Method and device for generating shared knowledge between human and machine in human-machine collaborative task, and medium

Abstract

The application discloses a human-machine collaborative task shared knowledge generation method and device and a medium. The method comprises the following steps: human-machine knowledge difference analysis; human-machine knowledge sharing criterion establishment; shared knowledge graph construction; and shared knowledge generation. The shared knowledge generation method can generate the knowledge required for each Agent in a multi-Agent team to make corresponding decisions, can effectively improve information transparency, can make the method have explainability, can effectively improve the trust degree of a human, can reduce the operation load of the human, and can efficiently and reasonably complete the human-machine interaction and collaboration process.

Description

Technical Field

[0001] This invention relates to the field of human-machine collaboration technology, specifically to a method, device, and medium for generating shared knowledge between humans and machines in human-machine collaborative tasks. Background Technology

[0002] With the rapid development of automation and robotics, human-machine collaborative operations, especially how to effectively integrate human intelligence with machine intelligence and mechanical efficiency to collaboratively complete typical tasks such as grasping and assembly, have always been a research focus for relevant scholars.

[0003] Existing technical methods mainly fall into two categories: reconstructing the robot's reasoning model and adding decision-making results to the robot. Reconstructing the robot's reasoning model primarily refers to modeling the psychological state of the human during a task. The robot infers the human's intentions by observing their behavior. If the human's behavior is outside the specified range, the robot updates its task plan and mental model to adapt to the task changes. This method has advantages such as high task matching and relatively low human workload, but it also has disadvantages such as poor interpretability and low human trust. Examples include a robot human-machine interaction method based on human behavior recognition and perception disclosed in Chinese Patent Publication No. CN114131610A on March 4, 2022, and a knowledge learning and transfer method for epidemic prevention robots disclosed in Chinese Patent Publication No. CN112231489A on January 15, 2021.

[0004] Enhancing the interpretability of robot decision-making outcomes is primarily achieved by assessing the interpretability of these outcomes across three axes: information abstraction, communication specificity, and communication comprehensiveness. Each interpretation originates from these three axes, selecting an appropriate position within each axis to form the most reasonable explanation. This method introduces a system that provides explanations for robot decision-making outcomes, allowing humans to understand the internal workings of how robots make decisions. However, it only allows robots to acquire human knowledge, not the other way around, preventing the sharing of human and machine knowledge and resulting in insufficient task completion efficiency. Examples include Chinese Patent CN114131602A, published on March 4, 2022, which discloses a method for accumulating robot operation knowledge in a human-machine collaborative curtain wall installation environment, and Chinese Patent CN106886408A, published on June 23, 2017, which discloses an intelligent decision-making system applied to robot software engineers.

[0005] Therefore, for human-machine collaborative tasks, it is necessary to design more efficient and reasonable methods for generating shared knowledge to fully leverage the advantages of humans and robots, avoid the risks to tasks and personnel caused by complex environments and uncertainties, and improve task performance. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, the present invention provides a method, device and medium for generating human-machine shared knowledge in human-machine collaborative tasks, in order to solve at least one of the above-mentioned technical problems.

[0007] According to one aspect of the present invention, a method for generating human-machine shared knowledge in a human-machine collaborative task is provided, comprising:

[0008] Human-machine knowledge difference analysis is used to analyze the knowledge of both humans and machines according to their strategies, information needs, and information representation methods, based on the human-centered principle of human-machine collaboration, to find the differences and determine the parts that can be converted.

[0009] The human-machine knowledge sharing principle is established to establish corresponding connections between human and robot knowledge based on the criteria of predictability, observability and instructibility, and to find corresponding knowledge conversion methods.

[0010] The shared knowledge graph construction is used to construct a shared knowledge graph based on the aforementioned analysis of human-machine knowledge differences and the establishment of human-machine knowledge sharing criteria.

[0011] Shared knowledge generation is used to search for relevant human and robot information needs and corresponding conversion methods in the constructed shared knowledge graph when the robot is unable to complete the task during task execution. The corresponding functions are then invoked in the system to complete the generation of shared knowledge.

[0012] The above technical solution adopts the human factors task analysis paradigm, combining the differences in knowledge required for humans and robots to make the same decisions in different tasks. Based on the knowledge required for different individuals to make the same decisions, a human-machine shared knowledge generation method is proposed, which enables humans and robots to have good task-level collaboration, functional-level coordination, and mutual understanding in communication. By determining the differences, commonalities, and convertibility of the knowledge required for humans and robots to perform tasks in real time during the task process, the possibility of human-machine collaboration is determined, so as to more fully generate shared knowledge in human-machine collaborative tasks, so as to achieve more reasonable collaboration between humans and robots to complete typical tasks such as grasping and path planning.

[0013] As a further technical solution, human-machine knowledge difference analysis also includes: analyzing the strategies of humans and robots; analyzing the information needs of humans and robots in the strategy stage where they can cooperate; and analyzing the differences in information representation methods between humans and robots under similar information needs.

[0014] The difference in human-machine knowledge is the reason why they cannot share knowledge. When analyzing this difference, the first step is to analyze the strategies of both humans and robots. Strategies determine the goals that humans and robots aim to achieve, and these goals form the basis for collaboration between the two. By analyzing the differences in their strategies, we can determine whether humans and robots can collaborate at a given moment. Collaboration is a prerequisite for knowledge sharing.

[0015] Furthermore, we analyze the information required by both in the strategy phase where they can collaborate. Due to the structural differences between humans and robots, their information needs differ in each strategy. If the needs of the collaborator can be clearly understood during the collaboration process, and targeted suggestions can be given, the cognitive load on humans can be effectively reduced.

[0016] Furthermore, we analyze the differences in information representation methods between humans and robots when faced with similar information needs. These differences in information representation methods are one of the reasons why humans and robots cannot understand each other. By analyzing these differences, we can identify the parts that can be converted between them.

[0017] Optionally, based on the criteria of predictability, observability, and instructibility between human and machine knowledge, corresponding connections are established between human and robot knowledge, and methods for converting between these knowledge types are sought accordingly. Specifically:

[0018] Human-robot situational awareness analysis. For humans and robots to achieve collaboration, predictability is essential; that is, their goals and intentions must be mutually known. To meet this predictability requirement, before strategy sharing, the robot, as an intelligent agent, should possess the ability to understand its own state and purpose.

[0019] Furthermore, this capability is constructed using the BDI model. B refers to the robot's current state, obtained by perceiving information about task changes during task execution; D refers to the robot's desired outcome of performing the task; and I refers to the robot's proposed action plan from the current state to the target state. In other words, each step of the robot's strategy has a starting state and a ending state. When the robot performs a task, if the state of B cannot reach the state of D, or if action plan I cannot be found, it will request human assistance. At this point, the relationship between the human and the robot will be predictable.

[0020] Furthermore, there should be observability and instructability between human and robot knowledge. Observability means that the robot's knowledge should be usable by humans, while instructability means that humans should be able to use their own knowledge to guide the robot.

[0021] Furthermore, achieving observability and instructibility requires meeting corresponding information needs and transforming information representation methods.

[0022] As a further technical solution, the constructed shared knowledge graph includes four ontology libraries: human knowledge, robot knowledge, task ontology, and transformation modules. During task execution, the shared knowledge graph can search for human and robot knowledge, find transformation methods through directed relationships, and thus generate shared knowledge.

[0023] Optionally, neo4j can be used to build a computer model of the knowledge base and relationships for the system to use.

[0024] As a further technical solution, the method also includes:

[0025] During the execution of human-robot collaborative tasks, when the robot is unable to solve the corresponding problem on its own, it uses BDI to determine its own strategic state.

[0026] Based on the robot's strategy information, search for the corresponding target information in the knowledge graph;

[0027] To search whether the corresponding purpose information exists within human knowledge.

[0028] As a further technical solution, if the corresponding target information exists within human knowledge, then...

[0029] Based on people's strategic information, identify the information needs under the corresponding strategies;

[0030] Based on the information needs under human strategies, we can determine whether there are observable and instructable relationships between human and robot knowledge, and then judge whether shared knowledge can be generated.

[0031] As a further technical solution, based on the ability to generate shared knowledge, we can find corresponding conversion methods.

[0032] As a further technical solution, the method also includes: after finding all the information, searching for the corresponding conversion function in the system, and inputting the corresponding knowledge into the function to complete the generation of shared knowledge.

[0033] According to one aspect of the present invention, a computer device is provided, the computer device including a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, it implements the steps of the human-computer shared knowledge generation method in the human-computer collaborative task.

[0034] According to one aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, wherein when the computer program is executed by a processor, the steps of the human-machine collaborative task knowledge sharing generation method are implemented.

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0036] (1) The shared knowledge generation method proposed in this invention can generate the knowledge required for each agent in a multi-agent team to make corresponding decisions. While meeting the needs of the task, it can effectively improve the transparency of information, make it explainable, thereby effectively improving people's trust, reducing the workload of people, and efficiently and reasonably completing the human-computer interaction and collaboration process.

[0037] (2) This invention uses predictability, observability and shareability to sort out the relationship between human and machine knowledge, takes into account the differences between human and machine, and conducts a difference analysis of human and machine knowledge. It uses a shared knowledge graph to build a knowledge sharing model, which can better integrate system resources and improve the speed of knowledge conversion during the process of shared knowledge generation, thereby effectively improving task performance. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of a human-machine collaborative knowledge generation method in a crawling task according to an embodiment of the present invention.

[0039] Figure 2 This is a diagram illustrating the similarities and differences in human-machine strategies according to an embodiment of the present invention.

[0040] Figure 3 This is a comparison diagram of information demand under the human-machine correspondence strategy according to an embodiment of the present invention.

[0041] Figure 4 This is a framework diagram of a human-machine collaborative system according to an embodiment of the present invention.

[0042] Figure 5 This is a flowchart of the BDI model according to an embodiment of the present invention.

[0043] Figure 6 This is a diagram illustrating the observable and guideable relationship between human-machine collaborative knowledge capture according to an embodiment of the present invention.

[0044] Figure 7 A model diagram of the body designed according to a preferred embodiment of the present invention.

[0045] Figure 8 The figure illustrates a preferred example of the knowledge extraction method according to the present invention.

[0046] Figure 9 A knowledge graph constructed according to a preferred embodiment of the present invention.

[0047] Figure 10 A flowchart for generating shared knowledge according to a preferred example of the present invention.

[0048] Figure 11This is a schematic diagram of the structure of a computing device according to an embodiment of the present invention. Detailed Implementation

[0049] The technical solutions of various 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.

[0050] Please see Figure 1 This invention is a shared knowledge generation method based on the knowledge differences between humans and robots. Taking a grasping task as an example, the specific process is as follows: First, the knowledge of humans and robots in performing tasks in the grasping task is analyzed separately. For humans, a cognitive task analysis based on the human factor paradigm is performed, and for robots, a meta-analysis of robot task strategies and methods is performed. Then, knowledge trees of humans and robots are generated in turn. Then, through difference and sharing analysis, encoding is performed, and finally a human-machine shareable knowledge tree is generated. This tree is then applied to human-machine collaborative tasks in contexts where human-machine shared knowledge can be generated, so as to effectively carry out human-machine collaboration.

[0051] Please see Figure 2 In the optimal grasping task, this paper first needs to obtain the strategies used by humans and robots in the grasping task, as shown in Tables 1 and 2; the information requirements in the task, as shown in Tables 3 and 4; and the representation methods of relevant information, as shown in Tables 5 and 6. Based on the human-centered principle, the differences between humans and robots in terms of strategies, information requirements, and information representation methods are analyzed to identify the differences.

[0052] Furthermore, the differences in strategies between humans and robots are analyzed. All collective actions are based on a common purpose, which is reflected in strategies. The strategic purposes of humans and robots in performing grasping tasks are laid out according to the task execution sequence. Figure 2 In the middle, the two ends of the double arrows represent that the goals to be achieved by the human and the robot are consistent under the corresponding strategies, and the corresponding human-machine strategies can be shared; otherwise, the goals are determined to be inconsistent.

[0053] Table 1. Human Scraping Strategies and Objectives

[0054]

[0055] Table 2 Robot Grasping Strategies and Objectives

[0056]

[0057]

[0058] Table 3 Information Needs During the Data Acquisition Process

[0059]

[0060] Table 4 Information Requirements During Robot Grasping

[0061]

[0062]

[0063] Table 5. Information Demand Attributes and Representation Methods During the Human Data Acquisition Process

[0064]

[0065] Table 6 Information Demand Attributes and Representation Methods During Robot Grasping Process

[0066]

[0067]

[0068] Please see Figure 3 Under the shareable strategy, the information needs of humans and robots at each strategy stage are aggregated. Figure 3 The left side shows the information requirements of a person under each strategy, while the right side shows the information requirements of a robot under each strategy. When the goals of humans and robots are aligned, their decision-making outputs contain the same type of information, but the required input information differs significantly.

[0069] Based on the summaries in Tables 5 and 6, Table 7 can be obtained. By analyzing the attributes and representation methods of the information required by both humans and robots, it can be seen that the information attributes required by both are color, position, shape, posture, and characters.

[0070] Table 7 Comparison of Human Representation Methods and Robot Representation Methods

[0071]

[0072] Please see Figure 4 Based on the analysis of the differences between human and machine knowledge, a human-machine collaborative system framework can be established to establish connections between human and machine knowledge and to find corresponding knowledge conversion methods.

[0073] Specifically, for effective human-robot collaboration to be successful, the knowledge shared by both humans and robots needs to be observable, predictable, and shareable. Observability means that others can know what the robot knows, including its understanding of the team, the task, and the environment. Predictability means that an individual's behavior should be sufficiently predictable, ensuring that team members can understand their intentions. Directability means that team members possess the ability to guide others. In the grasping task, the robot, as the primary executor, initiates the questioning. For humans and robots to collaborate, their knowledge should be predictable. As the on-site operator, the robot's perception of environmental information is superior to that of a human; therefore, its knowledge should be observable to humans. Humans, as assistants, guide the robot in performing the task; therefore, their knowledge should be directable.

[0074] In this context, human-machine knowledge must be predictable; the robot, as the initiator of the question, needs to know its own current state beforehand. This paper uses the BDI model to build this capability for the robot.

[0075] Please see Figure 5 In the BDI model, B refers to the robot's current state, obtained by perceiving changes in the task during its execution; D refers to the robot's desired outcome of the task; and I refers to the robot's proposed action plan from the current state to the target state. In other words, each step of the robot's strategy has a starting state and a ending state. When the robot cannot find an action plan I, or when B consistently fails to reach state D, it will request human assistance. In this case, the robot shares its strategy and objective information with the human to ensure predictability.

[0076] Please see Figure 6 The system determines whether the robot's knowledge satisfies observability and whether the human's knowledge satisfies instructibility, requiring analysis from two aspects: the attributes of the information and the object of the information. When both are consistent, the conditions are met. At this point, the necessary connections between the two sets of knowledge can be established. Once a relationship exists between human and machine knowledge, corresponding functions are needed to achieve sharing and conversion. Constructing these conversion functions will then complete the knowledge conversion.

[0077] Please see Figure 7 The shared knowledge graph ontology model includes human knowledge ontology, robot knowledge ontology, task ontology, and conversion module ontology. Based on the analysis of human-machine knowledge differences and human-machine knowledge relationships in the previous text, corresponding knowledge relationships can be constructed between the ontology and the ontology. The human-machine knowledge shared knowledge graph relationship is summarized in Table 8.

[0078] Table 8. Ontology Relationship Design

[0079]

[0080]

[0081] Please see Figure 8 At this point, the knowledge ontology can be organized into a corresponding triplet table for knowledge extraction from humans or robots. Human knowledge ontology includes two main categories: strategies and information needs, each with corresponding attribute information. Figure 8 Table 1 allows us to extract the strategy names (1) and the objective attributes (2) for each strategy. Furthermore, the strategies exhibit strong correlations, and based on the human knowledge ontology framework, we will also extract the strategy beliefs (3). Figure 8 Table 3 shows that information perception 5 and decision output 6 can be extracted and used as entity nodes in the knowledge graph. Furthermore, both information perception and decision output serve corresponding policies, thus the policies have a calling relationship with the corresponding information 4. Figure 8 Table 5 shows that the attributes and representation methods of the corresponding information can be extracted, which are the attribute information of the information perception and decision output instances.

[0082] Please see Figure 9 Based on the above extraction method, instance knowledge of the four types of ontology is extracted, and the knowledge graph is constructed using Neo4j to obtain the final human-computer crawling knowledge graph used in the system.

[0083] Please see Figure 10 Once the knowledge graph is constructed, it is integrated into the system along with related functions. Upon appropriate triggering, shared knowledge can be generated, resulting in the overall flowchart of this method. Specifically:

[0084] 1) First, the predictability requirement needs to be met.

[0085] In collaborative human-robot grasping in an unknown environment, the robot operates in the field and acts as the primary executor of the task. During task execution, the robot continuously updates and determines its current state (B) and the desired objective (D) using the BDI model. If an executable solution (I) cannot be found, or if B cannot reach D, the robot requests human assistance. Consistent objectives are the foundation of collaboration, and these objectives are embodied in the strategy. Based on predictability requirements, the robot displays its current state (B) and objective (D) to the human, enabling the human to understand the robot's status.

[0086] After the robot initiates a request, the system matches it with a strategy phase where the human would have the same objective when performing the task alone. If this condition is not met, the robot will be unable to collaborate with the human to complete the task or share the necessary knowledge, and the system will determine that the human must enter the command-level control mode to complete the task.

[0087] 2) After meeting the predictability requirement, observability needs to be met.

[0088] When a human has a strategy with the same objective as the robot at the current stage, the information required by the human and robot differs at their respective strategy stages. Therefore, accurate information recommendation is essential for effective collaboration. At this point, the system can use the human's acquired knowledge to determine the information needed under the current strategy. Simultaneously, it matches the robot's own information with information possessing the same attributes and objects as the required information, and searches for corresponding conversion methods to meet observability requirements and achieve knowledge sharing. If the robot cannot share matching information or lacks the corresponding conversion method, the information is processed by the human. If no match can be found, it is determined that shared knowledge cannot be generated, and the human then enters command-level control mode to complete the task.

[0089] 3) After meeting the observability requirement, the instructibility requirement needs to be met.

[0090] After acquiring shared knowledge from the robot, the human adds their own processed knowledge to make decisions. Once the human has made a decision, they need to output the result to the robot to provide guidance. Based on the previous analysis, under the premise of shared goals, the attributes of the information output by the human and machine will remain consistent with the object. Therefore, finding a suitable conversion method to transform the way human-machine knowledge is represented can satisfy the requirement of instructibility. If no suitable conversion method exists, it is determined that the generation of shared knowledge cannot be completed, and the human will subsequently enter a command-level control mode to complete the task. After completing the above three steps, shared knowledge will be generated, and the human and machine will collaborate to complete the task based on this knowledge.

[0091] Specifically, in the scenarios described above where shared knowledge generation cannot be completed, on the one hand, a human operator will take over the system to provide command-level control to the robot to complete the remaining tasks; on the other hand, to assist the human in decision-making, the system will also transmit information such as the human-robot related strategies and objectives that are currently unable to complete shared knowledge generation to the human operator. Furthermore, this information can be used to subsequently design new strategies and objectives for the human-robot team and integrate them into the human-robot shared knowledge generation method in the preferred examples described above. In other words, this method can be continuously improved and summarized, exhibiting scalability.

[0092] According to one aspect of the present invention, a computer device is provided, which may be an industrial control computer, a server, or a computer terminal.

[0093] The computer device includes a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, it implements the steps of the human-machine collaborative task human-machine shared knowledge generation method.

[0094] The computer device includes a processor, memory, and network interface connected via a system bus, wherein the memory may include non-volatile storage media and internal memory.

[0095] Non-volatile storage media can store operating systems and computer programs. These computer programs include program instructions that, when executed, cause the processor to perform any human-machine collaborative task involving the generation of shared knowledge.

[0096] The processor provides computing and control capabilities, supporting the operation of the entire computer device.

[0097] Internal memory provides an environment for the execution of computer programs stored in non-volatile storage media. When the computer program is executed by the processor, it enables the processor to execute any human-computer collaborative task's method for generating shared knowledge.

[0098] This network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 11 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0099] It should be understood that the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, a general-purpose processor can be a microprocessor or any conventional processor.

[0100] In one embodiment, the processor is configured to run a computer program stored in memory to perform the following steps:

[0101] Human-machine knowledge difference analysis is used to analyze the knowledge of both humans and machines according to their strategies, information needs, and information representation methods, based on the human-centered principle of human-machine collaboration, to find the differences and determine the parts that can be converted.

[0102] The human-machine knowledge sharing principle is established to establish corresponding connections between human and robot knowledge based on the criteria of predictability, observability and instructibility, and to find corresponding knowledge conversion methods.

[0103] The shared knowledge graph construction is used to construct a shared knowledge graph based on the aforementioned analysis of human-machine knowledge differences and the establishment of human-machine knowledge sharing criteria.

[0104] Shared knowledge generation is used to search for relevant human and robot information needs and corresponding conversion methods in the constructed shared knowledge graph when the robot is unable to complete the task during task execution. The corresponding functions are then invoked in the system to complete the generation of shared knowledge.

[0105] According to one aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, wherein when the computer program is executed by a processor, the steps of the human-machine collaborative task knowledge sharing generation method are implemented.

[0106] The computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiments, such as the hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, SmartMediaCard (SMC), Secure Digital (SD) card, or Flash Card equipped on the computer device.

[0107] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A method for generating shared knowledge between humans and machines in human-machine collaborative tasks, characterized in that, include: Human-machine knowledge difference analysis is used to analyze the knowledge of both humans and machines according to strategy, information needs, and information representation methods, based on the human-centered principle of human-machine collaboration, to find the differences and identify the parts that can be converted; during the execution of human-machine collaborative tasks, when the robot cannot solve the corresponding problem on its own, it determines its own strategy state through BDI; and searches for the corresponding target information in the knowledge graph based on the robot's strategy information. To search whether the relevant information for the purpose is available in human knowledge; If the corresponding purpose information is available in human knowledge, then the information needs under the corresponding strategy can be found based on the human strategy information. Based on the information needs under human strategies, we can find out whether there are observable and instructable relationships between human and robot knowledge, and then determine whether shared knowledge can be generated. The human-machine knowledge sharing principle is established to establish corresponding connections between human and robot knowledge based on the criteria of predictability, observability and instructibility, and to find corresponding conversion methods between knowledge. The shared knowledge graph construction is used to construct a shared knowledge graph based on the aforementioned analysis of human-machine knowledge differences and the establishment of human-machine knowledge sharing criteria. Shared knowledge generation is used to search for relevant human and robot information needs and corresponding conversion methods in the constructed shared knowledge graph when the robot is unable to complete the task during task execution. The corresponding functions are then invoked in the system to complete the generation of shared knowledge.

2. The method for generating shared human-machine knowledge in a human-machine collaborative task according to claim 1, characterized in that, Human-machine knowledge difference analysis also includes: analyzing the strategies of humans and robots; analyzing the information needs of humans and robots in the stage of strategy that enables collaboration; and analyzing the differences in information representation methods between humans and robots under similar information needs.

3. The method for generating shared human-machine knowledge in a human-machine collaborative task according to claim 1, characterized in that, The constructed shared knowledge graph includes four ontology libraries: human knowledge, robot knowledge, task ontology, and transformation module.

4. The method for generating shared human-machine knowledge in a human-machine collaborative task according to claim 1, characterized in that, Based on the ability to generate shared knowledge, we can find corresponding transformation methods.

5. The method for generating shared human-machine knowledge in a human-machine collaborative task according to claim 4, characterized in that, The method further includes: after finding all the information, searching for the corresponding conversion function in the system, and inputting the corresponding knowledge into the function to complete the generation of shared knowledge.

6. A computer device, characterized in that, The computer device includes a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, it implements the steps of the human-machine collaborative task human-machine shared knowledge generation method as described in any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, it implements the steps of the human-machine collaborative task human-machine shared knowledge generation method as described in any one of claims 1 to 5.

Images