Real-time situation updating and analyzing method, device and equipment based on dynamic ontology

Abstract

The invention discloses a real-time situation updating and analyzing method, device and equipment based on a dynamic ontology. The method comprises the following steps: constructing a cross-domain joint combat static ontology; acquiring real-time battlefield situation data from different combat field sources; directionally capturing target data in the real-time battlefield situation data; performing multi-source data fusion on the multiple pieces of target data; performing data standardization processing to generate structured situation data; extracting identifiers of all situation instances, if the situation instances do not exist in the static ontology, adding the situation instances and attributes thereof to the static ontology, and performing fuzzy matching; the file in the Turtle format is converted into a graph object; updating the static ontology by adding, deleting and modifying the graph object; dynamically analyzing the updated content and forming knowledge accumulation, and taking the knowledge accumulation as an input variable of next multi-source data fusion; and performing reasoning verification on the static ontology and the dynamic ontology. According to the method, the automation degree is high, and the ability of situation analysis and auxiliary decision making is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of situation fusion and analysis decision-making, and in particular to a real-time situation updating and analyzing method based on dynamic ontology, a real-time situation updating and analyzing device based on dynamic ontology, and electronic equipment. Background Art

[0002] Knowledge engineering in the military field has always been a core issue that needs to be solved in system planning and combat command. Facing the problems of knowledge discovery and ontology construction in complex combat scenarios such as joint operations and cross-domain operations, domestic and foreign scholars have carried out some research. First, by identifying and recording the human knowledge used by senior military professionals to solve advanced problems, military knowledge can be classified, expressed, shared and reused, and integrated into the automation system, so that artificial intelligence can truly play a role in advanced military applications; second, in order to solve the problem of interoperability caused by the heterogeneity of multi-source data on the battlefield, grammatical differences and semantic differences can be eliminated based on the consensus framework of ontology.

[0003] However, the traditional ontology construction method is difficult to adapt to the current big data era:

[0004] From the perspective of the autonomous construction process of the ontology model, the degree of expert participation is still high, which limits the portability and learning ability of the ontology, and the degree of automation needs to be improved;

[0005] From the specific implementation of ontology model construction, the efficiency of data extraction, matching and fusion needs to be improved.

[0006] From the perspective of dynamic ontology updating, the accuracy of matching static ontology with target data needs to be improved.

[0007] From the perspective of the meaning of changes in the ontology model, in the military field, especially in the combat field, the ontology construction focuses on the changes in the inherent attributes of the entity, but ignores the construction of exogenous attributes such as the logical relationship, role cognition, and behavior of the entity in the combat process. Therefore, it is difficult to provide deeper semantic support for the cognition and understanding of the dynamic process of the continuous evolution of entities in the combat system over time. Summary of the invention

[0008] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes a real-time situation update and analysis method based on dynamic ontology, which has the advantages of high automation, high efficiency of data extraction, matching and fusion, high matching accuracy, etc., and has the ability of situation analysis and auxiliary decision-making.

[0009] The present invention also proposes a real-time situation updating and analyzing device based on dynamic ontology.

[0010] The invention also provides an electronic device.

[0011] According to the first aspect of the present invention, a real-time situation update and analysis method based on dynamic ontology is used for multi-source battlefield situation awareness and analysis decision-making in cross-domain joint combat scenarios, wherein the cross-domain joint combat scenarios include land, sea, air, network and electromagnetic fields, and the method includes:

[0012] Constructing a cross-domain joint combat static ontology, and converting the cross-domain joint combat static ontology into a Turtle format file;

[0013] Obtain multiple real-time battlefield situation data from different combat domain sources;

[0014] Directedly grabbing target data in the real-time battlefield situation data according to the grabbing constraint conditions;

[0015] Perform multi-source data fusion on the plurality of target data, and generate multi-source complementary fusion data through data redundancy processing, data association analysis and data merging, including equipment parameter redundancy processing, weapon mount association relationship establishment and similar resource merging and division;

[0016] Performing data standardization processing on the multi-source complementary fusion data to generate structured situation data, wherein the structured situation data has a consistent format and structure and is provided with an identifier for matching a battlefield situation instance;

[0017] Extract identifiers of all battlefield situation instances; if the battlefield situation instance does not exist in the cross-domain joint operations static ontology, add the battlefield situation instance and its attributes to the cross-domain joint operations static ontology, and perform fuzzy matching on the category of the battlefield situation instance according to edit distance similarity, pinyin similarity and cosine similarity;

[0018] Using a graph method to load and parse the Turtle format file, and convert the Turtle format file into a graph object;

[0019] Updating the cross-domain joint operations static ontology by adding, deleting and modifying the graph objects;

[0020] Dynamically analyzing the updated content of the cross-domain joint operations static ontology and forming knowledge accumulation, and using the knowledge accumulation as an input variable for the next multi-source data fusion;

[0021] The cross-domain joint combat static ontology and the updated dynamic ontology are subjected to reasoning verification, including equipment effectiveness analysis, combat readiness deployment, cost-benefit analysis and combat outcome prediction; wherein the dynamic analysis includes:

[0022] Design prompt words for the large language model, use the large language model to predict enemy strategies and intentions, and provide our own troop deployment and firepower allocation strategies.

[0023] According to the real-time situation update and analysis method based on dynamic ontology in the embodiment of the present invention, by establishing a general workflow for ontology construction and update, it is possible to realize the automatic standardized construction of data to ontology model, which is conducive to improving the degree of automation; the use of intelligent data integration and standardization technology can improve the efficiency of data extraction, matching and fusion; the use of multi-dimensional fuzzy weighted matching algorithm to improve the accuracy of matching between the static ontology of cross-domain joint operations and target data; the use of a large language model to analyze and interpret the dynamically updated part of the static ontology of cross-domain joint operations, and then give predictions and suggestions to guide the decision-making process.

[0024] In addition, the real-time situation update and analysis method based on dynamic ontology according to the embodiment of the present invention also has the following additional technical features:

[0025] According to some embodiments of the present invention, constructing a cross-domain joint operations static ontology includes:

[0026] Collect literature, expert knowledge and historical data based on needs analysis;

[0027] Extract and integrate effective data;

[0028] Carry out hierarchical design and relationship construction to form conceptual modeling;

[0029] Generate the cross-domain joint combat static ontology.

[0030] According to some embodiments of the present invention, the directional grasping includes:

[0031] Acquire a first readable file containing the grabbing constraint conditions and a second readable file containing the real-time battlefield situation data;

[0032] Reading the first readable file and converting the grabbing constraint condition into a first dictionary, reading the second readable file and converting the real-time battlefield situation data into a second dictionary;

[0033] For each set of constraint information in the first dictionary, extract the location information and attribute information of the target data, and construct a capture field using the location information and the attribute information;

[0034] For each of the captured fields, capture the target data in the second dictionary;

[0035] The captured target data is converted into a third readable file and outputted.

[0036] In some embodiments of the present invention, the multi-source data fusion includes:

[0037] Acquire a plurality of the third readable files, and convert the third readable files into a third dictionary;

[0038] For each key of any two of the third dictionaries, if the keys of the two third dictionaries exist and have the same key value, the key value is updated using the result of fusion using the complementary fusion function; otherwise, the key values ​​of the two third dictionaries are updated with the key value of the existing key.

[0039] According to some embodiments of the present invention, the data normalization process includes:

[0040] The multi-source complementary fusion data is flattened, cleaned, converted, deduplicated, detected as anomaly, and corrected.

[0041] According to some embodiments of the present invention, the fuzzy matching is performed according to the following formula:

[0042] s=A·s 1 +B·s 2 +C·s 3

[0043] Where s is the comprehensive similarity, s∈[0,1]; s 1 is the edit distance similarity; s 2 is the pinyin similarity; s 3 is the cosine similarity; A, B, C are weighted coefficients, A+B+C=1;

[0044] If the lengths of the strings are different, then A is greater than B; if the lengths of the strings are the same, then A is less than B.

[0045] According to some embodiments of the present invention, the prompt words of the large language model are "You are a senior military tactical analyst with rich battlefield situation awareness and analysis experience. Please carefully analyze the updated content of the cross-domain joint operations static ontology, evaluate the number and location of military units / equipment of both sides, intelligence data, communication status and combat objectives, and based on these analyses, speculate on the tactical actions that the enemy may take against us, and how we should formulate response strategies";

[0046] The analysis includes:

[0047] The enemy and our own combat objectives, changes in troop strength, changes in the number and location of weapons and equipment, changes in intelligence collection, communication capabilities, and other important situation information; the analysis results include:

[0048] Tactics that the enemy is preparing to take: Based on the enemy's dynamically updated situation information, the specific tactical actions that the enemy is preparing to take include raids, reconnaissance, and long-range strikes;

[0049] Our response strategy: Our response measures include strengthening air defense capabilities, strengthening reconnaissance and surveillance, and taking actions to confuse the enemy;

[0050] The situation data used for the dynamic analysis include: changes in the physical status of both the enemy and us, including alert status, standby status, intelligence collection activities, equipment status, and combat level.

[0051] According to the second aspect of the present invention, a real-time situation update and analysis device based on a dynamic ontology is used for multi-source battlefield situation perception and analysis decision-making in a cross-domain joint combat scenario, wherein the cross-domain joint combat scenario includes land, sea, air, network and electromagnetic fields, and the device includes: a construction module, wherein the construction module is used to construct a cross-domain joint combat static ontology and convert the cross-domain joint combat static ontology into a Turtle format file; an acquisition module, wherein the acquisition module is used to acquire multiple real-time battlefield situation data from different combat domains; a directional capture module, wherein the directional capture module directional captures target data in the real-time battlefield situation data according to capture constraints; a multi-source data fusion module, wherein the multi-source data fusion module is used to perform multi-source data fusion on multiple target data, and generate multi-source complementary fusion data through data redundancy processing, data association analysis and data merging, including equipment parameter redundancy processing, weapon mounting association relationship establishment and similar resource merging and division; a data standardization processing module, wherein the data standardization processing module is used to perform data standardization processing on the multi-source complementary fusion data to generate structured situation data, wherein the format and structure of the structured situation data are consistent and an identifier for matching battlefield situation instances is provided; an extraction module, wherein the extraction The module is used to extract identifiers of all battlefield situation instances. If the battlefield situation instance does not exist in the cross-domain joint operations static ontology, the battlefield situation instance and its attributes are added to the cross-domain joint operations static ontology, and the category of the battlefield situation instance is fuzzy matched according to the edit distance similarity, pinyin similarity and cosine similarity; an update module, the update module uses a graph method to load and parse the Turtle format file, and converts the Turtle format file into a graph object; the update module updates the cross-domain joint operations static ontology by adding, deleting and modifying the graph object; A dynamic analysis module, which is used to dynamically analyze the updated content of the cross-domain joint operations static ontology and form knowledge accumulation, and use the knowledge accumulation as the input variable for the next multi-source data fusion; wherein the dynamic analysis module is also used to design prompt words for the large language model, use the large language model to predict the enemy's strategy and intention, and give our troop deployment and firepower allocation strategy; a reasoning verification module, which is used to perform reasoning verification on the cross-domain joint operations static ontology and the updated dynamic ontology, including equipment effectiveness analysis, combat readiness deployment, cost-benefit analysis and combat outcome prediction.

[0052] According to the real-time situation update and analysis device based on dynamic ontology in the embodiment of the present invention, by establishing a general workflow for ontology construction and update, it is possible to realize the automatic standardized construction of data to ontology model, which is conducive to improving the degree of automation; the use of intelligent data integration and standardization technology can improve the efficiency of data extraction, matching and fusion; the use of multi-dimensional fuzzy weighted matching algorithm to improve the accuracy of matching between the static ontology of cross-domain joint operations and target data; the use of a large language model to analyze and interpret the dynamically updated part of the static ontology of cross-domain joint operations, and then give predictions and suggestions to guide the decision-making process.

[0053] According to some embodiments of the present invention, the building block comprises:

[0054] A demand analysis module, wherein the demand analysis module is used to perform demand analysis;

[0055] A static data collection module, wherein the static data collection module is used to collect literature, expert knowledge and historical data;

[0056] A static data fusion module, which is used to extract and fuse effective data;

[0057] A conceptual modeling module, wherein the conceptual modeling module is used to carry out hierarchical design and relationship construction;

[0058] An ontology construction module, wherein the ontology construction module is used to generate the cross-domain joint combat static ontology.

[0059] According to an embodiment of the third aspect of the present invention, an electronic device includes a processor and a memory, wherein the processor is connected to the memory, and the memory is used to store a computer program. When the computer program is executed by the processor, the real-time situation update and analysis method based on dynamic ontology as described in the embodiment of the first aspect of the present invention is implemented.

[0060] The electronic device according to the embodiment of the present invention can realize the automatic standardized construction of data to ontology model by establishing a general workflow for ontology construction and update, which is conducive to improving the degree of automation; the intelligent data integration and standardization technology is adopted to improve the efficiency of data extraction, matching and fusion; the multi-dimensional fuzzy weighted matching algorithm is used to improve the accuracy of matching between the static ontology of cross-domain joint operations and the target data; the large language model is used to analyze and interpret the dynamically updated part of the static ontology of cross-domain joint operations, and then predictions and suggestions are given to guide the decision-making process.

[0061] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 is a schematic diagram of a general workflow for ontology construction and update according to an embodiment of the present invention;

[0063] Figure 2 is an architecture diagram of a real-time situation update and analysis method based on dynamic ontology according to an embodiment of the present invention;

[0064] Figure 3 is a schematic diagram of multi-source data fusion according to an embodiment of the present invention;

[0065] Figure 4 is a structural block diagram of comprehensive similarity according to an embodiment of the present invention;

[0066] Figure 5 is a schematic diagram of upper and lower supports of a cross-domain joint operations static ontology according to an embodiment of the present invention;

[0067] Figure 6 It is a mapping relationship diagram between real-time battlefield situation data and cross-domain joint combat static ontology according to an embodiment of the present invention. DETAILED DESCRIPTION

[0068] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0069] The following describes a real-time situation update and analysis method based on a dynamic ontology according to an embodiment of the first aspect of the present invention with reference to the accompanying drawings. The method is used for multi-source battlefield situation awareness and analysis decision-making in cross-domain joint combat scenarios, which include land, sea, air, network and electromagnetic fields.

[0070] like Figure 1-Figure 6 As shown, the real-time situation update and analysis method based on dynamic ontology according to an embodiment of the present invention includes:

[0071] Collect literature, expert knowledge and historical data based on needs analysis;

[0072] Extract and integrate effective data;

[0073] Carry out hierarchical design and relationship construction to form conceptual modeling;

[0074] Perform semantic modeling based on entities, relationships, and events in combat scenarios to generate a cross-domain joint combat static ontology, and convert the cross-domain joint combat static ontology into a Turtle format file;

[0075] Obtain multiple real-time battlefield situation data from different combat domain sources;

[0076] Directedly capture target data in real-time battlefield situation data according to capture constraints;

[0077] Perform multi-source data fusion on multiple target data, generate multi-source complementary fusion data through data redundancy processing, data association analysis and data merging, including equipment parameter redundancy processing, weapon mount association relationship establishment and similar resource merging and division;

[0078] Perform data standardization on multi-source complementary fusion data to generate structured situation data. The format and structure of structured situation data are consistent and are equipped with identifiers for matching battlefield situation instances. Ensure data consistency and integrity, break down data silos, realize cross-system and cross-platform data exchange and sharing, and improve data integration efficiency and effectiveness.

[0079] Extract identifiers of all battlefield situation instances. If the battlefield situation instance does not exist in the cross-domain joint operations static ontology, add the battlefield situation instance and its attributes to the cross-domain joint operations static ontology, and perform fuzzy matching on the categories of the battlefield situation instance based on edit distance similarity, phonetic similarity, and cosine similarity.

[0080] Use graph method to load and parse Turtle format files, and convert Turtle format files into graph objects;

[0081] Update the cross-domain joint operations static ontology by adding, deleting and modifying graph objects;

[0082] Dynamically analyze the updated content of the static ontology of cross-domain joint operations and form knowledge accumulation, and use the knowledge accumulation as the input variable for the next multi-source data fusion to meet the needs of dynamic data changes and adaptive growth, and further improve the integration and utilization capabilities of battlefield data. Among them, dynamic analysis includes:

[0083] Design the prompt words of the large language model, use the large language model to predict the enemy's strategy and intention, and give our troop deployment and firepower allocation strategy. For example, the prompt words are "You are a senior military tactical analyst with rich battlefield situation awareness and analysis experience. Please carefully analyze the updated content of the cross-domain joint combat static ontology, evaluate the number and location of military units / equipment of both sides, intelligence data, communication status and combat objectives, and based on these analyses, speculate on the tactical actions that the enemy may take against us, and how we should formulate a response strategy."

[0084] The analysis includes:

[0085] The enemy and our own combat objectives, changes in troop strength, changes in the number and location of weapons and equipment, changes in intelligence collection, communication capabilities, and other important situation information; the analysis results include:

[0086] Tactics that the enemy is preparing to take: Based on the enemy's dynamically updated situation information, the specific tactical actions that the enemy is preparing to take include raids, reconnaissance, and long-range strikes;

[0087] Our response strategy: Our response measures include strengthening air defense capabilities, strengthening reconnaissance and surveillance, and taking actions to confuse the enemy;

[0088] Situational data used for dynamic analysis include: changes in the physical status of both the enemy and our side, including alert status, standby status, intelligence collection activities, equipment status, and combat level.

[0089] Below is an example of the output of a large language model.

[0090] Possible tactics adopted by the enemy: Based on the enemy's dynamic updates, they may prepare for rapid strikes, surprise attacks, and long-range strikes. For example, they may use tactical assault vehicles to conduct surprise attacks while using missile launchers to conduct long-range strikes. The enemy's increased radar activity may be to better monitor our actions and guide their strike operations.

[0091] Our response strategy: We should strengthen our air defense capabilities, deploy air defense missile systems, enhance our electronic warfare capabilities, strengthen monitoring, and take actions to confuse the enemy. Specifically, we should immediately strengthen our air defense capabilities on the front line, deploy more air defense missile systems and close-range defense weapons; at the same time, enhance our electronic warfare capabilities to interfere with and destroy the enemy's radar and communication systems; in addition, we should strengthen monitoring of enemy actions and use drones and other reconnaissance means to obtain more accurate intelligence so that we can respond in a timely manner. At the tactical level, we can consider taking actions to confuse the enemy, such as distracting the enemy through simulated attacks, or setting traps and ambushes on the enemy's possible attack routes.

[0092] In some specific embodiments of the present invention, the enemy: The distance between the enemy's tactical assault vehicles and our base has been shortened to 5 kilometers, the number of missile launchers has increased significantly, and the frequency and duration of the enemy's radar startup have increased, indicating that the enemy may be preparing for some form of military action. Our side: Our intelligence collection capabilities have improved, but the intensity of electronic interference on the front line is relatively high, which may affect the accuracy of intelligence and the stability of communications.

[0093] Finally, the cross-domain joint combat static ontology and the updated dynamic ontology are reasoned and verified, including equipment effectiveness analysis, combat readiness deployment, cost-benefit analysis, and combat outcome prediction.

[0094] According to the real-time situation update and analysis method based on dynamic ontology in the embodiment of the present invention, in view of the characteristics of situation data in complex scenarios, such as multi-source heterogeneity, complex and intertwined relationship attributes, discontinuity in time and space, and incomplete information, a general workflow for ontology construction and update is established, and a combat system is dynamically constructed with fragmented data as a supplement, which reduces the participation of experts, can realize the automatic standardized construction of data to ontology model, further improves the extensibility and portability of ontology, and is conducive to improving the degree of automation. The use of intelligent data integration and standardization technology can improve the efficiency of data extraction, matching and fusion. The multi-dimensional fuzzy weighted matching algorithm is used to improve the accuracy of matching between static ontology and target data in cross-domain joint operations. The large language model is used to analyze and interpret the dynamic update part of the static ontology of cross-domain joint operations, realizes the cognition and understanding of the dynamic process of the continuous evolution of entities over time in the combat system, and gives the prediction of enemy tactics and the choice of our strategy.

[0095] Specifically, establishing Figure 1 The general workflow of ontology construction and update is shown in the figure. The left side of the figure shows the static ontology construction process, and the right side shows the dynamic ontology update process. The white boxes (such as demand analysis, literature, expert knowledge, historical data, hierarchical design, relationship construction, concept modeling, ontology construction, and cross-domain joint operations static ontology) are the unique links of static ontology construction, the grid boxes (such as real-time situation, ontology update, dynamic ontology, dynamic analysis, knowledge accumulation) are the unique links of dynamic ontology update, and the gray long boxes (such as data collection, data fusion, and reasoning verification) are the common links of static ontology construction and dynamic ontology update.

[0096] The construction of the static ontology of cross-domain joint operations begins with demand analysis. By collecting literature, expert knowledge, and historical data, effective data is extracted and integrated. Then, two key steps, hierarchical design and relationship construction, are carried out to achieve conceptual modeling. Finally, with the help of ontology knowledge modeling tools, the ontology file model is generated, laying the foundation for subsequent reasoning verification.

[0097] On the basis of the static ontology of cross-domain joint operations, real-time battlefield situation data is collected, and fusion is carried out by using targeted capture of large-scale target data, multi-source data fusion, and data standardization processing. Relationship mapping and updating are carried out based on the multi-dimensional fuzzy weighted matching algorithm. With the help of a large language model, the updated content of the static ontology of cross-domain joint operations is analyzed to form a knowledge accumulation that includes the enemy's combat intentions, our own tactics and strategies, and resource allocation plans.

[0098] like Figure 2As shown in the figure, based on the static ontology of cross-domain joint operations, combined with the fragmented target data obtained from the real-time battlefield situation data, the dynamic update and improvement of the static ontology of cross-domain joint operations is realized to reflect the latest intelligence and battlefield situation. Specifically, semi-structured data is converted into structured data; support the addition, deletion and creation of subclass-parent class relationships of ontology classes; add, delete and build subordinate relationships of instance data attributes and object attributes; focus on supporting the addition of instance and attribute annotations, instance core attribute completion, construction of complex object relationships between instances, and adaptive matching of instance categories; intelligent analysis and decision-making include process logic, spatiotemporal relationships and situation analysis. For example, the input is the complete situation information in hierarchical JSON format and the static ontology of cross-domain joint operations in TTL format, and the output is the updated dynamic ontology in TTL format.

[0099] Combat scenarios involve different military services and departments, with many participating units and data from multiple sources. Different participating units may use different systems, equipment, and communication protocols, which leads to heterogeneous data formats and requires data fusion and conversion. Therefore, in view of the multi-source, heterogeneous, and fragmented data in combat scenarios, multi-source heterogeneous data fusion technology is used to support the dynamic update of subsequent entities and support the decision-making needs of command.

[0100] Data fusion usually includes steps such as data quality management, matching, alignment and integration. In the dynamic update stage of the static ontology of cross-domain joint operations, the received situation data is large in scale, hierarchically nested in structure, and has complex relationship attributes. It is necessary to carry out targeted capture of target data under large-scale situations, complementary fusion of multi-source data, and standardized processing of heterogeneous data to achieve multi-source fragmented data fusion, laying the foundation for the dynamic update of the static ontology of cross-domain joint operations.

[0101] In large-scale situation data, obtaining designated target information is the basis for dynamic ontology updating. The accuracy and comprehensiveness of the extracted data affect the completeness of the updated battlefield ontology knowledge model. According to some embodiments of the present invention, directional crawling includes:

[0102] Obtain a first readable file containing grabbing constraints and a second readable file containing real-time battlefield situation data, where the first readable file and the second readable file may be json files;

[0103] Reading a first readable file and converting the grabbing constraints into a first dictionary, reading a second readable file and converting the real-time battlefield situation data into a second dictionary;

[0104] For each set of constraint information in the first dictionary, extract the location information and attribute information of the target data, and construct a capture field using the location information and the attribute information;

[0105] For each captured field, capture the target data in the second dictionary;

[0106] The captured target data is converted into a third readable file and outputted. The third readable file may be a json file.

[0107] First, in order to better represent and process complex data structures, the constraint conditions and real-time battlefield situation data are converted into nested dictionary forms by using data encoding and decoding methods to achieve effective data organization and management. Secondly, the position key and constraint conditions in the data are extracted through information extraction and semantic understanding technology. The position key specifies the storage location of the target data, while the constraint conditions describe the screening requirements for the target data, such as the attribute information of the target data.

[0108] According to the location key and constraints obtained by parsing, the crawl field can be constructed. The crawl field defines the location and constraints of the target data to be crawled, which is the key parameter for subsequent data extraction. The construction of the crawl field can be assisted by technical means such as data query language or regular expressions to accurately locate and filter the target data.

[0109] Finally, through data extraction, the acquired target data is converted into JSON format to realize target data extraction in large-scale situation data scenarios.

[0110] In combat scenarios, there are many combat units and weapons and equipment, and the data sources are wide and may be imperfect. Multi-source data complementary fusion technology integrates and fuses multiple data files from different sources to improve the quality, accuracy and availability of data. In this way, the complementarity between different data sources can be fully utilized to obtain more comprehensive and global information.

[0111] In some embodiments of the present invention, multi-source data fusion includes:

[0112] Obtaining a plurality of third readable files, and converting the third readable files into a third dictionary;

[0113] For each key of any two third dictionaries, if the keys of the two third dictionaries exist and have the same key value, the key value is updated using the result of fusion using the complementary fusion function; otherwise, the key values ​​of the two third dictionaries are updated with the key value of the existing key. Figure 3 In the process, data redundancy processing, data association analysis, and data merging are performed on n data sources to generate multi-source complementary fusion data. For example, multiple subsystems and multiple engines output battlefield situation files, that is, multiple subsystems and multiple engines become data source nodes.

[0114] Data fused from multiple sources usually have differences in format, structure and quality. Before use, data cleaning, data conversion, deduplication, anomaly detection and error correction are usually required to ensure that the fused data from different sources have a consistent format and structure and solve data quality issues.

[0115] Real-time battlefield situation data is usually stored in a nested JSON format, which contains a complex hierarchical structure. The first step in standardizing heterogeneous data is data flattening, also known as data flattening. Data flattening is an important step in data preprocessing. It can be flattened into a table-like structure using a tree traversal algorithm or a recursive method and filled into the constructed data framework.

[0116] The construction rules of the cross-domain joint operations static ontology stipulate that the instance name is unique and does not contain illegal characters, such as "#", " / ", etc. In the situation information, instance names are allowed to be repeated, and there are no restrictions on naming. Therefore, before using situation data to update the cross-domain joint operations static ontology, data cleaning must be performed to number instances with the same name and remove illegal symbols.

[0117] After a series of processing, information consistent with the static ontology construction rules of cross-domain joint operations can be obtained, providing high-quality situation information for dynamic updates.

[0118] The cross-domain joint combat static ontology is established after highly conceptualized and abstracted combat experience, intelligence, and initial information. Therefore, the judgment of instances and the setting of class names will be affected by human subjectivity to a certain extent, making them different from the instances in the situation data. In order to solve the problem that the situation data information is difficult to match with the cross-domain joint combat static ontology, the ontology update technology based on fuzzy matching is the key to whether the situation information can be updated to the cross-domain joint combat static ontology. The information in the situation data to be updated to the cross-domain joint combat static ontology is mostly in the form of strings. Because of its wide naming range and diverse naming methods, the names contain special characters such as Chinese characters, letters, and symbols, which brings many challenges to accurate matching. Since the edit distance similarity between air-to-air missiles and ground-to-ground missiles is the same as the edit distance similarity between air-to-air missiles and air-to-air missiles, and the pinyin similarity cannot cope with the string matching problem of homophonic characters, the cosine similarity is used to make up for the limitations of the first two matching methods. Cosine similarity is based on text content and does not consider the order of characters. It only focuses on the frequency of word occurrence and the angle between vectors, and is not sensitive to the order of text arrangement. Therefore, Figure 4 As shown, fuzzy matching is performed according to the following formula:

[0119] s=A·s 1 +B·s 2 +C·s 3

[0120] In the formula, s is the comprehensive similarity, s∈[0,1]; the larger s is, the higher the similarity between the two. When s=1, it means that the two are completely consistent. 1 is the edit distance similarity; s 2 is the pinyin similarity; s 3 is the cosine similarity; A, B, C are weighted coefficients, A+B+C=1. If the lengths of the strings are different, A is greater than B; if the lengths of the strings are the same, A is less than B. In this way, the accuracy of the comparison can be improved, the defects can be compensated for, the similarity between the two strings can be measured relatively comprehensively, and finally the fuzzy matching of the static ontology of cross-domain joint operations can be realized.

[0121] The cross-domain joint operations static ontology is built based on professional knowledge and industry rules, combined with initial data and typical scenarios. By defining ontology classes and attributes, the hierarchy, relationship and constraints between knowledge can be established. The cross-domain joint operations static ontology is mainly used to represent and store domain knowledge. It cannot automatically adapt to new knowledge and situation changes, so a dynamic ontology is needed to realize knowledge update. The dynamic update is mainly the relationship between the cross-domain joint operations static ontology and real-time battlefield situation data.

[0122] In the simulation and deduction system, the relationship between the cross-domain joint operations static ontology and the real-time battlefield situation data is bidirectional. On the one hand, the situation data is the source of knowledge for constructing the cross-domain joint operations static ontology; on the other hand, the relationship in the cross-domain joint operations static ontology also provides support for the association and fusion of situation data, making autonomous data association possible. For example, the synonymous relationship between the DBID (DataBase Identifier, the unique identifier of the database) in the real-time battlefield situation data and the class to which the instance belongs can guide the mapping of the corresponding fields. The cognitive ability of the cross-domain joint operations static ontology provides certain support for decision-making, and the execution results during system operation will also provide feedback to the cross-domain joint operations static ontology to further optimize and improve the quality of knowledge. Therefore, if Figure 5 As shown, a good cross-domain joint combat static ontology can support system applications upward and integrate situational data downward, playing an important role in the construction of combat system models.

[0123] The difficulty of dynamic adaptive updating of the static ontology of cross-domain joint operations lies in the mapping relationship between real-time battlefield situation data and ontology information. The present invention adopts a full-scale update method to update the static ontology of cross-domain joint operations based on patterns and rules. The artificially defined rules are matched with the standardized real-time battlefield situation data to identify the subject, object or predicate of the instance data triple. The class in the static ontology of cross-domain joint operations is an abstraction and classification of entities, which usually does not change when the situation changes. It is mainly used to determine the category to which the instance data belongs in the dynamic update; the instance in the static ontology of cross-domain joint operations is the focus of dynamic update, including the addition and value update of instances and their attributes. For the update of data attributes and relationship attributes, a set of attributes is pre-defined based on the static ontology of cross-domain joint operations, and fuzzy matching and update are performed with the attribute key values ​​in the real-time battlefield situation data. If there are attributes that are not defined in the static ontology of cross-domain joint operations in the real-time battlefield situation data, for the completeness of the model, they are first added to the annotation module, and then reminders and warnings are made, and their attribute categories are determined after semantic understanding and judgment.

[0124] The dynamic ontology is implemented on the basis of the cross-domain joint operations static ontology. The cross-domain joint operations static ontology is built based on the Resource Description Framework (RDF) model. Any entity that may exist in the battlefield can be represented as a resource in the RDF model using the W3C standard. The semi-structured situation data after fragmented fusion and heterogeneous standardization is in a two-layer nested form. The data primary key is the unique identifier of the instance, and the internal value contains the instance name and attributes. The cross-domain joint operations static ontology is mapped to the semi-structured situation data according to the knowledge understanding, and the attribute alignment and relationship alignment are achieved according to some key identifiers. After determining the mapping object of the situation data, it can be accurately updated to the cross-domain joint operations static ontology to realize the dynamic update of the cross-domain joint operations static ontology, that is, the dynamic ontology. Construct a rule-based mapping relationship between real-time battlefield situation data and the cross-domain joint operations static ontology, such as Figure 6 shown.

[0125] The dynamic nature of the cross-domain joint operations static ontology update lies in the fact that the model can be dynamically adjusted and updated according to changes in external situation data to keep it consistent and accurate with the actual situation. When new concepts, relationships, or attributes appear, the cross-domain joint operations static ontology can be expanded and updated by adding corresponding triples to reflect new situation changes. The adaptability of the cross-domain joint operations static ontology update lies in the use of fuzzy matching, which can achieve adaptive and accurate updates of different types of instances and their attributes. The two complement each other. The updated cross-domain joint operations static ontology can better reflect the battlefield situation, further enhance the ability to integrate and utilize battlefield data, and support the commander's tactical decision-making.

[0126] During the combat process, data sources are fragmented, dynamic, low-quality, and disordered, which poses a severe challenge to data analysis and knowledge discovery for complex combat processes. To this end, the present invention proposes a real-time situation update and analysis method based on dynamic ontology, using the cross-domain joint combat static ontology as the data model layer to form a normative constraint for data fusion; it allows the cross-domain joint combat static ontology to be dynamically adjusted based on a variety of data matching methods when new data sources are imported, realizing the dynamic linking of the cross-domain joint combat static ontology, seamless fusion of data, and reflecting the spatiotemporal relationship between combat formations, combat objects, and combat elements.

[0127] According to the second aspect of the present invention, the real-time situation update and analysis device based on dynamic ontology is used for multi-source battlefield situation awareness and analysis decision-making in cross-domain joint combat scenarios, and the cross-domain joint combat scenarios include land, sea, air, network and electromagnetic fields. The device includes: a construction module, an acquisition module, a directional capture module, a multi-source data fusion module, a data standardization processing module, an extraction module, an update module, a dynamic analysis module and an inference verification module. For example, the real-time situation update and analysis device based on dynamic ontology can be an open source large language model application development platform, and the general workflow for ontology construction and update is constructed based on the open source large language model application development platform, which can be used. Figure 1 The complex tasks in the ontology are decomposed into multiple subtasks, reducing the complexity of each link, thereby improving the efficiency of ontology updating.

[0128] Specifically, the construction module is used to construct a cross-domain joint combat static ontology and convert the cross-domain joint combat static ontology into a Turtle format file. The acquisition module is used to obtain multiple real-time battlefield situation data from different combat fields. The directional capture module directional captures the target data in the real-time battlefield situation data according to the capture constraints. The multi-source data fusion module is used to perform multi-source data fusion on multiple target data, and generate multi-source complementary fusion data through data redundancy processing, data association analysis and data merging, including equipment parameter redundancy processing, weapon mounting association relationship establishment and similar resource merging and division. The data standardization processing module is used to perform data standardization processing on multi-source complementary fusion data to generate structured situation data. The format and structure of the structured situation data are consistent and are provided with identifiers for matching battlefield situation instances. The extraction module is used to extract the identifiers of all battlefield situation instances. If the battlefield situation instance does not exist in the cross-domain joint combat static ontology, the battlefield situation instance and its attributes are added to the cross-domain joint combat static ontology, and the categories of the battlefield situation instances are fuzzy matched according to the edit distance similarity, pinyin similarity and cosine similarity. The update module uses a graph method to load and parse the Turtle format file, converting the Turtle format file into a graph object; the update module updates the cross-domain joint operations static ontology by adding, deleting and modifying the graph object. The dynamic analysis module is used to dynamically analyze the updated content of the cross-domain joint operations static ontology and form knowledge accumulation, and use the knowledge accumulation as the input variable for the next multi-source data fusion. Among them, the dynamic analysis module can be constructed based on the prompt engineering technology of the large language model, used to design the prompt words of the large language model, use the large language model to predict the enemy's strategy and intention, and give our troop deployment and firepower allocation strategy. The reasoning verification module is used to perform reasoning verification on the cross-domain joint operations static ontology and the updated dynamic ontology, including equipment effectiveness analysis, combat readiness deployment, cost-benefit analysis and combat result prediction.

[0129] According to the real-time situation update and analysis device based on dynamic ontology in the embodiment of the present invention, by establishing a general workflow for ontology construction and update, it is possible to realize the automatic standardized construction of data to ontology model, which is conducive to improving the degree of automation; the use of intelligent data integration and standardization technology can improve the efficiency of data extraction, matching and fusion; the use of multi-dimensional fuzzy weighted matching algorithm to improve the accuracy of matching between the static ontology of cross-domain joint operations and target data; the use of a large language model to analyze and interpret the dynamically updated part of the static ontology of cross-domain joint operations, and then give predictions and suggestions to guide the decision-making process.

[0130] According to some embodiments of the present invention, the construction module includes: a demand analysis module, a static data collection module, a static data fusion module, a concept modeling module, and an ontology construction module. The demand analysis module is used to perform demand analysis. The static data collection module is used to collect literature, expert knowledge, and historical data. The static data fusion module is used to extract and fuse effective data. The concept modeling module is used to carry out hierarchical design and relationship construction. The ontology construction module is used to generate a static ontology for cross-domain joint operations.

[0131] The electronic device according to the third aspect of the present invention comprises a processor and a memory. The processor and the memory are connected, and the memory is used to store a computer program. When the computer program is executed by the processor, the real-time situation update and analysis method based on dynamic ontology as described in the first aspect of the present invention is implemented.

[0132] The electronic device according to the embodiment of the present invention can realize the automatic standardized construction of data to ontology model by establishing a general workflow for ontology construction and update, which is conducive to improving the degree of automation; the intelligent data integration and standardization technology is adopted to improve the efficiency of data extraction, matching and fusion; the multi-dimensional fuzzy weighted matching algorithm is used to improve the accuracy of matching between the static ontology of cross-domain joint operations and the target data; the large language model is used to analyze and interpret the dynamically updated part of the static ontology of cross-domain joint operations, and then predictions and suggestions are given to guide the decision-making process.

[0133] Other structures and operations of the electronic device according to the embodiment of the present invention are known to those skilled in the art and will not be described in detail here.

[0134] In the description of the present invention, it is to be understood that the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0135] In the description of the present invention, "a first feature" or "a second feature" may include one or more of the features. A first feature "above" or "below" a second feature may include direct contact between the first and second features, or may include contact between the first and second features through another feature between them instead of direct contact. A first feature "above", "above" or "above" a second feature may include the first feature being directly above or obliquely above the second feature, or may simply mean that the first feature is higher in level than the second feature.

[0136] It should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0137] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "specific embodiments", "examples" or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0138] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A real-time situation update and analysis method based on dynamic ontology, used for multi-source battlefield situation awareness and analysis decision-making in cross-domain joint combat scenarios, wherein the cross-domain joint combat scenarios include land, sea, air, network and electromagnetic fields, characterized in that: The method comprises: Constructing a cross-domain joint combat static ontology, and converting the cross-domain joint combat static ontology into a Turtle format file; Obtain multiple real-time battlefield situation data from different combat domain sources; Directedly grabbing target data in the real-time battlefield situation data according to the grabbing constraint conditions; Perform multi-source data fusion on the plurality of target data, and generate multi-source complementary fusion data through data redundancy processing, data association analysis and data merging, including equipment parameter redundancy processing, weapon mount association relationship establishment and similar resource merging and division; Performing data standardization processing on the multi-source complementary fusion data to generate structured situation data, wherein the structured situation data has a consistent format and structure and is provided with an identifier for matching a battlefield situation instance; Extract identifiers of all battlefield situation instances; if the battlefield situation instance does not exist in the cross-domain joint operations static ontology, add the battlefield situation instance and its attributes to the cross-domain joint operations static ontology, and perform fuzzy matching on the category of the battlefield situation instance according to edit distance similarity, pinyin similarity and cosine similarity; Using a graph method to load and parse the Turtle format file, and convert the Turtle format file into a graph object; Updating the cross-domain joint operations static ontology by adding, deleting and modifying the graph objects; Dynamically analyzing the updated content of the cross-domain joint operations static ontology and forming knowledge accumulation, and using the knowledge accumulation as an input variable for the next multi-source data fusion; The cross-domain joint combat static ontology and the updated dynamic ontology are subjected to reasoning verification, including equipment effectiveness analysis, combat readiness deployment, cost-benefit analysis and combat outcome prediction; wherein the dynamic analysis includes: Design prompt words for the large language model, use the large language model to predict enemy strategies and intentions, and provide our own troop deployment and firepower allocation strategies.

2. The real-time situation update and analysis method based on dynamic ontology according to claim 1 is characterized in that: The construction of the cross-domain joint combat static ontology includes: Collect literature, expert knowledge and historical data based on needs analysis; Extract and integrate effective data; Carry out hierarchical design and relationship construction to form conceptual modeling; Generate the cross-domain joint combat static ontology.

3. The real-time situation update and analysis method based on dynamic ontology according to claim 1 is characterized in that: The directional crawling includes: Acquire a first readable file containing the grabbing constraint conditions and a second readable file containing the real-time battlefield situation data; Reading the first readable file and converting the grabbing constraint condition into a first dictionary, reading the second readable file and converting the real-time battlefield situation data into a second dictionary; For each set of constraint information in the first dictionary, extract the location information and attribute information of the target data, and construct a capture field using the location information and the attribute information; For each of the captured fields, capture the target data in the second dictionary; The captured target data is converted into a third readable file and outputted.

4. The real-time situation update and analysis method based on dynamic ontology according to claim 3 is characterized in that: The multi-source data fusion includes: Acquire a plurality of the third readable files, and convert the third readable files into a third dictionary; For each key of any two of the third dictionaries, if the keys of the two third dictionaries exist and have the same key value, the key value is updated using the result of fusion using the complementary fusion function; otherwise, the key values ​​of the two third dictionaries are updated with the key value of the existing key.

5. The real-time situation update and analysis method based on dynamic ontology according to claim 1 is characterized in that: The data standardization process includes: The multi-source complementary fusion data is flattened, cleaned, converted, deduplicated, detected as anomaly, and corrected.

6. The real-time situation update and analysis method based on dynamic ontology according to claim 1 is characterized in that: The fuzzy matching is performed according to the following formula: s=A·s1+B·s2+C·s3 Where s is the comprehensive similarity, s∈[0,1]; s1 is the edit distance similarity; s2 is the pinyin similarity; s3 is the cosine similarity; A, B, C are weighted coefficients, A+B+C=1; If the lengths of the strings are different, then A is greater than B; if the lengths of the strings are the same, then A is less than B.

7. The real-time situation update and analysis method based on dynamic ontology according to claim 1 is characterized in that: The prompt of the large language model is "You are a senior military tactical analyst with rich battlefield situation awareness and analysis experience. Please carefully analyze the updated content of the cross-domain joint operations static ontology, evaluate the number and location of military units / equipment of both sides, intelligence data, communication status and combat objectives, and based on these analyses, speculate on the tactical actions that the enemy may take against us, and how we should formulate response strategies"; The analysis includes: The enemy and our own combat objectives, changes in troop strength, changes in the number and location of weapons and equipment, changes in intelligence collection, communication capabilities, and other important situation information; the analysis results include: Tactics that the enemy is preparing to take: Based on the enemy's dynamically updated situation information, the specific tactical actions that the enemy is preparing to take include raids, reconnaissance, and long-range strikes; Our response strategy: Our response measures include strengthening air defense capabilities, strengthening reconnaissance and surveillance, and taking actions to confuse the enemy; The situation data used for the dynamic analysis include: changes in the physical status of both the enemy and us, including alert status, standby status, intelligence collection activities, equipment status, and combat level.

8. A real-time situation update and analysis device based on dynamic ontology, used for multi-source battlefield situation awareness and analysis decision-making in cross-domain joint combat scenarios, wherein the cross-domain joint combat scenarios include land, sea, air, network and electromagnetic fields, characterized in that: The device comprises: A construction module, wherein the construction module is used to construct a cross-domain joint combat static ontology and convert the cross-domain joint combat static ontology into a Turtle format file; An acquisition module, the acquisition module is used to acquire multiple real-time battlefield situation data from different combat fields; A directional grabbing module, which directionally grabs target data in the real-time battlefield situation data according to grabbing constraints; A multi-source data fusion module, which is used to perform multi-source data fusion on the plurality of target data, and generate multi-source complementary fusion data through data redundancy processing, data association analysis and data merging, including equipment parameter redundancy processing, weapon mount association relationship establishment and similar resource merging and division; A data standardization processing module, the data standardization processing module is used to perform data standardization processing on the multi-source complementary fusion data to generate structured situation data, the structured situation data has a consistent format and structure and is provided with an identifier for matching a battlefield situation instance; An extraction module, the extraction module is used to extract identifiers of all the battlefield situation instances, if the battlefield situation instance does not exist in the cross-domain joint operations static ontology, then add the battlefield situation instance and its attributes to the cross-domain joint operations static ontology, and perform fuzzy matching on the category of the battlefield situation instance according to edit distance similarity, pinyin similarity and cosine similarity; An update module, wherein the update module uses a graph method to load and parse the Turtle format file and convert the Turtle format file into a graph object; the update module updates the cross-domain joint operations static ontology by adding, deleting and modifying the graph object; A dynamic analysis module, which is used to dynamically analyze the updated content of the cross-domain joint operations static ontology and form knowledge accumulation, and use the knowledge accumulation as an input variable for the next multi-source data fusion; wherein the dynamic analysis module is also used to design prompt words for a large language model, use the large language model to predict enemy strategies and intentions, and give our own troop deployment and firepower allocation strategies; A reasoning verification module is used to perform reasoning verification on the cross-domain joint combat static ontology and the updated dynamic ontology, including equipment effectiveness analysis, combat readiness deployment, cost-benefit analysis and combat outcome prediction.

9. The real-time situation updating and analyzing device based on dynamic ontology according to claim 8 is characterized in that: The building blocks include: A demand analysis module, wherein the demand analysis module is used to perform demand analysis; A static data collection module, wherein the static data collection module is used to collect literature, expert knowledge and historical data; A static data fusion module, which is used to extract and fuse effective data; A conceptual modeling module, wherein the conceptual modeling module is used to carry out hierarchical design and relationship construction; An ontology construction module, wherein the ontology construction module is used to generate the cross-domain joint combat static ontology.

10. An electronic device, characterized in that: The electronic device includes a processor and a memory, the processor is connected to the memory, the memory is used to store a computer program, and when the computer program is executed by the processor, the real-time situation update and analysis method based on dynamic ontology as described in any one of claims 1-7 is implemented.

Images