weight Calculation Component; 25. Agency Rights Calculation Unit; 3. Static task allocation module; 31. Agency rights matching unit; 32. Priority matching unit; 33. Workload matching unit; 34. Task allocation unit; 4. Task dynamic optimization module; 41. Task detection unit; 42. Optimization calculation unit; 43. Feedback unit. Detailed Implementation

[0022] 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.

[0023] Human-machine collaboration technology allocates tasks to humans and collaborative robots based on the complexity of the tasks, the capabilities of humans and machines, and production needs. This leverages the advantages of humans in complex decision-making, creative thinking, and situational understanding, while utilizing the capabilities of collaborative robots in data processing, pattern recognition, and execution efficiency to jointly solve complex problems.

[0024] However, existing task allocation methods mainly focus on the cost and time of task completion, often assigning tasks to entities with lower costs or shorter time requirements, without matching tasks with the entity's situation to ensure the safety and efficiency of the entity in completing the task.

[0025] like Figures 1 to 4 As shown, an intelligent task matching method and system for human-machine collaboration includes: like Figures 1 to 2 As shown, the intelligent task matching system for human-machine collaboration includes: a task splitting module 1, a delegation rights evaluation module 2, a static task allocation module 3, and a dynamic task optimization module 4. The task splitting module 1 breaks down tasks into a series of independent atomic tasks; the delegation rights evaluation module 2 evaluates the proportion of human and collaborative robot delegation rights in the atomic tasks; the static task allocation module 3 performs human-machine matching based on delegation rights and completes task allocation; and the dynamic task optimization module 4 performs dynamic optimization based on task details. Specifically, the task splitting module 1 breaks down the task into a series of independent atomic tasks. It obtains the actual content of the task, identifies and splits the dependencies between the content, draws an acyclic graph to clarify the dependencies, and finally verifies the split tasks to determine their executableness, thus generating atomic tasks. The task splitting module 1 achieves atomic task splitting based on static splitting. Static splitting achieves task splitting by determining the workflow. Although static splitting has lower flexibility, it facilitates the screening of atomic tasks, thereby determining the correctness of the atomic task splitting and the correctness of the dependencies between atomic tasks. The delegation evaluation module 2 evaluates the proportion of human and collaborative robot delegation in the atomic tasks. It calculates the delegation of a task to determine whether the atomic task is more suitable for human or collaborative robot processing, improving the accuracy of task allocation and the efficiency of task processing. In this application, delegation refers to both human and collaborative robots. Delegation is a comprehensive evaluation index used to measure the degree of matching between human and collaborative robot delegation for a certain atomic task. The allocation of atomic tasks is achieved through the degree of matching of delegation. The static task allocation module 3 performs human-machine matching and task allocation based on agency rights. The static task allocation unit assigns atomic tasks to humans and collaborative robots based on the agency rights of atomic tasks, improving the matching degree between humans, collaborative robots, and atomic tasks, thereby increasing task completion efficiency. While pursuing economic or efficiency goals, task allocation also emphasizes the health and safety of humans and the rationality of resource allocation, which is conducive to creating a good working environment and improving the comfort of human workers. The dynamic task optimization module 4 performs dynamic optimization based on work details. During the human-machine collaborative task execution process, work data is collected, and the task allocation scheme is dynamically adjusted according to the existing work data to ensure the rationality of task allocation. During the human-machine collaborative operation, the system generates feedback information through work data and dynamically adjusts the task allocation scheme based on the feedback information, thereby better adapting to environmental changes and improving task completion efficiency.

[0026] In this embodiment, the delegation evaluation module 2 includes a task feature unit 21, a task risk unit 22, a task security unit 23, a hierarchical analysis unit 24, and a delegation calculation unit 25. The task feature unit 21 performs feature analysis and extraction on atomic tasks; the task risk unit 22 evaluates the risks during manual execution; the task security unit 23 evaluates the security of task results from both manual and collaborative robot perspectives; the hierarchical analysis unit 24 calculates delegation weights based on task features, task risks, and task security; and the delegation calculation unit 25 calculates delegation rights based on these delegation weights. Specifically, the task feature unit 21 performs feature analysis and extraction on the atomic tasks. Based on the content of the atomic tasks, the task feature unit 21 extracts features to determine the flexibility, complexity, technicality, repetitiveness, urgency, and other characteristics of the tasks, and calculates the delegation rights based on these task characteristics. The task risk unit 22 evaluates the risks during manual execution. When performing atomic tasks, humans are affected by physiological and psychological fatigue, which can lead to health risks under prolonged work, thus affecting the completion of the atomic tasks. The intensity, complexity, and duration of the tasks can all cause physiological fatigue, leading to prolonged reaction time, decreased action accuracy, and even an increased probability of operational errors. In addition, the psychological pressure brought by the task, such as the worry about the consequences of failure in high-risk tasks or the time pressure in urgent tasks, may also cause emotional fluctuations or a decline in cognitive function, weakening the executor's decision-making ability. The task safety unit 23 evaluates the safety of the results of human and collaborative robot tasks. The task safety unit 23 judges the safety of human execution and collaborative robot execution during the execution of the atomic task, thereby determining a safer execution agent. For example, in the chemical industry, the task involves the handling of high temperature, high pressure, toxic or corrosive substances, and the safety of using a collaborative robot agent is higher. In medical surgery scenarios, the accuracy and reliability of the task execution are crucial, and the safety of using a human agent is higher. The hierarchical analysis unit 24 calculates the proxy weight based on task characteristics, task risks, and task security. It constructs a hierarchical analysis model based on these characteristics and calculates the proxy weight through hierarchical analysis. The proxy weight calculation unit 25 calculates the proxy weight based on the proxy weight and determines the proxy weight, thereby determining whether the atomic task is more suitable for human or collaborative robot proxy. The proxy weight calculation unit calculates the proxy weight based on grey theory. When evaluating the matching degree between the proxy and the atomic task based on the proxy weight, due to differences in personal experience, cognitive perspective, and understanding of the evaluation object, there is often a high degree of subjectivity and the evaluation result is somewhat ambiguous. Using grey numbers and whitening weight functions in grey theory can effectively reduce the influence of subjective factors, more scientifically analyze the weight ratio of each indicator, and make the evaluation result more accurate and objective.

[0027] In this embodiment, the hierarchical analysis unit 24 includes a hierarchical analysis model component 241, a judgment matrix component 242, a verification component 243, and a global weight calculation component 244. The hierarchical analysis model component constructs a model based on task characteristics, task risks, and task security. The judgment matrix component 242 generates a matrix based on historical matching data for judging each level of the hierarchical analysis model. The verification component 243 calculates local weights based on the judgment matrix and verifies the consistency of the judgment matrix. The global weight calculation component 244 synthesizes global weights based on local weights. Specifically, the hierarchical analysis model component constructs a model based on task characteristics, task risks, and task safety. It decomposes the research objective layer by layer into multiple constituent factors, constructing a top-down ordered hierarchical structure model. The judgment matrix component 242 generates a matrix based on historical matching data for each level of the hierarchical analysis model. Historical matching data provides judgments on the relative importance of each factor at each level of the indicator system, and these judgments are numerically represented using an exponential scale to form a judgment matrix. This judgment matrix is ​​formed by pairwise comparisons of the importance of each factor at the same level relative to the factor at the next higher level. The verification component 243 calculates local weights based on the judgment matrix and verifies its consistency. It uses a sum-product method to calculate the local weights of each indicator and calculates the consistency ratio, thereby verifying whether the judgment matrix meets the consistency requirements. Consistency verification avoids distortion of weight calculation results due to arbitrariness or contradictions in the judgment process, thus improving the scientific rigor and reliability of decision-making. The global weight calculation component 244 synthesizes global weights based on local weights, obtaining the global weights of each secondary indicator through weight synthesis.

[0028] In this embodiment, the static task allocation module 3 includes a proxy matching unit 31, a priority matching unit 32, a workload matching unit 33, and a task allocation unit 34. The proxy matching unit 31 matches tasks to humans and collaborative robots based on proxy rights. The priority matching unit 32 matches tasks based on the task priorities of humans and collaborative robots. The workload matching unit 33 matches tasks based on the human's physical condition and workload. The task allocation unit 34 allocates tasks based on the multi-dimensional matching results. Specifically, the workload matching unit 33 and the delegation matching unit 31 match tasks to humans and collaborative robots based on delegation rights. The delegation matching unit 31 obtains the delegation rights calculation results of each atomic task and then classifies the atomic tasks into human tasks and collaborative robot tasks. The priority matching unit 32 matches tasks based on the priority of human and collaborative robot tasks. The priority matching unit 32 judges the priority of each atomic task of human and collaborative robot tasks to determine the order of task execution. The workload matching unit 33 matches tasks based on human physical fitness and workload. The workload matching unit 33 obtains human age and physical examination data to evaluate human physical health and then matches the workload of atomic tasks in the corresponding category to ensure human physical and mental health. By ensuring that the workload matches the human physical fitness, the high workload of the human avoids affecting the human's physical and mental health. The task allocation unit 34 allocates tasks based on the multi-dimensional matching results. The task allocation unit 34 combines delegation rights matching, priority matching, and workload matching to determine the atomic task allocation results, thereby completing the task allocation of atomic tasks and delegation.

[0029] In this embodiment, the dynamic task optimization includes a task detection unit 41, an optimization calculation unit 42, and a feedback unit 43; the task detection unit 41 detects and collects work data from both human and collaborative robots; the optimization calculation unit 42 performs optimization calculations on the weights of tasks of the same type based on the work data; and the feedback unit 43 sends the optimization calculation results to the static task matching module to match and optimize unmatched tasks. Specifically, the task detection unit 41 detects and collects work data from both human and collaborative robot agents, including data such as work time, efficiency, results, and benefits, to determine the matching degree between atomic tasks and agents. The optimization calculation unit 42 optimizes the weights of similar tasks based on the work data. This unit connects to the hierarchical analysis model (AHP) and adjusts the AHP based on the work data to calculate the agent rights. The feedback unit 43 sends the optimization calculation results to the static task matching module to match and optimize unmatched tasks. The feedback unit 43 compares the recalculated agent rights with the previously calculated ones. If the agent rights deviate beyond a set threshold, the new agent rights are sent to the static task matching module, which then reassigns unassigned tasks to ensure efficient task completion.

[0030] Figure 1This is a main framework diagram of the intelligent task matching system of the present invention. The task splitting module 1 splits the task into atomic tasks and then analyzes them through the task feature unit 21, task risk unit 22 and task security unit 23. The analysis results are sent to the hierarchical analysis unit 24 to construct a hierarchical analysis model to calculate the proxy weight. Then, the proxy right calculation unit 25 calculates the proxy right and sends it to the proxy right matching unit 31. The proxy right matching unit 31, priority matching unit 32 and workload matching unit 33 send the matching results to the task allocation unit 34. The task allocation unit 34 completes the task allocation. The task detection unit 41 collects data during the work and sends it to the optimization calculation unit 42 to optimize the proxy right. The feedback component sends the new proxy right to the proxy right matching unit 31 to re-match and adjust the unassigned tasks, and completes the optimization.

[0031] Figure 2 This is a detailed framework diagram of the intelligent task matching system of the present invention. The task splitting module 1 splits the task into atomic tasks, which are then analyzed by the task feature unit 21, task risk unit 22, and task security unit 23. The analysis results are sent to the hierarchical analysis unit 24. In the hierarchical analysis unit 24, the hierarchical analysis model component constructs a hierarchical analysis model based on the analysis data, the judgment matrix component 242 constructs a judgment matrix based on the hierarchical analysis model, the verification component 243 calculates local weights based on the judgment matrix and verifies consistency, and the global weight calculation component 244 calculates the global weight based on the local weights and sends it to the proxy right calculation unit 25. The proxy right calculation unit 25 calculates the proxy right and sends it to the proxy right calculation unit 25. The matching unit 31 performs matching based on proxy rights; the priority matching unit 32 performs matching based on the priority of atomic tasks; and the workload matching unit 33 performs matching based on artificial physical health conditions. The matching results of the above three units are sent to the task allocation unit 34, which allocates atomic tasks. Then, the task detection unit 41 detects and collects real-time work data. The optimization calculation unit 42 adjusts the hierarchical analysis model based on the work data and calculates new proxy rights. The feedback component compares the new and old proxy rights. If the difference exceeds the set limit, the new proxy rights are sent to the proxy rights matching unit 31 to re-match and allocate unassigned tasks.

[0032] A method for intelligent matching of work tasks for human-machine collaboration includes the following steps: Step 1: Task splitting module 1 splits the task into a series and further breaks it down into atomic tasks; Step 2: The delegation evaluation module 2 evaluates and judges the human-machine delegation rights of the atomic task; Step 3: The static task allocation module 3 matches and allocates tasks based on the proxy rights and the actual content of the atomic tasks. Step 4: Task Dynamic Optimization Module 4 collects work data in real time and optimizes the allocation of unassigned tasks based on the work data; Specifically, the task decomposition module 1 breaks down tasks into a series of tasks through a workflow, with one workflow step corresponding to one series. Then, the tasks within a series are further subdivided into atomic tasks. The delegation rights evaluation module 2 obtains atomic tasks and then analyzes and evaluates them to calculate the delegation rights of the atomic tasks. The static task allocation module 3 first completes the allocation of manual delegation tasks and collaborative robot delegation tasks based on the delegation rights. Then, it matches the corresponding delegation according to the actual content of the atomic tasks of the corresponding category to complete the task allocation. The dynamic task optimization module 4 adjusts the delegation rights based on actual work data, and then reallocates the unallocated tasks according to the new delegation rights.

[0033] Furthermore, step 2 above also includes the following steps: Step 2.1: Task feature unit 21 performs feature extraction and analysis based on the task content; Step 2.2: Task risk unit 22 evaluates the risks to human personnel during task execution; Step 2.3: Task security unit 23 evaluates the security of the task execution process and results; Step 2.4: Based on the above analysis and evaluation, the hierarchical analysis unit constructs a hierarchical analysis model to calculate the proxy weights; Step 2.5: The proxy calculation unit 25 calculates the proxy rights of each atomic task based on the proxy weight; Specifically, the task feature unit 21 determines the basic characteristics of the task based on its content, such as flexibility, complexity, technicality, repetitiveness, and urgency; the task risk unit 22 evaluates the psychological and physiological risks of manual execution based on the task's complexity, repetitiveness, and other characteristics; the task safety unit 23 evaluates the safety of both manual and collaborative robot execution of the atomic task; the hierarchical analysis unit 24 constructs a hierarchical analysis model based on the task features, task risks, and task safety, and calculates the agent weights through hierarchical analysis; and the agent weight calculation unit 25 determines the agent weights based on the agent weights to determine whether the atomic task is more suitable for manual or collaborative robot execution.

[0034] Furthermore, step 2.4 above also includes the following steps: Step 2.4.1: Hierarchical analysis model component 241 constructs the hierarchical analysis model; Step 2.4.2: Judgment matrix component 242 generates a judgment model based on historical data and in combination with each layer of the hierarchical analysis model; Step 2.4.3: Component 243 performs a sum-product calculation on the judgment matrix to determine local weights and checks for consistency. Step 2.4.4: If the consistency check fails, make manual adjustments based on historical data until the check passes. Step 2.4.5: After the verification is passed, the global weight calculation component 244 will synthesize the local weights to calculate the global matrix; Specifically, the hierarchical analysis model component decomposes task characteristics, task risks, and task safety into multiple constituent factors layer by layer, constructing a top-down ordered hierarchical analysis model; the judgment matrix component 242, based on historical task allocation, task process, task results, and other data, compares the importance of each factor in the same level of the hierarchical analysis model with the factors in the previous level to form a judgment matrix; the verification component 243 uses the sum-product method to calculate the local weights of each indicator and calculate the consistency ratio, thereby verifying whether the judgment matrix meets the consistency requirements; if the consistency verification fails, it is manually adjusted, and the judgment matrix is ​​adjusted manually in combination with historical data to ensure that the consistency verification passes, thereby improving the reliability of the decision; after the verification passes, the global weight calculation component 244 obtains the global weight through weight synthesis.

[0035] Furthermore, step 3 above also includes the following steps: Step 3.1: The proxy matching unit 31 obtains the proxy calculation results and classifies the atomic tasks accordingly; Step 3.2: Priority matching unit 32 performs priority judgment on the classified atomic tasks to determine the task allocation order; Step 3.3: The workload matching unit 33 acquires human data and matches it with the workload of each task to ensure the physiological and mental health of each human. Step 3.4: Task allocation unit 34 acquires multivariate matching data and completes task allocation; Specifically, the proxy matching unit 31 obtains the proxy calculation results of each atomic task, and then divides the atomic tasks into human tasks and collaborative robot tasks for proxy matching; the priority matching unit 32 judges the priority of each atomic task of the human task and the collaborative robot task to determine the order of task execution; the workload matching unit 33 obtains the age and physical examination data of the human to evaluate the human's physical health, and then matches it with the workload of the atomic tasks in the corresponding category to ensure the human's physical and mental health; the task allocation unit 34 combines the proxy matching, priority matching and workload matching to determine the atomic task allocation results, thereby completing the task allocation of atomic tasks and proxies.

[0036] Furthermore, step 4 above also includes the following steps: Step 4.1: After the task begins, the task detection unit 41 detects and collects the work data of each human and collaborative robot. Step 4.2: The optimization calculation unit 42 acquires the working data and connects it to the hierarchical analysis unit 24 to adjust and calculate the model; Step 4.3: The feedback component sends the proxy rights to the proxy rights matching unit 31; Step 4.4: Task allocation unit 34 reassigns and adjusts the unassigned tasks; Specifically, the task detection unit 41 collects work data from human agents and collaborative robot agents to determine whether the current type of atomic task needs to have its agency rights adjusted; the optimization calculation unit 42 accesses the hierarchical analysis model and optimizes the model using work data, thereby recalculating the agency rights; the feedback unit 43 sends the new agency rights to the agency rights matching unit 31 to classify the human agents and collaborative robot agents that have not been assigned atomic tasks; and the task allocation unit 34 reallocates the unassigned tasks based on the new agency rights matching results, completing the optimization adjustment.

[0037] Figure 3 This is a flowchart of the main process of the intelligent task matching method of the present invention. The flowchart describes the main process: the task splitting module 1 decomposes the task into a series of tasks based on the workflow, with one workflow link corresponding to one series. Then, the tasks within the series are further subdivided into atomic tasks. The atomic tasks are sent to the proxy right evaluation module 2. The proxy right evaluation module 2 analyzes and evaluates the atomic tasks to calculate the proxy right of the atomic tasks. The proxy right is then sent to the task static allocation module 3. The task static allocation module 3 first completes the allocation of manual proxy tasks and collaborative robot proxy tasks based on the proxy right. Then, it matches the corresponding proxy according to the priority and workload of the corresponding atomic tasks to complete the task allocation. During the actual execution of the task, the task dynamic optimization module 4 adjusts the proxy right based on the work data to reduce matching errors. The task static allocation module 3 re-allocates the unallocated tasks according to the new proxy right.

[0038] Figure 4 The following is a detailed flowchart of the intelligent task matching method of the present invention: the task splitting module 1 splits the task into atomic tasks, and then the task feature unit 21 judges the basic characteristics of the task based on the content of the atomic task; the task risk unit 22 evaluates the psychological and physiological risks of manual execution based on the complexity, repetition and other characteristics of the task; the task security unit 23 evaluates the security of the atomic task during the execution process of both manual execution and collaborative robot execution. Task feature unit 21, task risk unit 22 and task security unit 23 send the analysis results to hierarchical analysis unit 24. Hierarchical analysis unit 24 constructs a hierarchical analysis model based on the above data and calculates the proxy weight through hierarchical analysis. Proxy weight calculation unit 25 calculates the proxy weight based on the proxy weight and sends it to proxy weight matching unit 31. The proxy matching unit 31 obtains the proxy calculation result and then divides the atomic task into human task and collaborative robot task for matching; the priority matching unit 32 matches based on the priority of the atomic task; the workload matching unit 33 matches based on the human physical health condition, and sends the matching results of the above three units to the task allocation unit 34. The task allocation unit 34 determines the atomic task allocation result, thereby completing the task allocation of atomic task and proxy. The task detection unit 41 collects work data from human agents and collaborative robot agents to determine whether the current type of atomic task needs to have its agency rights adjusted; the optimization calculation unit 42 accesses the hierarchical analysis model and optimizes the hierarchical analysis model using work data, thereby recalculating the agency rights; the feedback unit 43 sends the new agency rights to the agency rights matching unit 31 to classify the human agents and collaborative robot agents that have not been assigned atomic tasks; the task allocation unit 34 reallocates the unassigned tasks based on the new agency rights matching results, completing the optimization adjustment.

[0039] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A work task intelligent matching system for human-machine collaboration, characterized in that, include: Task splitting module (1), delegation evaluation module (2), static task allocation module (3), and dynamic task optimization module (4); The task splitting module (1) breaks down the task into a series of independent atomic tasks; The proxy evaluation module (2) evaluates the proportion of human and collaborative robot proxy rights in atomic tasks; The static task allocation module (3) performs human-machine matching and completes task allocation based on the agency right; The task dynamic optimization module (4) performs dynamic optimization based on the task details.

2. The intelligent task matching system according to claim 1, characterized in that: The agency evaluation module (2) includes a task feature unit (21), a task risk unit (22), a task security unit (23), a hierarchical analysis unit (24), and an agency calculation unit (25). The task feature unit (21) performs feature analysis and extraction on the atomic task; The task risk unit (22) evaluates the risks during manual execution; The task safety unit (23) evaluates the safety of the results of human and collaborative robot tasks; The hierarchical analysis unit (24) calculates the proxy weight based on task characteristics, task risks, and task security. The agency right calculation unit (25) calculates agency right based on agency weight.

3. The intelligent task matching system according to claim 2, characterized in that: The hierarchical analysis unit (24) includes a hierarchical analysis model component (241), a judgment matrix component (242), a verification component (243), and a global weight calculation component (244). The hierarchical analysis model component is built based on task characteristics, task risks, and task security. The judgment matrix component (242) generates a matrix based on historical matching data for judgments at each level of the hierarchical analysis model; The verification component (243) calculates local weights based on the judgment matrix and verifies the consistency of the judgment matrix; The global weight calculation component (244) obtains the global weight based on the local weight synthesis.

4. The intelligent task matching system according to claim 1, characterized in that: The static task allocation module (3) includes a proxy matching unit (31), a priority matching unit (32), a workload matching unit (33), and a task allocation unit (34). The agency matching unit (31) matches tasks to human and collaborative robots based on agency rights; The priority matching unit (32) matches tasks based on the task priorities of humans and collaborative robots; The workload matching unit (33) matches tasks based on human physical fitness and workload; The task allocation unit (34) allocates tasks based on the multidimensional matching results.

5. The intelligent task matching system according to claim 1, characterized in that: The task dynamic optimization includes a task detection unit (41), an optimization calculation unit (42), and a feedback unit (43). The task detection unit (41) detects and collects work data from both human and collaborative robots; The optimization calculation unit (42) optimizes the weights of tasks of the same type based on the working data; The feedback unit (43) sends the optimization calculation results to the task static matching module to perform matching optimization for unmatched tasks.

6. A method for intelligent matching of work tasks for human-machine collaboration, used in the intelligent matching system for work tasks as described in any one of claims 1-5, characterized in that, Includes the following steps: Step 1, Task splitting module (1) splits the task into a series and further breaks it down into atomic tasks; Step 2, Proxy Evaluation Module (2) evaluates and judges the human-machine proxy rights of the atomic task; Step 3: The static task allocation module (3) performs task matching and allocation based on the proxy rights and the actual content of the atomic tasks; Step 4: Task Dynamic Optimization Module (4) Real-time collection of work data and optimization allocation of unassigned tasks based on work data.

7. The intelligent task matching method according to claim 6, characterized in that: Step 2 above also includes the following steps: Step 2.1, Task Feature Unit (21) performs feature extraction and analysis based on the task content; Step 2.2, Task Risk Unit (22) evaluates the risks to human personnel during task execution; Step 2.3, Task security unit (23) evaluates the security of the task execution process and results; Step 2.4: Based on the above analysis and evaluation, the hierarchical analysis unit constructs a hierarchical analysis model to calculate the proxy weights; Step 2.5, the proxy calculation unit (25) calculates the proxy rights of each atomic task based on the proxy weight.

8. The intelligent task matching method according to claim 7, characterized in that: Step 2.4 above also includes the following steps: Step 2.4.1: Construct the hierarchical analysis model using the hierarchical analysis model component (241); Step 2.4.1: The judgment matrix component (242) generates a judgment model based on historical data and in combination with each layer of the hierarchical analysis model; Step 2.4.1: The verification component (243) calculates the local weights by performing a sum-product operation on the judgment matrix and verifies consistency. Step 2.4.1: If the consistency check fails, make manual adjustments based on historical data until the check passes. Step 2.4.1: After the verification is passed, the global weight calculation component (244) will calculate the global matrix by combining the local weights.

9. The intelligent task matching method according to claim 6, characterized in that: Step 3 above also includes the following steps: Step 3.1: The proxy matching unit (31) obtains the proxy calculation results and classifies the atomic tasks accordingly; Step 3.2: The priority matching unit (32) performs priority matching on the classified atomic tasks to determine the task allocation order; Step 3.3, the workload matching unit (33) acquires artificial data and matches it with the workload of each task to ensure the physiological and mental health of each artificial person; Step 3.4: Task allocation unit (34) obtains multivariate matching data to complete task allocation.

10. The intelligent task matching method according to claim 6, characterized in that: Step 4 above also includes the following steps: Step 4.1 After the task begins, the task detection unit (41) detects and collects the work data of each human and collaborative robot; Step 4.2: The optimization calculation unit (42) acquires working data and connects it to the hierarchical analysis unit (24) to adjust and calculate the model; Step 4.3: The feedback component sends the proxy rights to the proxy rights matching unit (31); Step 4.4: The task allocation unit (34) reallocates and adjusts the unassigned tasks.

Images