Disaster warning system and method

Abstract

The present application provides a kind of disaster early warning system and method, the system includes: multiple intelligent agent equipment and multiple publishing equipment, large language model is arranged in intelligent agent equipment;The i-th target meteorological monitoring station is respectively corresponding with each intelligent agent equipment in the i-th intelligent agent equipment and each publishing equipment in the i-th publishing equipment, the i-th intelligent agent equipment is respectively connected with the i-th target meteorological monitoring station and the i-th publishing equipment, and each intelligent agent equipment is connected with each intelligent agent equipment.The present application provides a kind of disaster early warning system and method, can be based on large language model, without human intervention and expert judgment, more objective, more accurate and more efficient to determine whether to collect climate data for climate monitoring station release disaster warning or disaster broadcast, avoid the influence of the efficiency and accuracy of data processing to the data provided by different meteorological monitoring stations due to the different climate data format, quality and accuracy of different climate monitoring stations.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric digital data processing, and particularly relates to a disaster early warning system and method. BACKGROUND

[0002] Earthquake, tsunami, tornado, forest fire and debris flow disasters have brought great harm to people's production and life, therefore, timely and accurate disaster early warning can minimize casualties and property losses, improve the efficiency and effectiveness of coping with disasters, and protect the stability and development of society.

[0003] In the related art, a traditional disaster early warning system can obtain data collected by earthquake monitoring stations, tsunami monitoring stations and other meteorological monitoring stations, and then can determine whether to issue a disaster early warning based on the data collected by the meteorological monitoring stations.

[0004] However, in the process of determining whether to issue a disaster early warning based on the data collected by the meteorological monitoring stations, the traditional disaster early warning system usually needs expert intervention and judgment, resulting in that the traditional disaster early warning system has strong subjectivity and low efficiency in disaster early warning, the disaster early warning lacks standardization and uniformity, and the disaster early warning may not be timely. Therefore, how to improve the accuracy and efficiency of disaster early warning is a technical problem to be solved in the field. SUMMARY

[0005] The present application provides a disaster early warning system and method to solve the defect that it is difficult to accurately and efficiently perform disaster early warning in the prior art, and to improve the accuracy and efficiency of disaster early warning.

[0006] The application provides a disaster early warning system, comprising: a plurality of agent devices and a plurality of publishing devices, wherein a large language model is deployed in the agent devices; an ith target meteorological monitoring station corresponds to an ith agent device in each of the agent devices and an ith publishing device in each of the publishing devices, the ith agent device is in communication connection with the ith target meteorological monitoring station and the ith publishing device, the agent devices are in communication connection, i is an element of {1, 2, 3, …, N-1, N}, and N represents the total number of target meteorological monitoring stations; the ith agent device is used for, in the case of obtaining meteorological data collected by the ith target meteorological monitoring station, performing disaster prediction on the meteorological data, inputting the disaster prediction result corresponding to the meteorological data into the large language model in the ith agent device after obtaining the disaster prediction result corresponding to the meteorological data in the form of natural language description, obtaining a first target control instruction output by the large language model in the ith agent device, and then sending the first target control instruction to the ith publishing device, wherein the first target control instruction is a first control instruction used for instructing the publishing device to remain in a silent state, a second control instruction carrying disaster early warning information corresponding to the meteorological data, or a third control instruction carrying disaster broadcast information corresponding to the meteorological data; and the ith publishing device is used for, in the case of receiving the first target control instruction, publishing the disaster broadcast information corresponding to the meteorological data, publishing the disaster early warning information corresponding to the meteorological data, or remaining in a silent state in response to the first target control instruction.

[0007] According to the disaster early warning system provided by the application, the ith agent device is further used for, in the case that the current time is a target time, obtaining meteorological information of a region where the ith target meteorological monitoring station is located in the Internet between the current time and a previous target time, inputting the meteorological information into the large language model in the ith agent device, obtaining a second target control instruction output by the large language model in the ith agent device, and then sending the second target control instruction to the ith publishing device, wherein the second target control instruction is the first control instruction, a fourth control instruction carrying disaster early warning information corresponding to the meteorological information, or a fifth control instruction carrying disaster broadcast information corresponding to the meteorological information, and any two adjacent target times are separated by a first preset time length; and the ith publishing device is further used for, in the case of receiving the second target control instruction, publishing the disaster early warning information corresponding to the meteorological information, publishing the disaster broadcast information corresponding to the meteorological information, or remaining in a silent state in response to the second target control instruction.

[0008] According to the disaster early warning system, the first intelligent agent device is further configured to, in a case where disaster information published by a jth intelligent agent device of the intelligent agent devices is received, send a sixth control instruction carrying second natural language information to the ith publishing device, if it is determined that the jth target meteorological monitoring station is a trusted target meteorological monitoring station corresponding to the ith target meteorological monitoring station, the disaster information published by the jth intelligent agent device belongs to the same disaster field as meteorological data collected by the ith target meteorological monitoring station, and the meteorological data collected by the ith target meteorological monitoring station matches the disaster information published by the jth intelligent agent device successfully, the second natural language information indicates that the disaster information published by the jth intelligent agent device is trusted information, j is an element in a set {1, 2, 3, …, N-1, N}, i is not equal to j, and the disaster information includes disaster broadcast information and disaster early warning information; and the ith publishing device is further configured to, in a case where the sixth control instruction is received, publish the second natural language information in response to the sixth control instruction.

[0009] According to the disaster early warning system, the first intelligent agent device is further configured to, in a case where disaster information published by a jth intelligent agent device of the intelligent agent devices is received, send a sixth control instruction carrying second natural language information to the ith publishing device, if it is determined that the jth target meteorological monitoring station is a trusted target meteorological monitoring station corresponding to the ith target meteorological monitoring station, the disaster information published by the jth intelligent agent device belongs to the same disaster field as meteorological data collected by the ith target meteorological monitoring station, and the meteorological data collected by the ith target meteorological monitoring station matches the disaster information published by the jth intelligent agent device successfully, the second natural language information indicates that the disaster information published by the jth intelligent agent device is trusted information, j is an element in a set {1, 2, 3, …, N-1, N}, i is not equal to j, and the disaster information includes disaster broadcast information and disaster early warning information; and the ith publishing device is further configured to, in a case where the sixth control instruction is received, publish the second natural language information in response to the sixth control instruction.

[0010] The disaster early warning system provided in the application, the i th agent device is further configured to, in the case that the disaster information published by the j th agent device among the agent devices is received, if it is determined that the j th target meteorological monitoring station is not a trusted target meteorological monitoring station corresponding to the i th target meteorological monitoring station, send the sixth control instruction or the seventh control instruction to the i th publishing device based on the number of the second natural language information and the number of the third natural language information received by the i th agent device within a target period, the starting time of the target period being the time when the i th agent device receives the disaster information published by the j th agent device, and the length of the target period being a second preset length.

[0011] The disaster early warning system provided in the application, the i th agent device is further configured to, in the case that the disaster information published by the j th agent device among the agent devices is received, if it is determined that the j th target meteorological monitoring station is not a trusted target meteorological monitoring station corresponding to the i th target meteorological monitoring station, send the sixth control instruction or the seventh control instruction to the i th publishing device based on the number of the second natural language information and the number of the third natural language information received by the i th agent device within a target period, the starting time of the target period being the time when the i th agent device receives the disaster information published by the j th agent device, and the length of the target period being a second preset length.

[0012] According to the disaster early warning system, the first intelligent agent device is further configured to, in a case where disaster information published by a jth intelligent agent device of the intelligent agent devices is received, send the first control instruction to the ith publishing device if it is determined that the jth target meteorological monitoring station is a trusted target meteorological monitoring station corresponding to the ith target meteorological monitoring station, the disaster information published by the jth intelligent agent device is in the same disaster field as the ith target meteorological monitoring station, meteorological data collected by the ith target meteorological monitoring station and the disaster information published by the jth intelligent agent device are not successfully matched, and a distance between the jth target meteorological monitoring station and the ith target meteorological monitoring station is not less than a distance threshold, j is an element of {1, 2, 3, …, N-1, N}, and i is not equal to j.

[0013] According to the disaster early warning system, the first intelligent agent device is further configured to, after sending the sixth control instruction to the ith publishing device, increase a trusted value of the jth target meteorological monitoring station relative to the ith target meteorological monitoring station by 1, and decrease the trusted value of the jth target meteorological monitoring station relative to the ith target meteorological monitoring station by 1 in a case where the seventh control instruction is sent to the ith publishing device; the first intelligent agent device is further configured to, in a case where it is determined that the trusted value of the jth target meteorological monitoring station relative to the ith target meteorological monitoring station is 0, eliminate the jth target meteorological monitoring station from the trusted target meteorological monitoring stations corresponding to the ith target meteorological monitoring station.

[0014] According to the disaster early warning system, the first intelligent agent device is further configured to, in a case where it is determined that the jth target meteorological monitoring station is not an alternative target meteorological monitoring station corresponding to the ith target meteorological monitoring station and the sixth control instruction is sent to the ith publishing device, determine the jth target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the ith target meteorological monitoring station; the first intelligent agent device is further configured to, after determining the jth target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the ith target meteorological monitoring station, determine the jth target meteorological monitoring station as a trusted target meteorological monitoring station corresponding to the ith target meteorological monitoring station in a case where a cumulative number of times of sending the sixth control instruction to the ith publishing device exceeds a number threshold.

[0015] According to the disaster early warning system provided in the application, the first intelligent agent device is further configured to, in a case where disaster information published by a jth intelligent agent device of the intelligent agent devices is received, if it is determined that the jth target meteorological monitoring station is a trusted target meteorological monitoring station corresponding to the ith target meteorological monitoring station, input the disaster information published by the jth intelligent agent device into a large language model in the ith intelligent agent device, and obtain a third target control instruction output by the large language model in the ith intelligent agent device, the third target control instruction being the first control instruction, an eighth control instruction carrying disaster early warning information corresponding to the disaster information published by the jth intelligent agent device, or a ninth control instruction carrying disaster broadcast information corresponding to the disaster information published by the jth intelligent agent device, j∈{1, 2, 3,..., N-1, N}, i≠j; and the ith publishing device is further configured to, in a case where the third target control instruction is received, publish the disaster early warning information corresponding to the disaster information published by the jth intelligent agent device, publish the disaster broadcast information corresponding to the disaster information published by the jth intelligent agent device, or keep a mute state, in response to the third target control instruction.

[0016] The application further provides a disaster early warning method implemented based on the disaster early warning system described in any of the above, comprising: in a case where meteorological data collected by any target meteorological monitoring station is obtained, performing disaster prediction on the meteorological data to obtain a natural language description of a disaster prediction result corresponding to the meteorological data; inputting the disaster prediction result corresponding to the meteorological data into a large language model in an intelligent agent device corresponding to the any target meteorological monitoring station, and obtaining a first target control instruction output by the large language model, the first target control instruction being a first control instruction for instructing a publishing device to keep a mute state, a second control instruction carrying disaster early warning information corresponding to the meteorological data, or a third control instruction carrying disaster broadcast information corresponding to the meteorological data; and sending the first target control instruction to a publishing device corresponding to the any target meteorological monitoring station, so that the publishing device corresponding to the any target meteorological monitoring station publishes the disaster broadcast information corresponding to the meteorological data, publishes the disaster early warning information corresponding to the meteorological data, or keeps a mute state, in response to the first target control instruction.

[0017] The application further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the disaster early warning method described in any of the above when executing the program.

[0018] The application further provides a non-transitory computer readable storage medium having a computer program stored thereon, wherein the computer program is executable on a processor to implement the disaster early warning method described in any of the above.

[0019] The application further provides a computer program product comprising a computer program which, when executed by a processor, implements the disaster early warning method according to any one of the above.

[0020] The disaster early warning system and method provided by the application, the disaster early warning system comprises a plurality of agent devices and a plurality of publishing devices, a large language model is deployed in the agent devices, and the large language model can be used to more objectively, more accurately and more efficiently determine whether to publish a disaster early warning or a disaster broadcast for climate data collected by a climate monitoring station without human intervention and expert judgment, so that the efficiency and accuracy of data processing of data provided by different weather monitoring stations are not affected by different climate data formats, quality and accuracy collected by different climate monitoring stations, the accuracy, objectivity, uniformity and publishing efficiency of disaster early warning publishing can be effectively improved, so that personnel casualties and property losses caused by disasters can be better reduced, and the efficiency and effect of disaster response can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0022] Figure 1 is a structural schematic diagram of the disaster early warning system provided by the application.

[0023] Figure 2 is a connection schematic diagram of the disaster early warning system and the agent device provided by the application.

[0024] Figure 3 is a data interaction schematic diagram of the disaster early warning system provided by the application.

[0025] Figure 4 is a variation schematic diagram of the parameter for adjusting the determination weight in the disaster early warning system provided by the application.

[0026] Figure 5 is a flow schematic diagram of the disaster early warning system provided by the application.

[0027] Figure 6 is a structural schematic diagram of the electronic device provided by the application. DETAILED DESCRIPTION

[0028] In order to make the objects, technical solutions and advantages of the present application clearer, the following will clearly and completely describe the technical solutions in the present application in conjunction with the drawings in the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0029] In the description of the application, it should be noted that, unless otherwise explicitly specified and limited, the terms "mounting", "connection", "connecting" should be understood in a broad sense, for example, can be fixed connection, can also be detachable connection, or integrally connected; can be mechanical connection, can also be electrical connection; can be directly connected, can also be indirectly connected through an intermediate medium, can be the communication inside two elements. For those of ordinary skill in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0030] In the description of the present application, the terms "first", "second" and the like are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second" and the like are generally of a kind and do not limit the number of objects, for example, the first object can be one or more. In addition, in the description of the present application, "and / or" means at least one of the connected objects, and the character " / " generally means that the front and rear associated objects are in an "or" relationship.

[0031] It should be noted that the conventional disaster warning system in the related art integrates data provided by earthquake monitoring stations, tsunami monitoring stations and other meteorological monitoring stations to monitor the activities of disasters such as earthquakes and tsunamis in real time. Among them, the earthquake monitoring station, the tsunami monitoring station and other meteorological monitoring stations can use advanced equipment such as broadband seismograph and submarine seismograph to capture relevant data on the ground and seabed in real time.

[0032] After the conventional disaster warning system obtains the data collected by each meteorological monitoring station, for earthquake warning, the conventional disaster warning system can use the seismic wave inversion algorithm to analyze the seismic wave data captured by the earthquake monitoring station in detail, so as to accurately calculate the location and depth of the earthquake source and predict the potential impact of the earthquake. After the conventional disaster warning system predicts the location and depth of the earthquake source and predicts the potential impact of the earthquake, experts can determine whether to issue an earthquake warning based on the location and depth of the earthquake source and the potential impact of the earthquake.

[0033] For tsunami warning, the traditional disaster warning system can simulate the tsunami wave monitored by the tsunami monitoring station using the shallow water wave model, predict the propagation path and possible height of the tsunami wave, and further predict the arrival time and impact range of the tsunami. After the traditional disaster warning system predicts the propagation path and possible height of the tsunami wave and the arrival time and impact range of the tsunami, experts can determine whether a tsunami warning needs to be issued based on the propagation path and possible height of the tsunami wave and the arrival time and impact range of the tsunami.

[0034] Once the traditional disaster warning system determines that a disaster warning needs to be issued, the traditional disaster warning system can quickly convey disaster warning information to the public and government agencies through various channels such as television, radio, the Internet, and mobile applications, ensuring timely response.

[0035] In addition, automated emergency response systems can automatically respond to disasters based on the disaster warning information, such as automatically stopping train operation and closing natural gas pipelines, which are also deployed in some areas to further reduce disaster risks.

[0036] International cooperation plays a crucial role in this process, and data sharing between global monitoring networks significantly improves the accuracy and efficiency of the traditional disaster warning system's warning issuance. For example, the Pacific Tsunami Warning System (PTWS) as a traditional disaster warning system can enhance the monitoring and warning capabilities of potential tsunami threats in the Pacific by coordinating resources and information from different countries and agencies. The goal of these systems is to use technological advances and global cooperation to improve the timeliness and effectiveness of disaster response, thereby significantly reducing casualties and property losses caused by earthquakes and tsunamis.

[0037] However, the traditional disaster warning system faces several key problems in achieving efficient and accurate disaster warning, affecting its overall performance and response speed. First, when the traditional disaster warning system relies on data collected by multiple meteorological monitoring stations to determine whether to issue a disaster warning, expert intervention and judgment are required in the data analysis and decision-making process. For example, when the traditional disaster warning system uses seismic wave inversion algorithms to analyze seismic wave data captured by seismic monitoring stations, experts usually manually select which data collected by seismic monitoring stations to use.

[0038] However, on the one hand, the traditional disaster warning system relies on expert intervention and judgment in the decision-making process, not only slowing down the response speed of disaster warning, leading to the inability to issue disaster warnings in time in emergency situations, but also leading to the subjectivity of disaster warnings, lack of standardization and uniformity in disaster warning issuance, and reducing the accuracy of disaster warning issuance.

[0039] On the other hand, there is a lack of an efficient real-time automatic data interaction system between the earthquake monitoring stations and the tsunami monitoring stations, resulting in insufficient data utilization and affecting the timeliness and accuracy of the early warning.

[0040] In addition, the formats, qualities and accuracies of the data provided by different meteorological monitoring stations are not completely the same, which causes the traditional disaster early warning system to have difficulty in obtaining data, and the differences between the data provided by different meteorological monitoring stations make it difficult to develop a unified standard to evaluate and use these data, which seriously affects the efficiency and accuracy of the traditional disaster early warning system in data processing of the data provided by each meteorological monitoring station, and further affects the accuracy and efficiency of disaster early warning.

[0041] With the continuous addition of new meteorological monitoring stations and / or new equipment in existing meteorological monitoring stations, how to maintain the consistency and reliability of the data collected by each meteorological monitoring station, and how to evaluate the quality of the data collected by each new meteorological monitoring station and / or new equipment in existing meteorological monitoring stations have become a big challenge.

[0042] The following will be combined Figures 1-4 to describe the disaster early warning system provided by the present application.

[0043] Figure 1 is a structural schematic diagram of the disaster early warning system provided by the present application, as Figure 1 shown, the disaster early warning system 101 comprises a plurality of agent devices 102 and a plurality of publishing devices 103, and a large language model is deployed in the agent devices 102.

[0044] Figure 2 is a connection schematic diagram of the disaster early warning system provided by the present application and the agent device. As Figure 2 shown, the i-th target meteorological monitoring station corresponds to the i-th agent device 102 in each agent device 102 and the i-th publishing device 103 in each publishing device 103, respectively, the i-th agent device 102 is in communication connection with the i-th target meteorological monitoring station 104 and the i-th publishing device 103, respectively, and each agent device 102 is in communication connection, i∈{1,2,3,...,N-1,N}, N represents the total number of target meteorological monitoring stations 104.

[0045] The i-th intelligent agent device 102 is configured to, in a case where meteorological data collected by the i-th target meteorological monitoring station 104 is acquired, perform disaster prediction on the meteorological data, input the disaster prediction result corresponding to the meteorological data into a large language model in the i-th intelligent agent device 102 after obtaining the disaster prediction result corresponding to the meteorological data in the natural language description, obtain a first target control instruction output by the large language model in the i-th intelligent agent device 102, and then send the first target control instruction to the i-th publishing device 103. The first target control instruction is a first control instruction for instructing the publishing device 103 to remain in a silent state, a second control instruction carrying disaster warning information corresponding to the meteorological data, or a third control instruction carrying disaster broadcast information corresponding to the meteorological data.

[0046] The i-th publishing device 103 is configured to, in a case where the first target control instruction is received, publish disaster broadcast information corresponding to the meteorological data, publish disaster warning information corresponding to the meteorological data, or remain in a silent state in response to the first target control instruction.

[0047] Specifically, the disaster in the embodiment of the present application can include but is not limited to earthquake, tsunami, tornado, forest fire, and mudslide, etc.

[0048] The meteorological monitoring station in the embodiment of the present application is a comprehensive station integrating multiple devices and systems, which is used for observing and recording various meteorological elements. The meteorological monitoring station can collect meteorological data through various devices and sensors for meteorological forecasting, disaster warning, agricultural and ecological environment monitoring, etc.

[0049] The target meteorological monitoring station 104 in the embodiment of the present application can include a plurality of meteorological monitoring stations determined based on actual needs, for example, all meteorological monitoring stations in a certain administrative region can be determined as the target meteorological monitoring station 104 in the embodiment of the present application. The target meteorological monitoring station 104 in the embodiment of the present application is not specifically limited.

[0050] The target meteorological monitoring station 104 in the embodiment of the present application includes but is not limited to an earthquake monitoring station, a tsunami monitoring station, an air quality monitoring station, a radar station, a satellite ground station, a wind field station, and an upper air observation station, etc. Moreover, the target meteorological monitoring station 104 in the embodiment of the present application can include multiple meteorological monitoring stations of the same type, for example, the target meteorological monitoring station 104 can include multiple earthquake monitoring stations, multiple tsunami monitoring stations, and multiple radar stations, which can be respectively arranged in different regions.

[0051] The target meteorological monitoring station 104 in the embodiment of the present application corresponds to one intelligent agent device 102 in the disaster warning system 101, and the intelligent agent device 102 in the disaster warning system 101 corresponds to one publishing device 103.

[0052] It should be noted that the agent device 102 in the embodiment of the present application can be an electronic device carrying an agent. The agent can refer to a virtual entity that can perceive the environment and take actions to achieve a specific goal, with a certain degree of autonomy and decision-making ability. The agent in the embodiment of the present application can realize intelligent behavior such as autonomous learning, data processing and decision-making through algorithms and models.

[0053] The agent device 102 in the embodiment of the present application further includes a processor, a plurality of instruction interfaces and a plurality of communication interfaces.

[0054] Optionally, the agent device 102 in the embodiment of the present application can include a plurality of communication interfaces, which are a first communication interface, a second communication interface, a third communication interface and a fourth communication interface. The agent device 102 can be communicatively connected with the corresponding target meteorological monitoring station 104 through the first communication interface, can be communicatively connected with the corresponding publishing device 103 through the second communication interface, can be communicatively connected with each target agent device 102 through the third communication interface, and can access the Internet through the fourth communication interface.

[0055] Optionally, any agent device 102 in the embodiment of the present application can be communicatively connected with each target agent device 102 through the third communication interface and a secure communication network.

[0056] It should be noted that a large language model is deployed in each agent device 102 in the embodiment of the present application. The large language model (Large Language Model, LLM) refers to a deep learning model trained by a large amount of text data, which can understand and generate natural language text. The large language model in the embodiment of the present application can be a language model trained by self-supervision based on the Transformer architecture, which can receive text information and generate long replies by generating the next word of the reply sentence through the decoder loop.

[0057] It should be noted that the i-th target meteorological monitoring station 104 can represent any target meteorological monitoring station 104 in the target meteorological monitoring station 104, the i-th agent device 102 can represent the agent device 102 corresponding to the i-th target meteorological monitoring station 104, and the i-th publishing device 103 can represent the publishing device 103 corresponding to the i-th agent device 102.

[0058] It should be noted that the disaster warning system 101 provided by the present application needs to initialize each agent device 102 after starting, so as to be used for disaster warning.

[0059] In the initialization phase, for each agent device 102, an initial prompt information described in natural language can be input to the large language model in each agent device 102. The initial prompt information can include the following: each control instruction, the meaning of each control instruction, and the instruction interface corresponding to each control instruction, the task target described in natural language, the task parameter setting described in natural language, and the required output form of the large language model. Among them, the task target includes making a decision based on the input natural language information, determining whether to issue a disaster warning or a disaster broadcast for the natural language information, and in the case of determining to issue a disaster warning for the natural language information, generating disaster warning information corresponding to the natural language information described in natural language, and in the case of determining to issue a disaster broadcast for the natural language information, generating disaster broadcast information corresponding to the natural language information described in natural language. The initial prompt information can also include the reasoning process and step decomposition of the current task, the instructions and corresponding parameters needed to be completed, and the next task to be completed.

[0060] It should be noted that the instruction set including each control instruction in the embodiment of the present application is a set of instructions that can be increased or decreased by the user, which specifies all external instructions that can be called by the large language model. The instruction set includes a termination control instruction, which is used to end the execution of the agent device 102 when the large language model in any agent device 102 outputs the termination control instruction. The instruction set can also include control instructions for issuing a disaster broadcast and a disaster warning, and a control instruction indicating that the issuing device 103 is in a silent state. In addition, the instruction set can also include control instructions for searching meteorological information on the Internet, and can also include control instructions for using search engines, browsing web pages, executing codes, file I / O, calling mathematical tools, etc. The target meteorological monitoring station 104 can also add control instructions in the instruction set by writing personalized instructions and executing codes.

[0061] After initializing the agent device 102, the user can input natural language description indicating the end of initialization to the large language model in the agent device 102, which can start running. When the large language model in the agent device 102 receives the start information, the agent device 102 starts running.

[0062] Figure 3 is a data interaction schematic diagram of the disaster warning system provided by the present application. As Figure 3As shown, for the ith agent device 102, in a case where the ith agent device 102 receives meteorological data collected by the ith target meteorological monitoring station 104, the agent in the ith agent device 102 can call an operation resource to perform disaster prediction on the meteorological data, and obtain a disaster prediction result corresponding to the meteorological data. The operation resource can be a processor in the ith agent device 102, and the processor is deployed with a disaster prediction model. The disaster prediction model can be composed of a preset judgment network, a memory judgment network, and an interactive memory storage. The judgment network is a multi-layer Transformer network, and the output is a scalar credibility value.

[0063] For example, in a case where the ith target meteorological monitoring station 104 is an earthquake monitoring station, if the ith agent device 102 receives seismic wave data collected by the ith target meteorological monitoring station 104, the agent in the ith agent device 102 can call an operation resource to perform seismic wave inversion on the seismic wave data, and obtain a natural language description of the earthquake source location and the earthquake magnitude as the disaster prediction result corresponding to the seismic wave data.

[0064] For another example, in a case where the ith target meteorological monitoring station 104 is a tsunami monitoring station, if the ith agent device 102 receives offshore wave data collected by the ith target meteorological monitoring station 104, the agent in the ith agent device 102 can call an operation resource to perform shallow water wave model calculation on the offshore wave data, and obtain a natural language description of the tsunami height and the tsunami arrival time as the disaster prediction result corresponding to the offshore wave data.

[0065] After the ith agent device 102 obtains the natural language description of the disaster prediction result corresponding to the meteorological data, the agent in the ith agent device 102 can input the disaster prediction result into a large language model in the ith agent device 102.

[0066] Based on the disaster prediction result, the large language model in the ith agent device 102 can determine whether a disaster warning or a disaster broadcast needs to be issued for the meteorological data. In a case where it is determined that a disaster warning needs to be issued for the meteorological data, the large language model generates disaster warning information corresponding to the meteorological data. In a case where it is determined that a disaster broadcast needs to be issued for the meteorological data, the large language model generates disaster broadcast information corresponding to the meteorological data. Then, the large language model outputs a first control instruction for instructing the publishing device 103 to remain in a mute state, a second control instruction indicating that a disaster warning needs to be issued for the meteorological data and carrying the disaster warning information corresponding to the meteorological data, or a third control instruction indicating that a disaster broadcast needs to be issued for the meteorological data and carrying the disaster broadcast information corresponding to the meteorological data.

[0067] It should be noted that the disaster warning information corresponding to the meteorological data and the disaster broadcast information corresponding to the meteorological data in the embodiment of the application are both natural language information described in natural language.

[0068] It should be noted that the first target instruction in the embodiment of the application can be a first control instruction, a second control instruction or a third control instruction.

[0069] After the agent in the i th agent device 102 obtains the first target control instruction output by the large language model in the i th agent device 102, the first target control instruction can be sent to the i th publishing device 103 through a corresponding instruction interface.

[0070] The i th publishing device 103 remains in a mute state in the case of receiving the first control instruction sent by the i th agent device 102; the i th publishing device 103 can publish the disaster warning information corresponding to the meteorological data to other agent devices 102 in the disaster warning system 101 and target user terminals in the case of receiving the second control instruction sent by the i th agent device 102; the i th publishing device 103 can publish the disaster broadcast information corresponding to the meteorological data to other agent devices 102 in the disaster warning system 101 in the case of receiving the third control instruction sent by the i th agent device 102.

[0071] It should be noted that the target user terminal in the embodiment of the application can be predefined based on actual conditions, for example, all user terminals in the region where the i th target meteorological monitoring station 104 is located can be determined as target user terminals.

[0072] The disaster warning system in the embodiment of the application includes a plurality of agent devices and a plurality of publishing devices, and a large language model is deployed in the agent devices, which can determine whether to publish disaster warnings or disaster broadcasts for climate data collected by climate monitoring stations based on the large language model without human intervention and expert judgment, more objectively, more accurately and more efficiently, avoid the influence of different climate data formats, quality and accuracy collected by different climate monitoring stations on the efficiency and accuracy of data processing of data provided by different meteorological monitoring stations, and effectively improve the accuracy, objectivity, uniformity and publishing efficiency of disaster warning publishing, so as to better reduce personnel casualties and property losses caused by disasters and improve the efficiency and effect of coping with disasters.

[0073] As an optional embodiment, the i th intelligent agent device 102 is further configured to, in a case where the current time is a target time, acquire weather information of a region where the i th target weather monitoring station 104 is located in the Internet between the current time and a last target time, input the weather information into a large language model in the i th intelligent agent device 102, acquire a second target control instruction output by the large language model in the i th intelligent agent device 102, and then send the second target control instruction to the i th publishing device 103. The second target control instruction is the first control instruction, a fourth control instruction carrying disaster warning information corresponding to the weather information, or a fifth control instruction carrying disaster broadcast information corresponding to the weather information. Any two adjacent target times are separated by a first preset time length.

[0074] The i th publishing device 103 is further configured to, in a case where the second target control instruction is received, publish disaster warning information corresponding to the weather information, publish disaster broadcast information corresponding to the weather information, or remain in a mute state in response to the second target control instruction.

[0075] Specifically, the intelligent agent device 102 in the embodiment of the present application can periodically acquire weather information in the Internet, and determine whether to issue disaster warning or disaster broadcast for the weather information based on the weather information.

[0076] It should be noted that the region where the i th target weather monitoring station 104 is located can be determined according to administrative division in the embodiment of the present application. In the embodiment of the present application, the region covered by each target weather monitoring station 104 can be uniformly divided according to the total region covered by each target weather monitoring station 104, and then the region where each target weather monitoring station 104 is located is determined based on the region.

[0077] It should be noted that the first preset time length between any two target times in the embodiment of the present application can be determined according to actual conditions and / or prior knowledge, and the first preset time length is not specifically limited in the embodiment of the present application.

[0078] Optionally, the first preset time length can be 1 second or several seconds.

[0079] In a case where the intelligent agent in the i th intelligent agent device 102 determines that the current time is a target time, the intelligent agent can call a control instruction for searching weather information in the Internet or a control instruction for using a search engine, browsing a webpage, executing code, file I / O, calling a mathematical tool, etc., to acquire weather information of a region where the i th target weather monitoring station 104 is located in the Internet between the current time and a last target time.

[0080] It should be noted that the meteorological information of the region where the i-th target meteorological monitoring station 104 is located in the Internet between the current time and the last target time obtained by the agent in the i-th agent device 102 is usually natural language information described in natural language. If the meteorological information of the region where the i-th target meteorological monitoring station 104 is located in the Internet between the current time and the last target time obtained by the agent in the i-th agent device 102 is not natural language information described in natural language, the agent in the i-th agent device 102 can call the operation resource to convert the meteorological information of the region where the i-th target meteorological monitoring station 104 is located in the Internet between the current time and the last target time into natural language information described in natural language.

[0081] After the agent in the i-th agent device 102 obtains the meteorological information of the region where the i-th target meteorological monitoring station 104 is located in the Internet between the current time and the last target time, the agent can input the meteorological information of the region where the i-th target meteorological monitoring station 104 is located in the Internet between the current time and the last target time into the large language model in the i-th agent device 102.

[0082] Based on the above meteorological information, the large language model in the i-th agent device 102 can determine whether a disaster warning or a disaster broadcast needs to be issued for the above meteorological information, and in the case of determining that a disaster warning needs to be issued for the above meteorological information, generate disaster warning information corresponding to the above meteorological information, and in the case of determining that a disaster broadcast needs to be issued for the above meteorological information, generate disaster broadcast information corresponding to the above meteorological information, and then output a first control instruction for instructing the publishing device 103 to remain in a mute state, a fourth control instruction indicating that a disaster warning needs to be issued for the above meteorological information and carrying the disaster warning information corresponding to the above meteorological information, or a fourth control instruction indicating that a disaster broadcast needs to be issued for the above meteorological information and carrying the disaster broadcast information corresponding to the above meteorological information.

[0083] It should be noted that the disaster warning information corresponding to the above meteorological information and the disaster broadcast information corresponding to the above meteorological data in the embodiment of the present application are natural language information described in natural language.

[0084] It should be noted that the second target instruction in the embodiment of the present application can be the first control instruction, the fourth control instruction or the fifth control instruction.

[0085] After the agent in the i-th agent device 102 obtains the second target control instruction output by the large language model in the i-th agent device 102, the agent can send the above second target control instruction to the i-th publishing device 103 through a corresponding instruction interface.

[0086] The i-th publishing device 103 can keep silent in response to the first control instruction in a case where the i-th publishing device 103 receives the first control instruction sent by the i-th intelligent agent device 102. The i-th publishing device 103 can publish disaster warning information corresponding to the meteorological information to other intelligent agent devices 102 in the disaster warning system 101 and to a target user terminal in response to the fourth control instruction in a case where the i-th publishing device 103 receives the fourth control instruction sent by the i-th intelligent agent device 102. The i-th publishing device 103 can publish disaster broadcast information corresponding to the meteorological information to other intelligent agent devices 102 in the disaster warning system 101 in response to the fifth control instruction in a case where the i-th publishing device 103 receives the fifth control instruction sent by the i-th intelligent agent device 102.

[0087] The intelligent agent device in the embodiment of the present application can also periodically acquire climate information in the Internet, and based on a large language model, more accurately and efficiently determine whether disaster broadcast or disaster warning needs to be published for the climate information, so that the climate monitoring station can maximize the acquisition of effective data from different data sources, and further improve the accuracy, objectivity, uniformity and publishing efficiency of disaster warning publication based on multi-source information.

[0088] As an optional embodiment, the i-th intelligent agent device 102 is further configured to, in a case where the i-th intelligent agent device 102 receives disaster information published by a j-th intelligent agent device 102 among the intelligent agent devices 102, if it is determined that the j-th target meteorological monitoring station 104 is a trusted target meteorological monitoring station 104 corresponding to the i-th target meteorological monitoring station 104, the disaster information published by the j-th intelligent agent device 102 belongs to the same disaster field as the meteorological data collected by the i-th target meteorological monitoring station 104, and the meteorological data collected by the i-th target meteorological monitoring station 104 matches the disaster information published by the j-th intelligent agent device 102 successfully, send a sixth control instruction carrying second natural language information to the i-th publishing device 103, the second natural language information indicating that the disaster information published by the j-th intelligent agent device 102 is trusted information, j∈{1,2,3,...,N-1,N}, i≠j, and the disaster information includes disaster broadcast information and disaster warning information.

[0089] The i-th publishing device 103 is further configured to, in a case where the i-th publishing device 103 receives the sixth control instruction, publish the second natural language information in response to the sixth control instruction.

[0090] Specifically, in the embodiment of the present application, the j-th intelligent agent device 102 can represent any intelligent agent device 102 among the intelligent agent devices 102 except the i-th intelligent agent device 102.

[0091] It should be noted that the trusted meteorological monitoring station corresponding to the i-th target meteorological monitoring station 104 can be determined according to actual conditions and / or prior knowledge in the embodiment of the present application.

[0092] The number of trusted meteorological monitoring stations corresponding to the i-th target meteorological monitoring station 104 can be multiple, and for the i-th target meteorological monitoring station 104, the trusted meteorological monitoring station corresponding to the i-th target meteorological monitoring station 104 has a higher credibility of the disaster broadcast information or the disaster warning information published.

[0093] In the case where the i-th intelligent agent device 102 receives the disaster information published by the j-th intelligent agent device 102, the intelligent agent in the i-th intelligent agent device 102 can determine whether the j-th target meteorological monitoring station 104 is a trusted meteorological monitoring station corresponding to the i-th target meteorological monitoring station 104 by querying the trusted meteorological monitoring station list of the i-th target meteorological monitoring station 104. The trusted meteorological monitoring station list of the i-th target meteorological monitoring station 104 includes each trusted meteorological monitoring station corresponding to the i-th target meteorological monitoring station 104.

[0094] It should be noted that the disaster information published by the j-th intelligent agent device 102 can be disaster broadcast information or disaster warning information. The disaster information published by the j-th intelligent agent device 102 is natural language information described in natural language.

[0095] After the intelligent agent in the i-th intelligent agent device 102 determines that the j-th target meteorological monitoring station 104 is a trusted meteorological monitoring station corresponding to the i-th target meteorological monitoring station 104, the intelligent agent in the i-th intelligent agent device 102 can determine whether the disaster information published by the j-th intelligent agent device 102 and the meteorological data collected by the i-th target meteorological monitoring station 104 belong to the same disaster field according to the data type and the data use, etc.

[0096] It should be noted that the disaster field in the embodiment of the present application can include but is not limited to the earthquake field, the tsunami field, the tornado field, the forest fire field, and the debris flow field, etc.

[0097] After the intelligent agent in the i-th intelligent agent device 102 determines that the disaster information published by the j-th intelligent agent device 102 and the meteorological data collected by the i-th target meteorological monitoring station 104 belong to the same disaster field, the intelligent agent in the i-th intelligent agent device 102 can also determine whether the disaster information published by the j-th intelligent agent device 102 and the meteorological data collected by the i-th target meteorological monitoring station 104 can be successfully matched through data simulation.

[0098] For example, the agent in the i th agent device 102 can output a control instruction for calling the simulation software in the same disaster field through the instruction interface, and perform simulation calculation on the disaster information published by the j th agent device 102 and the meteorological data collected by the i th target meteorological monitoring station 104. By comparing the simulation results corresponding to the disaster information published by the j th agent device 102 and the simulation results corresponding to the meteorological data collected by the i th target meteorological monitoring station 104, it is determined whether the disaster information published by the j th agent device 102 and the meteorological data collected by the i th target meteorological monitoring station 104 can be matched successfully. If the similarity of the simulation results corresponding to the disaster information published by the j th agent device 102 and the simulation results corresponding to the meteorological data collected by the i th target meteorological monitoring station 104 exceeds the similarity threshold, it is determined that the disaster information published by the j th agent device 102 and the meteorological data collected by the i th target meteorological monitoring station 104 are matched successfully, otherwise, it is determined that the disaster information published by the j th agent device 102 and the meteorological data collected by the i th target meteorological monitoring station 104 are not matched successfully.

[0099] If the agent in the i th agent device 102 determines that the disaster information published by the j th agent device 102 and the meteorological data collected by the i th target meteorological monitoring station 104 are matched successfully, the agent in the i th agent device 102 can consider that the credibility of the disaster information published by the j th agent device 102 is relatively high, and the agent in the i th agent device 102 can send a sixth control instruction to the i th publishing device 103. The above-mentioned sixth control instruction carries second natural language information described in natural language, indicating that the disaster information published by the j th agent device 102 is trusted information.

[0100] The i th publishing device 103 can publish the above-mentioned second natural language information to other agent devices 102 in the disaster warning system 101 when receiving the sixth control instruction sent by the i th agent device 102.

[0101] As an optional embodiment, the i-th agent device 102 is further configured to, in response to receiving the disaster information published by the j-th agent device 102, if it is determined that the j-th target meteorological monitoring station 104 is a trusted target meteorological monitoring station 104 corresponding to the i-th target meteorological monitoring station 104, the disaster information published by the j-th agent device 102 is in the same disaster field as the i-th target meteorological monitoring station 104, and the meteorological data collected by the i-th target meteorological monitoring station 104 does not match the disaster information published by the j-th agent device 102, but the distance between the j-th target meteorological monitoring station 104 and the i-th target meteorological monitoring station 104 is less than the distance threshold, send a seventh control instruction carrying third natural language information to the i-th publishing device 103, the third natural language information indicates that the disaster information published by the j-th agent device 102 is not trusted information, j ∈ {1, 2, 3,..., N-1, N}, and i ≠ j.

[0102] The i-th publishing device 103 is further configured to, in response to receiving the seventh control instruction, publish the third natural language information.

[0103] Specifically, in response to the i-th agent device 102 receiving the disaster information published by the j-th agent device 102, if the agent in the i-th agent device 102 determines that the j-th target meteorological monitoring station 104 is a trusted meteorological monitoring station corresponding to the i-th target meteorological monitoring station 104, the disaster information published by the j-th agent device 102 is in the same disaster field as the meteorological data collected by the i-th target meteorological monitoring station 104, but the disaster information published by the j-th agent device 102 does not match the meteorological data collected by the i-th target meteorological monitoring station 104, the agent in the i-th agent device 102 can determine whether the distance between the i-th target meteorological monitoring station 104 and the j-th target meteorological monitoring station 104 is less than the distance threshold.

[0104] It should be noted that the distance threshold in the embodiment of the present application can be determined according to actual conditions and / or prior knowledge. The distance threshold is not specifically limited in the embodiment of the present application.

[0105] Optionally, the distance threshold can be in the range of 400 km to 600 km, for example, the distance threshold can be 400 km, 500 km or 600 km.

[0106] Preferably, the distance threshold is 500 km.

[0107] If the agent in the i-th agent device 102 determines that the distance between the i-th target meteorological monitoring station 104 and the j-th target meteorological monitoring station 104 is less than the distance threshold, the agent in the i-th agent device 102 can consider that the credibility of the disaster information published by the j-th agent device 102 is low, and the agent in the i-th agent device 102 can send a seventh control instruction to the i-th publishing device 103, where the seventh control instruction carries third natural language information in natural language, indicating that the disaster information published by the j-th agent device 102 is not trusted information.

[0108] The i-th publishing device 103 can publish the third natural language information to other agent devices 102 in the disaster warning system 101 when receiving the seventh control instruction sent by the i-th agent device 102.

[0109] As an optional embodiment, the i-th agent device 102 is further configured to, when receiving the disaster information published by the j-th agent device 102 among the agent devices 102, if it is determined that the j-th target meteorological monitoring station 104 is not a trusted target meteorological monitoring station 104 corresponding to the i-th target meteorological monitoring station 104, send a sixth control instruction or a seventh control instruction to the i-th publishing device 103 based on the number of second natural language information and the number of third natural language information received by the i-th agent device 102 within a target period, where the start time of the target period is the time when the i-th agent device 102 receives the disaster information published by the j-th agent device 102, and the length of the target period is a second preset time length.

[0110] Specifically, when the i-th agent device 102 receives the disaster information published by the j-th agent device 102, if the agent in the i-th agent device 102 determines that the j-th target meteorological monitoring station 104 is not a trusted meteorological monitoring station corresponding to the i-th target meteorological monitoring station 104, the agent in the i-th agent device 102 can determine the credibility of the disaster information published by the j-th agent device 102 by numerical calculation based on the number of second natural language information and the number of third natural language information received by the i-th agent device 102 from other agent devices 102 within a second preset time length from the time when the i-th agent device 102 receives the disaster information published by the j-th agent device 102, and then send a sixth control instruction or a seventh control instruction to the i-th publishing device 103 based on the credibility of the disaster information published by the j-th agent device 102.

[0111] It should be noted that the second preset time length in the embodiment of the present application can be determined according to actual conditions and / or prior knowledge. The second preset time length in the embodiment of the present application is not specifically limited.

[0112] Optionally, the value range of the second preset time length can be 8 minutes to 12 minutes, for example, the value of the second preset time length can be 8 minutes, 10 minutes or 12 minutes.

[0113] Preferably, the value of the second preset time length can be 10 minutes.

[0114] As an optional embodiment, the ith agent device 102 is specifically configured to, based on the number of the second natural language information received by the ith agent device 102 within the target period, calculate a confirmation value corresponding to the disaster information published by the jth agent device 102, based on the number of the third natural language information received by the ith agent device 102 within the target period, calculate a negation value corresponding to the disaster information published by the jth agent device 102, for each second natural language information received by the ith agent device 102 within the target period, traverse to judge whether the target meteorological monitoring station 104 corresponding to the agent device 102 publishing each second natural language information is a trusted target meteorological monitoring station 104 corresponding to the ith target meteorological monitoring station 104, update the confirmation value corresponding to the disaster information published by the jth agent device 102, traverse to judge whether the target meteorological monitoring station 104 corresponding to the agent device 102 publishing each third natural language information is a trusted target meteorological monitoring station 104 corresponding to the ith target meteorological monitoring station 104, update the negation value corresponding to the disaster information published by the jth agent device 102 based on whether the target meteorological monitoring station 104 corresponding to the agent device 102 publishing each third natural language information is a trusted target meteorological monitoring station 104 corresponding to the ith target meteorological monitoring station 104, and further send a sixth control instruction or a seventh control instruction to the ith publishing device 103 based on the updated confirmation value and the negation value corresponding to the disaster information published by the jth agent device 102 and the ratio threshold.

[0115] Specifically, the agent in the ith agent device 102 calculates a determination value C corresponding to the disaster information published by the jth agent device 102 based on the number of the second natural language information published by other agent devices 102 and the trusted weight corresponding to the second natural language information received by the ith agent device 102 within a second preset time length from the time when the disaster information published by the jth agent device 102 is received.

[0116] The agent in the i th agent device 102 calculates a negative value N corresponding to the disaster information published by the j th agent device 102 based on the number of third natural language information published by other agent devices 102 received by the i th agent device 102 and the trust weight corresponding to the third natural language information within a second preset time length from the time when the i th agent device 102 receives the disaster information published by the j th agent device 102.

[0117] It should be noted that the trust weight corresponding to the second natural language information and the trust weight corresponding to the third natural language information in the embodiment of the application can be predefined based on prior knowledge and / or actual situation. The trust weight corresponding to the second natural language information and the trust weight corresponding to the third natural language information in the embodiment of the application are not specifically limited.

[0118] For each second natural language information received by the i th agent device 102 within the target period, it can be determined whether the target meteorological monitoring station 104 corresponding to the agent device 102 publishing each second natural language information is a trusted target meteorological monitoring station 104 corresponding to the i th target meteorological monitoring station 104. If the target meteorological monitoring station 104 corresponding to the agent device 102 publishing any second natural language information is not a trusted target meteorological monitoring station 104 corresponding to the i th target meteorological monitoring station 104, the confirmation value C corresponding to the disaster information published by the j th agent device 102 is increased by 1. If the target meteorological monitoring station 104 corresponding to the agent device 102 publishing any second natural language information is a trusted target meteorological monitoring station 104 corresponding to the i th target meteorological monitoring station 104, the confirmation value C corresponding to the disaster information published by the j th agent device 102 is increased by the first value corresponding to the agent device 102 publishing any second natural language information.

[0119] For example, for any second natural language information received by the i th agent device 102 within the target period, if the agent device 102 publishing the above-mentioned second natural language information is the m th agent device 102 among the agent devices 102, in the case that the m th target meteorological monitoring station 104 is not a trusted target meteorological monitoring station 104 corresponding to the i th target meteorological monitoring station 104, the confirmation value C corresponding to the disaster information published by the j th agent device 102 is increased by 1, and in the case that the m th target meteorological monitoring station 104 is a trusted target meteorological monitoring station 104 corresponding to the i th target meteorological monitoring station 104, the confirmation value C corresponding to the disaster information published by the j th agent device 102 is increased by Wherein, m∈{1,2,3,...,N-1,N}, i≠j≠m.

[0120] Wherein, represents a trust value of the mth target meteorological monitoring station 104 relative to the ith target meteorological monitoring station 104.

[0121] It should be noted that, in the embodiments of the present application can be used to measure the degree of reliance of the ith target meteorological monitoring station 104 on the mth target meteorological monitoring station 104, the trust value of the mth target meteorological monitoring station 104 relative to the ith target meteorological monitoring station 104 is higher, the higher the degree of reliance of the ith target meteorological monitoring station 104 on the mth target meteorological monitoring station 104.

[0122] It should be noted that, in the case that the mth target meteorological monitoring station 104 is a trusted target meteorological monitoring station 104 corresponding to the ith target meteorological monitoring station 104, the determined as the first value corresponding to the mth intelligent agent device 102.

[0123] For each third natural language information received by the ith intelligent agent device 102 within the target period, it can be determined whether the target meteorological monitoring station 104 corresponding to the intelligent agent device 102 that publishes each third natural language information is a trusted target meteorological monitoring station 104 corresponding to the ith target meteorological monitoring station 104. If the target meteorological monitoring station 104 corresponding to the intelligent agent device 102 that publishes any third natural language information is not a trusted target meteorological monitoring station 104 corresponding to the ith target meteorological monitoring station 104, the negative value N corresponding to the disaster information published by the jth intelligent agent device 102 is increased by 1. If the target meteorological monitoring station 104 corresponding to the intelligent agent device 102 that publishes any third natural language information is a trusted target meteorological monitoring station 104 corresponding to the ith target meteorological monitoring station 104, the confirmation value N corresponding to the disaster information published by the jth intelligent agent device 102 is increased by the second value corresponding to the intelligent agent device 102 that publishes any third natural language information.

[0124] For example, for any third natural language information received by the ith intelligent agent device 102 within the target period, if the intelligent agent device 102 that publishes the above third natural language information is the nth intelligent agent device 102 among the intelligent agent devices 102, in the case that the nth target meteorological monitoring station 104 is not a trusted target meteorological monitoring station 104 corresponding to the ith target meteorological monitoring station 104, the negative value N corresponding to the disaster information published by the jth intelligent agent device 102 is increased by 1. In the case that the nth target meteorological monitoring station 104 is a trusted target meteorological monitoring station 104 corresponding to the ith target meteorological monitoring station 104, the negative value N corresponding to the disaster information published by the jth intelligent agent device 102 is increased by wherein n∈{1,2,3,...,N-1,N}, i≠j≠m.

[0125] in, This represents the confidence value of the nth target meteorological monitoring station 104 relative to the ith target meteorological monitoring station 104.

[0126] Similarly, in the embodiments of the present invention It can be used to measure the degree of confidence of the i-th target meteorological monitoring station 104 over the n-th target meteorological monitoring station 104, and the confidence value of the n-th target meteorological monitoring station 104 relative to the i-th target meteorological monitoring station 104. The higher the value, the greater the confidence level of the i-th target meteorological monitoring station 104 in the n-th target meteorological monitoring station 104.

[0127] It should be noted that, in the case that the nth target meteorological monitoring station 104 is a credible target meteorological monitoring station 104 corresponding to the ith target meteorological monitoring station 104, in this embodiment of the invention, the target meteorological monitoring station 104 can be... The second value is determined to be the value corresponding to the nth intelligent agent device 102.

[0128] For each second natural language message received by the i-th intelligent agent device 102 within the target time period, the determination value C and denial value N corresponding to the disaster information released by the j-th intelligent agent device 102 are updated by iterating through and determining whether the target meteorological monitoring station 104 corresponding to the intelligent agent device 102 that released each second natural language message is a trusted target meteorological monitoring station 104 corresponding to the i-th target meteorological monitoring station 104, and by iterating through and determining whether the target meteorological monitoring station 104 corresponding to the intelligent agent device 102 that released each third natural language message is a trusted target meteorological monitoring station 104 corresponding to the i-th target meteorological monitoring station 104. final and denial value N final Then, the updated determination value C can be based on the disaster information released by the j-th intelligent agent device 102. final and denial value N final Whether the ratio is greater than the ratio threshold T, the credibility of the disaster information released by the j-th intelligent agent device 102 is determined.

[0129] It should be noted that, in this embodiment of the invention, the credibility of the disaster information released by the j-th intelligent agent device 102 can be determined by the following formula.

[0130] δ=max[p,tanh(C dic / k)]

[0131] C final =δ·C dic +(1-δ)·C others

[0132] N final = δ · N dic + (1 - δ) · N others

[0133]

[0134] wherein δ represents a parameter for adjusting the decision weight, C dic represents the sum of the trust values of each target meteorological monitoring station 104 corresponding to the trusted meteorological monitoring station of the i-th target meteorological monitoring station 104, among the agent devices 102 receiving each second natural language information published by the i-th agent device 102 within the target period, relative to the i-th target meteorological monitoring station 104; N dic represents the sum of the trust values of each target meteorological monitoring station 104 corresponding to the trusted meteorological monitoring station of the i-th target meteorological monitoring station 104, among the agent devices 102 receiving each third natural language information published by the i-th agent device 102 within the target period, relative to the i-th target meteorological monitoring station 104; k represents a parameter for adjusting the change rate, and p represents an upper limit of the parameter to avoid δ approaching 1; C ithers represents the sum of the confirmation values of each target meteorological monitoring station 104 not corresponding to the trusted meteorological monitoring station of the i-th target meteorological monitoring station 104, among the agent devices 102 receiving each second natural language information published by the i-th agent device 102 within the target period. final represents the sum of the denial values of each target meteorological monitoring station 104 not corresponding to the trusted meteorological monitoring station of the i-th target meteorological monitoring station 104, among the agent devices 102 receiving each third natural language information published by the i-th agent device 102 within the target period.

[0135] Optionally, in the embodiment of the present application, the value of k is 200, and the value of p is 0.5.

[0136] The meaning of the above formula includes that when the trust value of a certain trusted meteorological monitoring station corresponding to the i-th target meteorological monitoring station 104 becomes very large, the influence of the trusted meteorological monitoring station on the i-th target meteorological monitoring station 104 will become very large, and even if many other target meteorological monitoring stations 104 give opposite judgments, the trusted meteorological monitoring station will still win. Therefore, when the value of C dic is large enough, δ can be used to reduce the influence of C dic on the final decision, and the larger the value of C dic is, the smaller the weight of δ is, so that the final result can be more balanced in any case.

[0137] Figure 4 The variation diagram of the parameter for adjusting the determination weight in the disaster early warning system provided by the present application is shown in the following figure. The variation of the adjustment determination weight δ is shown in the following figure. Figure 4

[0138] The agent in the i th agent device 102 determines the updated determination value C final and the ratio of the negation value N final of the disaster information published by the j th agent device 102 based on the above formula. If the ratio is greater than the ratio threshold value T, the agent in the i th agent device 102 can consider that the disaster information published by the j th agent device 102 has a high degree of credibility, and the agent in the i th agent device 102 can send a sixth control instruction to the i th publishing device 103. The above sixth control instruction carries second natural language information in natural language, which indicates that the disaster information published by the j th agent device 102 is trusted information.

[0139] The agent in the i th agent device 102 determines the updated determination value C final and the ratio of the negation value N final of the disaster information published by the j th agent device 102 based on the above formula. If the ratio is not greater than the ratio threshold value T, the agent in the i th agent device 102 can consider that the disaster information published by the j th agent device 102 has a low degree of credibility, and the agent in the i th agent device 102 can send a seventh control instruction to the i th publishing device 103. The above seventh control instruction carries third natural language information in natural language, which indicates that the disaster information published by the j th agent device 102 is not trusted information.

[0140] As an optional embodiment, the i th agent device 102 is further configured to, in a case where the disaster information published by the j th agent device 102 among the agent devices 102 is received, send a first control instruction to the i th publishing device 103 if it is determined that the j th target meteorological monitoring station 104 is a trusted target meteorological monitoring station 104 corresponding to the i th target meteorological monitoring station 104, the disaster information published by the j th agent device 102 belongs to the same disaster field as the i th target meteorological monitoring station 104, the meteorological data collected by the i th target meteorological monitoring station 104 and the disaster information published by the j th agent device 102 are not matched successfully, and the distance between the j th target meteorological monitoring station 104 and the i th target meteorological monitoring station 104 is not less than a distance threshold value, j ∈ {1, 2, 3,..., N-1, N}, i ≠ j

[0141] ​It should be noted that the agent in the i th agent device 102 determines that the j th target meteorological monitoring station 104 is a trusted meteorological monitoring station corresponding to the i th target meteorological monitoring station 104, the disaster information published by the j th agent device 102 and the meteorological data collected by the i th target meteorological monitoring station 104 belong to the same disaster field, the disaster information published by the j th agent device 102 and the meteorological data collected by the i th target meteorological monitoring station 104 are not matched successfully, and the distance between the i th target meteorological monitoring station 104 and the j th target meteorological monitoring station 104 is not less than the distance threshold, the agent in the i th agent device 102 can send the first control instruction to the i th publishing device 103.

[0142] The i th publishing device 103 can keep silent state in response to the first control instruction when receiving the first control instruction sent by the i th agent device 102.

[0143] The agent recognition in the embodiment of the application realizes the data interaction between the meteorological monitoring stations automatically through the interaction between the agent devices corresponding to the target meteorological monitoring stations, without manual intervention and expert judgment, so that each meteorological monitoring station can maximize the effective data published by other meteorological monitoring stations, and the efficiency and objectivity of disaster warning are improved.

[0144] As an optional embodiment, the i th agent device 102 is further configured to, when receiving the disaster information published by the j th agent device 102 among the agent devices 102, if it is determined that the j th target meteorological monitoring station 104 is a trusted target meteorological monitoring station 104 corresponding to the i th target meteorological monitoring station 104, input the disaster information published by the j th agent device 102 into the large language model in the i th agent device 102, obtain the third target control instruction output by the large language model in the i th agent device 102, and the third target control instruction is the first control instruction, the eighth control instruction carrying the disaster warning information corresponding to the disaster information published by the j th agent device 102, or the ninth control instruction carrying the disaster broadcast information corresponding to the disaster information published by the j th agent device 102.

[0145] The i th publishing device 103 is further configured to, when receiving the third target control instruction, publish the disaster warning information corresponding to the disaster information published by the j th agent device 102, publish the disaster broadcast information corresponding to the disaster information published by the j th agent device 102, or keep silent state in response to the third target control instruction.

[0146] Specifically, in a case where the i th agent device 102 receives disaster information published by the j th agent device 102, if the agent in the i th agent device 102 determines that the j th target meteorological monitoring station 104 is a trusted meteorological monitoring station corresponding to the i th target meteorological monitoring station 104, the agent in the i th agent device 102 can input the disaster information published by the j th agent device 102 into the large language model in the i th agent device 102.

[0147] The large language model in the i th agent device 102 can determine whether a disaster warning or a disaster broadcast needs to be issued for the disaster information published by the j th agent device 102 based on the disaster information, and in a case where it is determined that a disaster warning needs to be issued for the disaster information, generate disaster warning information corresponding to the disaster information, and in a case where it is determined that a disaster broadcast needs to be issued for the disaster information, generate disaster broadcast information corresponding to the disaster information, and then output a first control instruction for instructing the publishing device 103 to remain in a silent state, an eighth control instruction indicating that a disaster warning needs to be issued for the disaster information and carrying the disaster warning information corresponding to the disaster information, or a ninth control instruction indicating that a disaster broadcast needs to be issued for the disaster information and carrying the disaster broadcast information corresponding to the disaster information.

[0148] It should be noted that the third target instruction in the embodiment of the present application can be the first control instruction, the eighth control instruction or the ninth control instruction.

[0149] After the agent in the i th agent device 102 obtains the third target control instruction output by the large language model in the i th agent device 102, the agent can send the third target control instruction to the i th publishing device 103 through a corresponding instruction interface.

[0150] In a case where the i th publishing device 103 receives the first control instruction sent by the i th agent device 102, the i th publishing device 103 can remain in a silent state in response to the first control instruction; in a case where the i th publishing device 103 receives the eighth control instruction sent by the i th agent device 102, the i th publishing device 103 can publish the disaster warning information corresponding to the disaster information to other agent devices 102 in the disaster warning system 101 and to target user terminals in response to the eighth control instruction; and in a case where the i th publishing device 103 receives the ninth control instruction sent by the i th agent device 102, the i th publishing device 103 can publish the disaster broadcast information corresponding to the disaster information to other agent devices 102 in the disaster warning system 101 in response to the ninth control instruction.

[0151] As an optional embodiment, the ith agent device 102 is further configured to increase the trust value of the jth target meteorological monitoring station 104 relative to the ith target meteorological monitoring station 104 by 1 after sending the sixth control instruction to the ith publishing device 103, and decrease the trust value of the jth target meteorological monitoring station 104 relative to the ith target meteorological monitoring station 104 by 1 in the case of the seventh control instruction sent to the ith publishing device 103.

[0152] The ith agent device 102 is further configured to eliminate the jth target meteorological monitoring station 104 from the corresponding trusted target meteorological monitoring station 104 of the ith target meteorological monitoring station 104 in the case that the trust value of the jth target meteorological monitoring station 104 relative to the ith target meteorological monitoring station 104 is determined to be 0.

[0153] Specifically, the agent in the ith agent device 102 sends the sixth control instruction to the ith publishing device 103, indicating that the agent in the ith agent device 102 considers the trust degree of the disaster information published by the jth agent device 102 to be relatively high. Therefore, after the agent in the ith agent device 102 sends the sixth control instruction to the ith publishing device 103, the trust value of the jth target meteorological monitoring station 104 relative to the ith target meteorological monitoring station 104 can be increased by 1.

[0154] On the contrary, the agent in the ith agent device 102 sends the seventh control instruction to the ith publishing device 103, indicating that the agent in the ith agent device 102 considers the trust degree of the disaster information published by the jth agent device 102 to be relatively low. Therefore, after the agent in the ith agent device 102 sends the seventh control instruction to the ith publishing device 103, the trust value of the jth target meteorological monitoring station 104 relative to the ith target meteorological monitoring station 104 can be decreased by 1.

[0155] The agent in the ith agent device 102 can eliminate the jth target meteorological monitoring station 104 from the corresponding trusted meteorological monitoring station of the ith target meteorological monitoring station 104 in the case that the trust value of the jth target meteorological monitoring station 104 relative to the ith target meteorological monitoring station 104 is determined to be 0.

[0156] ​​​As an optional embodiment, the i th intelligent agent device 102 is further configured to determine the j th target meteorological monitoring station 104 as a candidate target meteorological monitoring station corresponding to the i th target meteorological monitoring station 104 in a case that the i th intelligent agent device 102 determines that the j th target meteorological monitoring station 104 is not the candidate target meteorological monitoring station corresponding to the i th target meteorological monitoring station 104 and sends the sixth control instruction to the i th publishing device 103, and the i th intelligent agent device 102 is further configured to determine the j th target meteorological monitoring station 104 as a trusted target meteorological monitoring station corresponding to the i th target meteorological monitoring station 104 in a case that the number of times of sending the sixth control instruction to the i th publishing device 103 exceeds the threshold number of times after the j th target meteorological monitoring station 104 is determined as the candidate target meteorological monitoring station corresponding to the i th target meteorological monitoring station 104.

[0157] Specifically, the intelligent agent in the i th intelligent agent device 102 sends the sixth control instruction to the i th publishing device 103, indicating that the intelligent agent in the i th intelligent agent device 102 considers that the j th intelligent agent device 102 publishes disaster information with a high degree of trust, and thus, after the intelligent agent in the i th intelligent agent device 102 sends the sixth control instruction to the i th publishing device 103, the intelligent agent in the i th intelligent agent device 102 can determine whether the j th target meteorological monitoring station 104 is a candidate target meteorological monitoring station corresponding to the i th target meteorological monitoring station 104 by querying a candidate meteorological monitoring station list of the i th target meteorological monitoring station 104. The candidate meteorological monitoring station list of the i th target meteorological monitoring station 104 includes each candidate meteorological monitoring station corresponding to the i th target meteorological monitoring station 104.

[0158] If the intelligent agent in the i th intelligent agent device 102 determines that the j th target meteorological monitoring station 104 is not a candidate target meteorological monitoring station corresponding to the i th target meteorological monitoring station 104, the intelligent agent in the i th intelligent agent device 102 can determine the j th target meteorological monitoring station 104 as a candidate target meteorological monitoring station corresponding to the i th target meteorological monitoring station 104.

[0159] After the intelligent agent in the i th intelligent agent device 102 determines the j th target meteorological monitoring station 104 as a candidate target meteorological monitoring station corresponding to the i th target meteorological monitoring station 104, if the intelligent agent in the i th intelligent agent device 102 determines that the number of times of sending the sixth control instruction to the i th publishing device 103 exceeds the threshold number of times, the intelligent agent in the i th intelligent agent device 102 can determine the j th target meteorological monitoring station 104 as a trusted target meteorological monitoring station corresponding to the i th target meteorological monitoring station 104.

[0160] It should be noted that the threshold number of times in the embodiment of the present application can be predefined based on prior knowledge and / or actual conditions. The threshold number of times is not specifically limited in the embodiment of the present application.

[0161] Optionally, the value of the number threshold in the embodiment of the present application can be 8 to 12, for example, the value of the number threshold can be 8, 10 or 12.

[0162] Preferably, the value of the number threshold can be 10.

[0163] The embodiment of the present application can determine the credibility of the information provided by each climate monitoring station through the credibility determination of the agent device constantly evolving and updating, so that each target climate monitoring station can maximize the effective data of different data sources, and can also solve the cold start problem of the new climate monitoring station.

[0164] Figure 5 is the flowchart of the disaster warning system provided by the present application. The disaster warning method provided by the present application is realized based on the disaster warning system 101 as above. As shown in the figure, Figure 5 The method comprises the following steps: step 501, in the case of obtaining meteorological data collected by any target meteorological monitoring station 104, performing disaster prediction on the meteorological data to obtain a disaster prediction result corresponding to the meteorological data described in natural language.

[0165] Step 502, inputting the disaster prediction result corresponding to the meteorological data into a large language model in the agent device 102 corresponding to any target meteorological monitoring station 104, obtaining a first target control instruction output by the large language model, and the first target control instruction is a first control instruction for instructing the publishing device 103 to keep silent, a second control instruction carrying disaster warning information corresponding to the meteorological data, or a third control instruction carrying disaster broadcast information corresponding to the meteorological data.

[0166] Step 503, sending the first target control instruction to the publishing device 103 corresponding to any target meteorological monitoring station 104, so that the publishing device 103 corresponding to any target meteorological monitoring station 104 publishes disaster broadcast information corresponding to the meteorological data, publishes disaster warning information corresponding to the meteorological data or keeps silent in response to the first target control instruction.

[0167] It should be noted that the specific implementation steps of the disaster warning method provided by the present application can refer to the contents in the above embodiments, and the embodiment of the present application will not be repeated.

[0168] The embodiment of the present application can determine whether to issue a disaster warning or a disaster broadcast for the climate data collected by the climate monitoring station based on a large language model without human intervention and expert judgment, more objectively, more accurately and more efficiently, avoid the influence of different climate data formats, quality and accuracy collected by different climate monitoring stations on the efficiency and accuracy of data processing of the data provided by different meteorological monitoring stations, effectively improve the accuracy, objectivity, uniformity and efficiency of disaster warning issuance, thereby better reducing personnel casualties and property losses caused by disasters and improving the efficiency and effectiveness of disaster response.

[0169] Figure 6 An example of an entity structure diagram of an electronic device is shown in Figure 6 The electronic device can include a processor 610, a communications interface 620, a memory 630 and a communications bus 640, wherein the processor 610, the communications interface 620 and the memory 630 communicate with each other through the communications bus 640. The processor 610 can call the logical instructions in the memory 630 to execute the disaster warning method, which includes: in the case of obtaining meteorological data collected by any target meteorological monitoring station, performing disaster prediction on the meteorological data to obtain a natural language description of the disaster prediction result corresponding to the meteorological data; inputting the disaster prediction result corresponding to the meteorological data into a large language model in an agent device corresponding to any target meteorological monitoring station, obtaining a first target control instruction output by the large language model, the first target control instruction being a first control instruction for instructing a release device to remain in a silent state, a second control instruction carrying disaster warning information corresponding to the meteorological data or a third control instruction carrying disaster broadcast information corresponding to the meteorological data; sending the first target control instruction to a release device corresponding to any target meteorological monitoring station, so that the release device corresponding to any target meteorological monitoring station responds to the first target control instruction to release the disaster broadcast information corresponding to the meteorological data, release the disaster warning information corresponding to the meteorological data or remain in a silent state.

[0170] In addition, the logic instructions in the memory 630 described above can be implemented in the form of a software function unit and sold or used as an independent product, and can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0171] In another aspect, the present application also provides a computer program product, which comprises a computer program, the computer program can be stored on a non-transitory computer readable storage medium, and the computer program can be executed by a processor, so that the computer can execute the disaster warning method provided by the above-mentioned method, the method comprises: in the case of obtaining the meteorological data collected by any target meteorological monitoring station, performing disaster prediction on the meteorological data to obtain the disaster prediction result corresponding to the meteorological data described in natural language; inputting the disaster prediction result corresponding to the meteorological data into the large language model in the agent device corresponding to any target meteorological monitoring station, obtaining the first target control instruction output by the large language model, the first target control instruction is a first control instruction for instructing the release device to keep silent state, a second control instruction carrying disaster warning information corresponding to the meteorological data or a third control instruction carrying disaster broadcast information corresponding to the meteorological data; sending the first target control instruction to the release device corresponding to any target meteorological monitoring station, so that the release device corresponding to any target meteorological monitoring station releases the disaster broadcast information corresponding to the meteorological data, releases the disaster warning information corresponding to the meteorological data or keeps silent state in response to the first target control instruction.

[0172] In yet another aspect, the present application also provides a non-transitory computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the disaster warning method provided by any of the above methods, and the method comprises: in the case of obtaining meteorological data collected by any target meteorological monitoring station, performing disaster prediction on the meteorological data to obtain a disaster prediction result corresponding to the meteorological data in natural language; inputting the disaster prediction result corresponding to the meteorological data into a large language model in an agent device corresponding to any target meteorological monitoring station, obtaining a first target control instruction output by the large language model, the first target control instruction being a first control instruction for instructing a publishing device to remain in a silent state, a second control instruction carrying disaster warning information corresponding to the meteorological data, or a third control instruction carrying disaster broadcast information corresponding to the meteorological data; and sending the first target control instruction to a publishing device corresponding to any target meteorological monitoring station, so that the publishing device corresponding to any target meteorological monitoring station publishes disaster broadcast information corresponding to the meteorological data, publishes disaster warning information corresponding to the meteorological data, or remains in a silent state in response to the first target control instruction.

[0173] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e., they can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0174] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be implemented by means of software and the necessary general hardware platform, and of course, can also be implemented by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0175] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A disaster warning system, characterized by, Comprising: A plurality of agent devices and a plurality of publishing devices, wherein a large language model is deployed in the agent devices; i Each of the target meteorological monitoring stations corresponds to a first agent device in each of the agent devices and a first publishing device in each of the publishing devices. i The first agent device in each of the agent devices is respectively connected in communication with the first target meteorological monitoring station and the first publishing device in each of the publishing devices. i The first agent device in each of the agent devices is respectively connected in communication with the first target meteorological monitoring station and the first publishing device in each of the publishing devices. i The first agent device in each of the agent devices is respectively connected in communication with the first target meteorological monitoring station and the first publishing device in each of the publishing devices. i The first agent device in each of the agent devices is respectively connected in communication with the first target meteorological monitoring station and the first publishing device in each of the publishing devices. i The first agent device in each of the agent devices is respectively connected in communication with the first target meteorological monitoring station and the first publishing device in each of the publishing devices. The first agent device in each of the agent devices is respectively connected in communication with the first target meteorological monitoring station and the first publishing device in each of the publishing devices. N The first agent device in each of the agent devices The first i Intelligent agent device is used to obtain the meteorological data collected by the first i Target meteorological monitoring station, and perform disaster prediction on the meteorological data. After obtaining the natural language description of the disaster prediction result corresponding to the meteorological data, the meteorological data corresponding disaster prediction result is input into the large language model in the first i Intelligent agent device, and the first target control instruction output by the large language model in the first i Intelligent agent device is obtained, and then the first target control instruction is sent to the first i Publishing device. The first target control instruction is a first control instruction for instructing the publishing device to remain in a silent state, a second control instruction carrying disaster warning information corresponding to the meteorological data, or a third control instruction carrying disaster broadcast information corresponding to the meteorological data. The first target control instruction is used to control the first target device to release the weather data corresponding to the disaster broadcast information, release the weather data corresponding to the disaster warning information, or keep a mute state. i The first target control instruction is used to control the first target device to release the weather data corresponding to the disaster broadcast information, release the weather data corresponding to the disaster warning information, or keep a mute state. The first i The first j The first j The first i The first j The first i The first i The first j The first i The first j The first The first disaster broadcast information and disaster warning information. The first i The first publishing device is further configured to, in a case where the sixth control instruction is received, publish the second natural language information in response to the sixth control instruction.

2. The disaster warning system according to claim 1, characterized by, The first i Intelligent agent device is also used to acquire the weather information of the area where the first i Target meteorological monitoring station in the Internet between the current time and the last target time in the case of the current time being the target time, input the weather information into the large language model in the first i Intelligent agent device, acquire the second target control instruction output by the large language model in the first i Intelligent agent device, and then send the second target control instruction to the first i Publishing device. The second target control instruction is the first control instruction, a fourth control instruction carrying disaster warning information corresponding to the weather information, or a fifth control instruction carrying disaster broadcast information corresponding to the weather information. Any two adjacent target times are separated by a first preset time length. The first i The first publishing device is further configured to, in a case where the second target control instruction is received, publish disaster warning information corresponding to the weather information, publish disaster broadcast information corresponding to the weather information, or remain in a mute state in response to the second target control instruction.

3. The disaster warning system according to claim 1, characterized by, The first i The first j The first j The first i The first j The first i The first i The first j The first j The first i The first i The first j The first The first i The publishing device is also configured to, upon receiving the seventh control command, publish the third natural language information in response to the seventh control command.

4. The disaster warning system according to claim 3, characterized by The first i The first j In a case where the first j If it is determined that the first i target meteorological monitoring station, the first i The first i The first i The first j The first 5. The disaster warning system according to claim 4, characterized by The first i The first i The first j The first i The first j The first i The first i The first j The first i The first i The first j The first j The first i The first 6. The disaster alerting system of claim 1, wherein, The first i The first j The first j The first i The first j The first i The first i The first j The first j The first i The first i The first The first ​ 7. The disaster warning system according to claim 3, characterized by The first i Intelligent agent device is also used for increasing the trust value of the first i Target meteorological monitoring station relative to the first j Target meteorological monitoring station by 1 after sending the sixth control instruction to the first i Publishing device, and decreasing the trust value of the first i Target meteorological monitoring station relative to the first j Target meteorological monitoring station by 1 in the case of the seventh control instruction sent to the first i Publishing device. The first i Intelligent agent device is further configured to eliminate the first j Target meteorological monitoring station from the corresponding trusted target meteorological monitoring station of the first i Target meteorological monitoring station in a case where the trusted value of the first i Target meteorological monitoring station relative to the first j Target meteorological monitoring station is 0.

8. The disaster alerting system of claim 1, wherein, The first target meteorological monitoring station is determined as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station. i The first intelligent agent device is further configured to determine the first target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station. j The first intelligent agent device is further configured to determine the first target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station. i The first intelligent agent device is further configured to determine the first target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station. i The first intelligent agent device is further configured to determine the first target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station. j The first intelligent agent device is further configured to determine the first target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station. i The first intelligent agent device is further configured to determine the first target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station. i The first intelligent agent device is further configured to determine the first target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station. j The first intelligent agent device is further configured to determine the first target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station. i The first intelligent agent device is further configured to determine the first target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station. i The first intelligent agent device is further configured to determine the first target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station. j The first intelligent agent device is further configured to determine the first target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station. i The first intelligent agent device is further configured to determine the first target meteorological monitoring station as the alternative target meteorological monitoring station corresponding to the first target meteorological monitoring station.

9. The disaster warning system according to any one of claims 1 to 8, characterized by, The first i The first j The first j The first i The first j The first i The first i The first j The first j The first The first ​ The first i The publishing device is also configured to, upon receiving the third target control command, publish the third target control command in response to the third target control command. j The disaster warning information corresponding to the disaster information released by the intelligent agent device, and the release of the first j The disaster information released by the intelligent device corresponds to the disaster broadcast information or remains silent.

10. A disaster warning method based on the disaster warning system according to any one of claims 1 to 9, characterized by, In the case of acquiring any target meteorological monitoring station collected meteorological data, the meteorological data is used for disaster prediction, and the disaster prediction result corresponding to the meteorological data described in natural language is obtained; The disaster prediction result corresponding to the meteorological data is input into the large language model in the agent device corresponding to any target meteorological monitoring station, and the first target control instruction output by the large language model is obtained. The first target control instruction is a first control instruction for instructing the publishing device to remain in a silent state, a second control instruction carrying disaster warning information corresponding to the meteorological data, or a third control instruction carrying disaster broadcast information corresponding to the meteorological data; The first target control instruction is sent to the publishing device corresponding to any target meteorological monitoring station, so that the publishing device corresponding to any target meteorological monitoring station publishes disaster broadcast information corresponding to the meteorological data, publishes disaster warning information corresponding to the meteorological data, or remains in a silent state in response to the first target control instruction. ​

Images