Coal mining intelligent interaction integration system based on large language model

Abstract

The invention provides a coal mining intelligent interaction integration system based on a large language model, and belongs to the technical field of intelligent interaction. The coal mining intelligent interaction integration system based on the large language model comprises a human-computer interface assembly used for responding to a first request of a target user and sending the first request to a large language model assembly; the large-scale language model component generates a first insight according to the first request and sends the first insight to the large-scale language model agent component; the large language model agent component is used for sending the first insight to the knowledge graph component; the knowledge graph component is used for carrying out data query according to the first insight to obtain first data and sending the first data to the large language model agent component; and the robot assistant assembly is used for executing a distribution task, and the distribution task is a mining task. The problem that the coal mining operation efficiency is low can be solved.

Description

Technical Field

[0001] The present disclosure relates to the field of intelligent interaction technology, and in particular to a coal mining intelligent interaction integrated system based on a large language model. Background Art

[0002] With the rapid development of information technology, the demand for intelligent systems in the field of coal mining is becoming increasingly urgent. The coal mining industry has always faced many challenges, including safety hazards, environmental impacts, and low operational efficiency. In the traditional coal mining process, there are problems such as untimely information processing and lack of scientific basis for decision-making. At the same time, each link of coal mining often relies on manual experience for operation and judgment, and the efficiency of data analysis and integration is low. Therefore, the existing coal mining has the problem of low operating efficiency. Summary of the invention

[0003] The disclosed embodiment provides a coal mining intelligent interactive integration system based on a large language model to solve the problem of low operating efficiency.

[0004] The embodiment of the present disclosure provides a coal mining intelligent interactive integration system based on a large language model, including: Human-computer interface component, large language model component, large language model agent component, knowledge graph component and robot assistant component; A human-computer interface component, configured to respond to a first request from a target user and send the first request to a large language model component; The large language model component generates a first insight based on the first request and sends the first insight to the large language model proxy component; A large language model proxy component that sends first insights to the knowledge graph component; A knowledge graph component, configured to perform a data query based on the first insight, obtain first data, and send the first data to the large language model proxy component; The large language model proxy component is further used to generate an assignment task based on the first data and send the assignment task to the robot assistant component; The robot assistant component is used to perform the assigned task, and the assigned task is a mining task.

[0005] In an exemplary embodiment of the present disclosure, a large language model-based intelligent interactive integrated system for coal mining further includes a retrieval enhancement generation component; The large language model proxy component is further used to send the first data to the retrieval enhancement generation component; The retrieval enhancement generation component is used to generate a query response based on the first data and send the query response to the large language model proxy component.

[0006] In an exemplary embodiment of the present disclosure, a coal mining intelligent interactive integrated system based on a large language model further includes an environment modeling component: The environmental modeling component is connected to the large language model agent component and the knowledge graph component respectively, and is used to collect environmental data of the mining area and generate a high-fidelity environmental model of the mining area based on the environmental data; The environment modeling component is also used to update the environment data to the knowledge graph component.

[0007] In an exemplary embodiment of the present disclosure, the robot assistant component is further used to collect second data and send the second data to the large language model component; The large language model component is also used to process and analyze the second data for problem detection; The large language model proxy component is also used to adjust task allocation and update the environment modeling component after problem detection is completed.

[0008] In an exemplary embodiment of the present disclosure, the knowledge graph component is also used to analyze data and identify faults, obtain fault results, and send the fault results to the retrieval enhancement generation component; The retrieval enhancement generation component is also used to generate contextual responses and send the contextual responses to the large language model proxy component; A large language model agent component is also used to make recommendations.

[0009] In an exemplary embodiment of the present disclosure, the large language model proxy component is specifically used to: third data responsive to the large language model component; Analysis is performed based on the third data and tasks are assigned to the robot assistant components.

[0010] In an exemplary embodiment of the present disclosure, the human-machine interface component is used to receive a first request from a target user; The large language model component is also used to assign tasks to the robot assistant component based on the large language model proxy component in response to the first request; The HMI component is also used to display the updated task status.

[0011] In an exemplary embodiment of the present disclosure, the human-machine interface component is also used to display the mining environment.

[0012] In an exemplary embodiment of the present disclosure, the knowledge graph component includes a knowledge base, which includes equipment data, geological data, safety protocols, and historical performance indicators.

[0013] In one exemplary embodiment of the present disclosure, the robotic assistant assembly is also used to navigate a mining environment.

[0014] The beneficial effects of the coal mining intelligent interactive integrated system based on a large language model provided by the embodiment of the present disclosure are: The disclosed embodiments greatly improve the intelligence and efficiency of coal mining. The disclosed embodiments achieve seamless connection and intelligent scheduling from user needs to actual mining operations by integrating various components. The target user only needs to express the needs through natural language, and the system can quickly understand and generate effective insights, rely on the knowledge graph component to accurately query relevant data, and then intelligently assign mining tasks to the robot component for execution. The disclosed embodiments not only simplify the operating process and reduce human errors, but also significantly improve the response speed and operation accuracy, ensuring the safety and efficiency of coal mining. Therefore, the disclosed embodiments solve the problem of low efficiency of coal mining operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0016] Figure 1 It is a structural schematic diagram of a coal mining intelligent interactive integrated system based on a large language model provided by an embodiment of the present disclosure; Figure 2 It is a structural schematic diagram of another coal mining intelligent interactive integrated system based on a large language model provided by an embodiment of the present disclosure; Figure 3 It is an architecture diagram of a coal mining intelligent interactive integrated system based on a large language model provided by an embodiment of the present disclosure; Figure 4 is a flowchart of real-time monitoring and data collection provided by an embodiment of the present disclosure; Figure 5 is a schematic diagram of the mechanism of predictive maintenance and decision support provided by an embodiment of the present disclosure; Figure 6 is a schematic diagram of the autonomous operation of the robot assistant provided by an embodiment of the present disclosure; Figure 7 It is a process diagram of the task management and coordination functions of a large language model proxy component provided by an embodiment of the present disclosure; Figure 8 is a schematic diagram of a process of using 3D Gaussian sputtering to perform environmental modeling according to an embodiment of the present disclosure; Fig. 9 It is a sequence diagram of a coal mining intelligent interactive integration system based on a large language model provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0017] In order to enable people in the technical field to better understand the present solution, the technical solution in the embodiment of the present solution will be clearly described below in conjunction with the drawings in the embodiment of the present solution. Obviously, the described embodiment is an embodiment of a part of the present solution, not all of the embodiments. Based on the embodiments in the present solution, all other embodiments obtained by ordinary technicians in the field without creative work should fall within the scope of protection of the present solution.

[0018] The term "including" and any other variations in the specification and claims of this solution and the above drawings mean "including but not limited to", and is intended to cover non-exclusive inclusions and is not limited to the examples listed in the text. In addition, the terms "first" and "second" are used to distinguish different objects, not to describe a specific order.

[0019] The following is a detailed description of the implementation of the present disclosure in conjunction with the specific drawings: Figure 1 A schematic diagram of the structure of a coal mining intelligent interactive integrated system based on a large language model provided by an embodiment of the present disclosure. Figure 1 , the coal mining intelligent interactive integrated system based on a large language model includes: Human-computer interface component, large language model component, large language model agent component, knowledge graph component and robot assistant component; A human-computer interface component, configured to respond to a first request from a target user and send the first request to a large language model component; The large language model component generates a first insight based on the first request and sends the first insight to the large language model proxy component; A large language model proxy component that sends first insights to the knowledge graph component; A knowledge graph component, configured to perform a data query based on the first insight, obtain first data, and send the first data to the large language model proxy component; The large language model proxy component is further used to generate an assignment task based on the first data and send the assignment task to the robot assistant component; The robot assistant component is used to perform the assigned task, and the assigned task is a mining task.

[0020] In this embodiment, the human-computer interface component is connected to the large language model component, the large language model component is connected to the large language model proxy component, and the large language model proxy component is respectively connected to the knowledge graph component and the robot assistant component.

[0021] Target users as operators may include but are not limited to engineers and managers.

[0022] The human-machine interface component enables the interaction between the target user and each component of the system, and can provide real-time visualization of the mining environment, task status, and system performance to enhance situational awareness. For example, in a coal mining scenario, the target user (such as an operator such as an engineer or manager) can make a first request such as a demand, question, or operation instruction through the human-machine interface component. The first request can be a command sent by the target user through the human-machine interface component.

[0023] The large language model component uses advanced natural language processing to interpret commands, analyze technical documents, and process real-time data streams; it provides background insights and suggestions for target users and robot assistant components; it can facilitate communication between the system components and enhance the decision-making process. After receiving the first request sent by the human-machine interface component, the large language model component uses the language understanding and generation capabilities obtained through training based on a large amount of data to analyze and understand the first request and generate the corresponding first insights. The first insights can be insights processed by the large language model component.

[0024] The large language model proxy component acts as an intermediary to manage and coordinate tasks between the large language model component, knowledge graph component, retrieval enhancement generation component, and robot assistant component; ensures optimal resource allocation and efficient operation through adaptive learning and task management; and continuously improves the performance of the large language model proxy component based on feedback and operational data.

[0025] Exemplarily, on the one hand, the large language model proxy component passes the first insights generated by the large language model component to the knowledge graph component so as to perform a deeper data query in the knowledge graph; on the other hand, the large language model proxy component receives the first data returned by the knowledge graph component, and generates a distribution task based on the data and sends it to the robot assistant component. The first data may be the returned data after the data query.

[0026] The knowledge graph component builds a comprehensive representation of mining operations, including equipment, geological data, safety protocols, and historical performance indicators; it can efficiently retrieve and organize relevant information to provide decision support and predictive analysis; and it enhances the system’s ability to identify patterns, trends, and potential problems in real time.

[0027] For example, the knowledge graph component constructs a knowledge network in the field of coal mining, storing a large amount of information related to coal mining in a structured manner, including geological conditions, equipment information, mining processes, safety regulations, etc. After receiving the first insights from the large language model proxy component, the knowledge graph component performs data queries in the above knowledge network to mine more detailed and accurate first data.

[0028] The robotic assistant component uses an autonomous robotic assistant equipped with sensors, cameras, and artificial intelligence algorithms to perform various mining tasks; it is capable of navigating in challenging environments, operating machines, and performing inspection and maintenance activities; and it is integrated with the large language model component, the large language model agent component, and the knowledge graph component to perform complex tasks with minimal human intervention.

[0029] For example, the robot assistant component is specifically responsible for executing the assigned task, and the task is a mining task directly related to coal mining. After receiving the assigned task sent by the large language model agent component, the robot assistant can perform actual operations through hardware such as mechanical arms and mining equipment.

[0030] It can be concluded from the above that this embodiment greatly improves the intelligence and efficiency of coal mining. This embodiment realizes seamless connection and intelligent scheduling from user needs to actual mining operations by integrating various components. The target user only needs to put forward the demand through natural language, and the system can quickly understand and generate effective insights, rely on the knowledge graph component to accurately query relevant data, and then intelligently assign mining tasks to the robot component for execution. This embodiment not only simplifies the operation process and reduces human errors, but also significantly improves the response speed and operation accuracy, ensuring the safety and efficiency of coal mining. Therefore, this embodiment solves the problem of low efficiency of coal mining operations.

[0031] In one embodiment of the present disclosure, reference Figure 2 ,An intelligent interactive integrated system for coal mining based on a large language model also includes a retrieval enhancement generation component; The large language model proxy component is further used to send the first data to the retrieval enhancement generation component; The retrieval enhancement generation component is used to generate a query response based on the first data and send the query response to the large language model proxy component.

[0032] In this embodiment, the retrieval enhancement generation component is connected to the large language model proxy component, and the large language model proxy component is combined with the retrieval enhancement generation component to generate accurate and context-relevant query responses; support dynamic and real-time retrieval of information from the knowledge graph component and external data sources; provide precise and actionable insights to improve decision-making and problem solving.

[0033] When the large language model proxy component sends the first data obtained from the knowledge graph component, the search enhancement generation component starts working. This embodiment provides more comprehensive and accurate information supplement for the system by further processing the first data. Based on the received first data, the search enhancement generation component uses its own algorithm and stored information library for analysis.

[0034] From the above, we can conclude that the addition of the retrieval enhancement generation component enriches the depth and breadth of the system's information processing, while also providing more reference information for the large language model component, making the system more scientific and reasonable in generating insights and formulating tasks.

[0035] In one embodiment of the present disclosure, reference Figure 2 and Figure 8 , an intelligent interactive integrated system for coal mining based on a large language model also includes an environmental modeling component: The environmental modeling component is connected to the large language model agent component and the knowledge graph component respectively, and is used to collect environmental data of the mining area and generate a high-fidelity environmental model of the mining area based on the environmental data; The environment modeling component is also used to update the environment data to the knowledge graph component.

[0036] In this embodiment, the environmental modeling component uses 3D Gaussian sputtering technology to create a high-fidelity environmental model of the mining area; generates an accurate 3D representation of geological structures, equipment locations, and safety areas; and enhances situational awareness and decision-making capabilities by providing detailed visualization of the mining environment.

[0037] The environment modeling component is connected to the large language model proxy component, so that the high-fidelity environment model and related environment data generated by the environment modeling component can be used by the large language model proxy component. For example, when allocating mining tasks, the large language model proxy component can arrange the working path of the robot assistant component more reasonably based on the high-fidelity environment model, avoiding problems such as equipment collision or low mining efficiency due to unfamiliarity with the environment.

[0038] The environmental modeling component is connected to the knowledge graph component, and the environmental data is updated to the knowledge graph component, which helps to enrich the information base of the knowledge graph component. For example, when the environmental modeling component collects new data on geological structure changes in the mining area, it is updated to the knowledge graph component, so that the knowledge graph component can keep up with the times and reflect the actual situation of the mining area.

[0039] The environmental modeling component collects environmental data of the mining area through various sensors (such as lidar, infrared and seismic sensors, etc.). The above data includes but is not limited to the geological structure, topography, temperature, humidity and other information of the mining area.

[0040] Based on the collected multi-source environmental data, the environmental modeling component uses 3D Gaussian sputtering technology to construct a high-fidelity environmental model of the mining area. This model can accurately present the actual environment of the mining area, such as showing the distribution of coal seams, the location of geological faults, and environmental parameters of different areas.

[0041] For example, in Figure 8In the paper, the process of using the 3D Gaussian splash algorithm for environment modeling is demonstrated. Specifically, the process of collecting and processing environmental data to create a high-fidelity model is as follows: First, collect data, that is, use sensors to collect environmental data; Second, processing the data, i.e. applying the 3D Gaussian sputtering technique to the collected data; Third, generate the model, which creates a detailed 3D representation of the geological structure, equipment locations, and safety zones; Fourth, update the knowledge graph component, that is, integrate environmental data into the knowledge graph component; Fifth, provide models, that is, distribute high-fidelity models to large language model proxy components and robot assistant components; Sixth, enhanced awareness, that is, using models to improve situational awareness and decision-making capabilities.

[0042] As can be seen above, the environmental modeling component provides an intuitive and accurate environmental information foundation for the entire system. During the mining process, the system can adjust the mining strategy in real time based on the high-fidelity environmental model and the updated knowledge graph.

[0043] In one embodiment of the present disclosure, reference Figure 2 and Figure 4 , the robot assistant component is further used to collect second data and send the second data to the large language model component; The large language model component is also used to process and analyze the second data for problem detection; The large language model proxy component is also used to adjust task allocation and update the environment modeling component after problem detection is completed.

[0044] In this embodiment, the second data may be equipment performance data, geological condition data, and safety parameter data.

[0045] The robot assistant component sends the collected second data to the large language model component to provide real-time feedback information to the system so that the system can understand the task execution status and environmental changes in a timely manner.

[0046] After receiving the second data sent by the robot assistant, the large language model component uses its powerful language processing and data analysis capabilities to process and analyze the above data. The large language model component can identify key information in the data, extract useful features, and compare and associate them with existing knowledge and experience. For example, by analyzing the working parameters and mining progress data of the coal mining machine, it can be judged whether the current mining efficiency meets expectations and whether there are potential problems.

[0047] Based on the analysis of the second data, the large language model component performs problem detection. The large language model component can identify various potential problems, such as equipment failure, safety hazards, low mining efficiency, etc.

[0048] After the large language model component completes problem detection, the large language model proxy component adjusts task allocation based on the detection results. If a device failure is found, the large language model component can assign tasks to the robot component for repair.

[0049] The large language model proxy component is also responsible for feeding back problem detection results and task adjustment information to the environmental modeling component so that the environmental modeling component can update the high-fidelity environmental model of the mining area in a timely manner. For example, if a safety hazard is found in a certain area, the environmental modeling component can mark the area in the model and update the relevant environmental data to provide a more accurate basis for subsequent mining decisions.

[0050] For example, in Figure 4 In the paper, the real-time monitoring and data collection process within this system is outlined. Specifically, the steps involved in collecting and analyzing data from the mining environment are as follows: First, initialize the sensor and camera to activate the environmental sensor and camera; Second, collect data, that is, collect data on equipment performance, geological conditions and safety parameters; Third, process the data, that is, call the large language model component to process and analyze the collected data; Fourth, problem detection, which is identifying any potential problems; Fifth, generate alerts and recommendations, that is, provide real-time alerts and actionable insights if problems are detected; Sixth, adjust task allocation. The large language model proxy component reallocates tasks based on the analyzed data. Seventh, update the environment model, that is, use 3D Gaussian sputtering to reflect the current situation; Eighth, transmit data, that is, use the communication network to share data within the system.

[0051] From the above, it can be concluded that this embodiment significantly improves the intelligence and adaptability of the system by introducing the collaborative work of the robot assistant component, the large language model component and the large language model proxy component. This embodiment can detect potential problems in a timely manner, reduce manual intervention, and improve the efficiency and accuracy of problem detection.

[0052] In one embodiment of the present disclosure, reference Figure 2 and Figure 5 ,The knowledge graph component is also used to analyze data and identify faults, obtain fault results, and send the fault results to the retrieval enhancement generation component; The retrieval enhancement generation component is also used to generate contextual responses and send the contextual responses to the large language model proxy component; A large language model agent component is also used to make recommendations.

[0053] In this embodiment, the knowledge graph component uses its rich structured knowledge and powerful reasoning ability to conduct in-depth analysis of data from different channels. The knowledge graph component can compare the real-time collected data with known normal operating modes and failure modes to identify possible failures. For example, by analyzing the equipment's operating parameters, sensor data, and historical failure cases, it can be determined whether the equipment has signs of failure such as overheating, abnormal vibration, or performance degradation.

[0054] Once a fault is identified, the knowledge graph component will organize the fault results into a clear report format and send it to the search enhancement generation component. The above fault results can include detailed information such as fault type, location, severity, etc.

[0055] After receiving the fault results sent by the knowledge graph component, the retrieval enhancement generation component will search and analyze the huge information base based on the above results. The retrieval enhancement generation component will generate a response containing rich contextual information by combining historical fault solutions, expertise in related fields, and the latest research results. The above response can not only explain the cause of the fault, but also provide possible solutions and preventive measures.

[0056] The retrieval enhancement generation component sends the generated contextual response to the large language model proxy component to provide it with decision support.

[0057] After receiving the context response sent by the retrieval enhancement generation component, the large language model proxy component will comprehensively consider the operating status and goals of the entire system and make specific suggestions. At the same time, the large language model proxy component can also make suggestions for strengthening equipment monitoring and maintenance in the future based on the preventive measures in the context response to reduce the probability of similar failures recurring.

[0058] For example, in Figure 5 In the paper, the mechanism of predictive maintenance and decision support is explained. Specifically, the process of how to use historical data and real-time analysis to optimize maintenance planning and support decision-making is as follows: First, retrieve historical data, that is, access past performance data from the knowledge graph component; Second, analyzing the data, that is, using knowledge graph components to identify patterns and potential failures; Third, generate responses, i.e., the retrieval enhancement generation component generates contextual responses; Fourth, provide insights, i.e. the large language model component provides actionable insights and recommendations; Fifth, recommend measures, that is, propose preventive maintenance measures; Sixth, optimize the plan, that is, adjust the maintenance plan based on the insights; Seventh, assign tasks, that is, the large language model agent component assigns maintenance tasks to the robot assistant component; Eighth, task execution, i.e. the robot assistant component performs maintenance tasks; Ninth, log data, that is, record data in the knowledge graph component for future use.

[0059] From the above, it can be concluded that this embodiment significantly improves the efficiency and accuracy of problem diagnosis and solution generation by adopting the collaborative work of knowledge graph components, retrieval enhancement generation components and large language model proxy components. The deep analysis capability of the knowledge graph component quickly locks the root cause of the fault, laying a solid foundation for subsequent processing. The retrieval enhancement generation component constructs a response rich in contextual information through intelligent matching and generation, effectively enhancing the accuracy and pertinence of information communication. In the end, the large language model proxy component makes suggestions in a humane way, which not only improves the target user experience, but also promotes the intelligent and automated process of problem solving, bringing significant efficiency improvements and cost savings to industry applications.

[0060] In one embodiment of the present disclosure, reference Figure 2 and Figure 7 , the large language model proxy component is specifically used for: third data responsive to the large language model component; Analysis is performed based on the third data and tasks are assigned to the robot assistant components.

[0061] In this embodiment, the large language model proxy component always keeps an eye on the output of the large language model component, and once the large language model component generates and sends the third data, the large language model proxy component can quickly receive it. The third data can be an in-depth analysis result of a specific problem, a comprehensive assessment of the current coal mining scenario, or insights generated based on a new input request.

[0062] The third data is a key decision-making basis for the large language model agent component, which can contain key insights into the current mining status. The large language model agent component needs to accurately understand the meaning and potential impact of the third data in order to make reasonable task allocation decisions.

[0063] The large language model agent component performs a comprehensive analysis of the third-party data, which may include an assessment of the accuracy, timeliness, and relevance of the data.

[0064] At the same time, the large language model proxy component will analyze the current state of the entire system, and will also consider the operation of other components, the mining environment information provided by the environmental modeling component, the existing knowledge in the knowledge graph component, etc. Based on the analysis results of the third data, the large language model proxy component determines the specific assignment task.

[0065] The large language model proxy component also sets priorities based on the urgency and importance of the task. For example, if the third-party data indicates that a critical equipment has a serious fault, the large language model proxy component will immediately assign a high-priority maintenance task to the robot assistant component to ensure that production returns to normal as soon as possible. After determining the task and priority, the large language model proxy component sends clear and unambiguous task instructions to the robot assistant component. The above instructions should include detailed information such as the specific content of the task, the execution location, and the time requirements, so that the robot assistant component can accurately understand and efficiently perform the task. For example, "Dispatch robot R1 to coal mining area A to inspect and repair the coal mining machine failure numbered C1. The task must be completed within 2 hours." This instruction allows the robot assistant component to act quickly to ensure the smooth progress of coal mining.

[0066] For example, in Figure 7 In the paper, we outline the task management and coordination functions performed by the large language model agent component. Specifically, the process of how it interacts with other components to ensure optimal resource utilization and task execution is as follows: First, receiving data, that is, the large language model proxy component receives data input from the large language model component; Second, analyze needs, which means assessing operational needs and requirements; Third, assign tasks, i.e. assign tasks to robot assistant components based on the analysis; Fourth, coordinate with the knowledge graph component and the retrieval enhancement generation component, i.e., ensure that the tasks are contextually relevant and effectively managed; Fifth, adjust the priority, that is, dynamically adjust the task priority according to real-time data; Sixth, monitoring execution, that is, supervising the execution of tasks; Seventh, optimize resources, that is, optimize resource allocation and use; Eighth, provide feedback, i.e. pass performance data back to the large language model component for continuous improvement.

[0067] It can be concluded from the above that this embodiment significantly improves the flexibility and efficiency of task processing, ensures the accuracy and timeliness of task allocation, reduces manual intervention through intelligent task allocation, speeds up problem solving, and provides target users with a more efficient and convenient service experience.

[0068] In one embodiment of the present disclosure, reference Figure 2 and Fig. 9 , the human-machine interface component is used to receive a first request from a target user; The large language model component is also used to assign tasks to the robot assistant component based on the large language model proxy component in response to the first request; The HMI component is also used to display the updated task status.

[0069] In this embodiment, the target user sends a command (i.e., a first request) to the human-machine interface component, and the human-machine interface component forwards the first request to the large language model component. After receiving the first request, the large language model component processes it and generates insights. After receiving the insights, the large language model proxy component sends a query data request to the knowledge graph component, and the knowledge graph component returns the data to the large language model proxy component after the query is completed. The large language model proxy component requests context information from the retrieval enhancement generation component, the retrieval enhancement generation component retrieves relevant information and sends it to the knowledge graph component, and provides a context response to the large language model proxy component. The large language model proxy component assigns tasks to the robot assistant component, and the robot assistant component provides environmental data to the environment modeling component, so that the environment modeling component updates the environment model, and sends the updated model to the large language model proxy component.

[0070] Environmental sensors send environmental data to the large language model component, which sends the processed environmental data to the large language model proxy component, which provides feedback and insights to the target user. At the same time, the large language model proxy component updates the task status and visualization to the human-computer interface component.

[0071] From the above, we can conclude that the close collaboration between the human-machine interface component, the large language model component, and the robot assistant component greatly improves the target user experience. The target user initiates a request through the intuitive human-machine interface component, and the large language model component responds quickly and intelligently assigns tasks to the robot assistant, realizing an efficient task processing process. At the same time, the human-machine interface displays task status updates in real time, allowing the target user to keep track of the progress at any time, enhancing transparency and interactivity.

[0072] In one embodiment of the present disclosure, reference Figure 2 ,The human-machine interface components are also used to display the mining environment.

[0073] In one embodiment of the present disclosure, reference Figure 2 ,The knowledge graph component includes a knowledge base that includes equipment data, geological data, safety protocols, and historical performance indicators.

[0074] In one embodiment of the present disclosure, reference Figure 2 and Figure 6,The robotic assistant component is also used to navigate the mining environment.

[0075] exist Figure 6 In the paper, the autonomous operation of the robot assistant components is described. Specifically, how to perform tasks and how to interact with other components for guidance and coordination are as follows: First, the navigation environment, i.e., the robotic assistant component moves autonomously within the mining area; second, execution, which involves performing tasks such as drilling, blasting, and material handling; Third, receive guidance, that is, the large language model component provides real-time operation guidance; Fourth, coordination activities, where a large language model agent component manages and coordinates the activities of multiple robot assistants; Fifth, optimize execution, that is, ensure efficient task execution through continuous coordination; Sixth, using models, that is, using 3D environmental models for precise navigation and mission planning; Seventh, update status, that is, report the task status through the human-machine interface component for operator monitoring.

[0076] In one embodiment of the present disclosure, referring to Figure 3 , Figure 3 This is an architectural diagram of an intelligent interactive integrated system for coal mining based on a large language model.

[0077] exist Figure 3 In the system, the role of each component is defined, and the data and task flows between the components are planned. Specifically, the key components and functions of the system are as follows: A large language model component is used to process natural language commands, analyze data, and provide insights; A large language model agent component is used to manage tasks and coordinate activities between components; The knowledge graph component is used to organize and retrieve relevant information; The retrieval enhancement generation component is used to generate accurate responses and dynamically retrieve information; The robotic helper component is used to perform autonomous mining tasks; The role of 3D Gaussian sputtering for environment modeling is to create high-fidelity environment models; Advanced environmental sensors are used to collect real-time environmental data; Safety management systems are used to assess and warn of risks; Adaptive learning algorithms are used to enhance system performance over time; Human-machine interface components are used to provide visualization and allow operator commands; Energy management systems are used to optimize electricity usage; Communications networks are used to ensure seamless data transfer; Data analysis platforms are used to process and analyze data; Cloud integration is used to provide scalable data storage and processing; Teleoperation capabilities are used to support remote management of mining activities; Environmental impact assessments are used to monitor and ensure compliance with environmental regulations; Dynamic task scheduling is used to optimize task allocation; Real-time collaboration tools are used to facilitate communication and coordination; Fault detection and diagnostic systems are used to monitor and diagnose problems; Training and simulation modules are used to provide training for operators; Resource allocation optimization is used to optimize resource usage; Regulatory compliance monitoring is used to ensure safe and legal operations; Environmental monitoring and reporting is used to generate compliance reports; Enhanced security measures are used to protect against cyber threats.

[0078] Among them, the large language model component processes data and interacts with the large language model agent component; the large language model agent component coordinates with the knowledge graph component, the retrieval enhancement generation component and the robot assistant component; the 3D Gaussian sputtering and sensors provide environmental data; the security management system, energy management system and other systems support operational efficiency and safety.

[0079] The above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than to limit them. Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present disclosure.

Claims

1. An intelligent interactive integrated system for coal mining based on a large language model, characterized in that: include: Human-computer interface component, large language model component, large language model agent component, knowledge graph component and robot assistant component; The human-machine interface component is used to respond to a first request from a target user and send the first request to the large language model component; The large language model component generates a first insight according to the first request, and sends the first insight to the large language model proxy component; The large language model proxy component is used to send the first insight to the knowledge graph component; The knowledge graph component is used to perform data query according to the first insight, obtain first data, and send the first data to the large language model proxy component; The large language model proxy component is further used to generate an assignment task based on the first data, and send the assignment task to the robot assistant component; The robot assistant component is used to perform the assigned task, and the assigned task is a mining task.

2. The coal mining intelligent interactive integrated system based on a large language model as claimed in claim 1, characterized in that: Also included is a retrieval enhancement generation component; The large language model proxy component is further used to send the first data to the retrieval enhancement generation component; The retrieval enhancement generation component is used to generate a query response based on the first data, and send the query response to the large language model proxy component.

3. The coal mining intelligent interactive integrated system based on a large language model as claimed in claim 1, characterized in that: Also includes environmental modeling components: The environmental modeling component is connected to the large language model proxy component and the knowledge graph component respectively, and is used to collect environmental data of the mining area and generate a high-fidelity environmental model of the mining area according to the environmental data; The environment modeling component is also used to update the environment data to the knowledge graph component.

4. The coal mining intelligent interactive integrated system based on a large language model as claimed in claim 3, characterized in that: The robot assistant component is also used to collect second data and send the second data to the large language model component; The large language model component is also used to process and analyze the second data to perform problem detection; The large language model proxy component is also used to adjust task allocation and update the environment modeling component after the problem detection is completed.

5. The coal mining intelligent interactive integrated system based on a large language model as claimed in claim 2, characterized in that: The knowledge graph component is also used to analyze data and identify faults, obtain fault results, and send the fault results to the retrieval enhancement generation component; The retrieval enhancement generation component is also used to generate a context response and send the context response to the large language model proxy component; The large language model proxy component is also used to make recommendations.

6. The coal mining intelligent interactive integrated system based on a large language model as claimed in claim 2, characterized in that: The large language model proxy component is specifically used for: third data responsive to the large language model component; An analysis is performed based on the third data, and tasks are assigned to the robot assistant component.

7. The coal mining intelligent interactive integrated system based on a large language model as claimed in claim 1, characterized in that: The human-machine interface component is used to receive a first request from the target user; The large language model component is further used to respond to the first request and assign tasks to the robot assistant component based on the large language model proxy component; The human-machine interface component is also used to display updated task status.

8. The coal mining intelligent interactive integrated system based on a large language model as claimed in claim 1, characterized in that: The human-machine interface component is also used to display the mining environment.

9. The coal mining intelligent interactive integrated system based on a large language model as claimed in claim 1, characterized in that: The knowledge graph component includes a knowledge base, which includes equipment data, geological data, safety protocols, and historical performance indicators.

10. The coal mining intelligent interactive integrated system based on a large language model according to claim 1, characterized in that: The robotic assistant assembly is also used to navigate a mining environment.

Images