CN113268591A - Air target intention evidence judging method and system based on affair atlas - Google Patents

Abstract

The invention belongs to the technical field of aerial target intention discrimination, and in particular relates to an aerial target intention discrimination method and system based on an event map. Event trigger words and event elements corresponding to typical events, and analyze the air target flight activity text data to extract events, and obtain typical events and atypical events in their flight activities; The database is used to construct the aerial target flight activity event map; for the pending air target flight activity data, the event type is extracted based on the event map, and the air target intention is judged. The present invention conducts intention identification by means of the constructed affair graph, improves the overall grasp of the actual application environment situation of the aerial target, is conducive to timely management and control of the aerial target, and has a good application prospect.

Description

Air target intention evidence judging method and system based on affair atlas

Technical Field

The invention belongs to the technical field of air target intention judgment, and particularly relates to an air target intention evidence judgment method and system based on a physics map.

Background

With the continuous development of modern aircraft technology, the types of various application tasks performed are increasing. In practical application, the intention of the high-value aerial target is judged accurately in time, comprehensive control on the situation of the application environment can be ensured, and the real-time analysis and judgment on the situation of the application environment are facilitated so as to make a response in time. For both parties, the determination of the intent of an aerial target is complex, requiring a significant amount of time and labor. In the face of the flight activity text data of the redundant aerial target, the manual evidence judgment method has a relatively obvious error rate, so an intention judgment method which can assist personnel in judging the evidence and has a relatively good effect is urgently needed to be researched.

Two methods are mainly adopted in the current research on judging the target intention in the air: the method is characterized in that the method is a multi-attribute decision method and a knowledge-based reasoning method. The multi-attribute decision can comprehensively process a large amount of information but lacks reasoning capability, and the method based on knowledge reasoning has overlarge calculated amount and is difficult to guarantee completeness. For example, the air target intention recognition based on the deep neural network is an improvement on multi-attribute decision, and a rule between a characteristic state and an intention is obtained through self-training and is used for representing the corresponding relation between the characteristic state and the intention, so that the accuracy of the recognition intention is improved. For another example, an air target intention judgment model based on intuitive fuzzy generative rule Inference (IPR) and multi-attribute decision combines the advantages of two methods, namely multi-attribute decision and knowledge inference, so that the accuracy of the result is improved, but specific feasibility needs to be explored.

Disclosure of Invention

Therefore, the invention provides an air target intention evidence judging method and system based on a case atlas, flight activity events in the air target are extracted through the existing flight activity text data of the air target, the evolution rules and modes between the events are mined, the air target intention case atlas is constructed, the intention is judged by means of the constructed case atlas, the comprehensive grasp of the actual application environment situation of the air target is promoted, and the air target is controlled in time.

According to the design scheme provided by the invention, the air target intention evidence judging method based on the affair atlas comprises the following contents:

obtaining typical event types in the activity field of the aerial target according to the flight activity text data of the aerial target;

setting event trigger words and event elements corresponding to each type of typical events, analyzing text data of the flight activities of the aerial targets to extract the events, and acquiring typical events and atypical events in the flight activities;

marking label attributes on all events and combining a database to construct an air target flight activity event map;

and aiming at the flight activity data of the air target to be processed, judging and verifying the intention of the air target by extracting the event and matching the event type based on the event map.

As the air target intention judging method based on the case atlas, the character replacement and data cleaning are firstly carried out through the regular expression aiming at the air target flight activity text data, and then the typical event is obtained according to the word clustering method.

As the air target intention judging method based on the case map, further, in the typical event type obtaining process, the word segmentation, part of speech tagging and syntax analysis are carried out on the text data, the two-tuple of the main and the predicate relations and the moving and guest relations are extracted, the candidate trigger words of the flight activity event are determined, and a candidate trigger word set is formed; and filtering the trigger words in the candidate trigger word set through a preset filtering rule, and clustering words based on the semantic similarity of the words to form typical event types in the field of flight activities.

As the air target intention judging method based on the case map, further, the preset filtering rules include but are not limited to: general verbs and verb nouns in the set candidate trigger words are reserved, and other types of verbs are filtered out; and/or filtering out trigger words with the correlation degree smaller than a set threshold value according to the correlation degree of the candidate trigger words in the set and the air target flight activity field.

As the air target intention judging method based on the affair atlas, further, the acquired typical event types include: take-off, companion flight, warp stop, arrival, normal flight, deployment, hover, and landing.

As the air target intention judging method based on the affair atlas, the event type is further identified by analyzing the text data of the flight activity of the air target in the event extraction, and if the event type is a typical event type, the event extraction is carried out by naming an entity and matching a mode; if the event type is the atypical event type, obtaining the atypical event trigger words and the event elements through the relationship analysis between the syntactic components.

As the air target intention judging method based on the case map, further, aiming at the atypical event type, setting the core word as the event trigger word according to the analysis result of the relationship between the syntactic components, and combining the subject and the idiom of the object component to obtain the event element.

The air target intention judging method based on the case map further comprises the steps of uniquely numbering all events in the case map construction, excavating the relation among the air target flight activity events, marking a plurality of events in the same data through label attributes, setting sequential bearing relation among the events with the same label attributes, and constructing and storing the air target flight activity case map by using a map database.

As the air target intention evidence judging method based on the case atlas, the event extraction is firstly carried out on the air target flight activity data to be processed, and the event type is matched through the case atlas; and if the event type cannot be matched through the event map, fuzzy matching is carried out by utilizing the keywords to determine the event type membership.

Further, the invention also provides an air target intention evidence judging system based on a affair atlas, which comprises: a type acquisition module, an event extraction module, a map construction module and an intention matching module, wherein,

the type acquisition module is used for acquiring typical event types in the activity field of the aerial target according to the flight activity text data of the aerial target;

the event extraction module is used for extracting events by setting event trigger words and event elements corresponding to each type of typical events and analyzing text data of the flight activities of the aerial targets to obtain typical events and atypical events in the flight activities;

the map construction module is used for constructing an air target flight activity event map by labeling label attributes of all events and combining a map database;

and the intention matching module is used for matching the event type according to the flight activity data of the air target to be processed by event extraction and based on the event map and judging the intention of the air target.

The invention has the beneficial effects that:

according to the method, typical events in the field are found from text data of the flight activities of the air targets, unstructured text information is structurally represented, and target intention judgment is carried out by constructing an air target flight activity affair atlas and an air target intention affair atlas; a typical event in the air target flight activity field is found by adopting a field event word clustering method, so that the definition of the event type and the event argument role in the air target flight activity field is more definite; the constructed flight activity map and the intention affair map of the aerial target visually and clearly show the flight activity change of the aerial target and the incidence relation between the flight activity and the intention, are favorable for judging the intention of the aerial target, are convenient for monitoring the aerial target in real time, and have better application prospect.

Description of the drawings:

FIG. 1 is a schematic flow chart of an air target intention judging method based on a case map in the embodiment;

FIG. 2 is a flow diagram illustrating an exemplary clustering process of event types in an embodiment;

FIG. 3 is a schematic diagram of an event extraction process in an embodiment;

FIG. 4 is a diagram of an embodiment dependency parsing structure;

FIG. 5 is a schematic diagram of an object intention event graph in the examples.

The specific implementation mode is as follows:

in order to make the objects, technical solutions and advantages of the present invention clearer and more obvious, the present invention is further described in detail below with reference to the accompanying drawings and technical solutions.

The embodiment of the invention provides an air target intention evidence judging method based on a physical map, which comprises the following contents: obtaining typical event types in the activity field of the aerial target according to the flight activity text data of the aerial target; setting event trigger words and event elements corresponding to each type of typical events, analyzing text data of the flight activities of the aerial targets to extract the events, and acquiring typical events and atypical events in the flight activities; marking label attributes on all events and combining a database to construct an air target flight activity event map; and aiming at the flight activity data of the air target to be processed, judging and verifying the intention of the air target by extracting the event and matching the event type based on the event map.

Extracting flight activity events from the existing flight activity text data of the air target, mining evolution rules and modes between the events, constructing an air target intention event map, and performing intention judgment by the constructed event map. The rationality of judging the intent of an aerial target by constructing a situational map lies in: firstly, the flight activity of the aerial target has certain rules and modes, and the event graph can just describe the flight activity of the aerial target; second, the edges of the event graph represent the relationships between events, while the compliance and causal event graphs are suitable for studying the intent of flight objectives: the flight activity is regarded as a series of events with sequential relations, the series of events can be abstracted into a cause event, the intention is regarded as an effect event, and the causal relation between the events in the event map is regarded as the causal relation between the events in the event map. In application, the sequential relationship of the flight activity event of the aerial target and the causal relationship between the flight activity of the aerial target and the intention are mined, the event is visually displayed in a mode of a case map, the method can be applied to causal analysis of the flight activity and the intention of the aerial target, a new analysis tool and a new analysis approach are provided for judging the intention of the aerial target, support is provided for pre-judging the intention, and even the change of the flight activity of the aerial target can be assisted in pre-judging. The intention of the aerial target is judged and proved by using the 'affair atlas', and from a new perspective, the flight activity of the aerial target is shown in an atlas mode by using the advantages of the affair atlas, so that the intention judgment is assisted, and the accuracy of the judgment of personnel is improved.

Aiming at the situations that the definition of event types and event argument roles in the field of air target flight activities is not clear, the data volume of air target flight activity messages obtained from a public source is small, the judgment effect of target intentions is not ideal and the like due to small scale of a constructed case atlas, in the embodiment of the scheme, referring to fig. 1, typical events in the field are found by adopting a mode of clustering event words in the field of air target flight activities, and an event template is designed for each event; and aiming at the problem that the constructed physical map is small in scale, a target intention is obtained in a newly added keyword fuzzy matching mode.

As the air target intention judging method based on the affair atlas in the embodiment of the invention, further, aiming at the air target flight activity text data, firstly, character replacement and data cleaning are carried out through a regular expression, and then a typical event is obtained according to a word clustering method.

The method comprises the following steps of using a regular expression to replace characters, carrying out data cleaning on the flight activity text data of the aerial target, and using a field event word clustering method to find typical events of the flight activity field of the aerial target, wherein the specific steps comprise word segmentation, part of speech tagging, dependency syntactic analysis, verb fine classification, field relevance calculation, semantic similarity calculation and the like. Further, in the typical event type obtaining process, word segmentation, part of speech tagging and syntax analysis are carried out on text data, two-tuple of a main and predicate relation and a moving and guest relation is extracted, candidate trigger words of the flight activity event are determined, and a candidate trigger word set is formed; and filtering the trigger words in the candidate trigger word set through a preset filtering rule, and clustering words based on the semantic similarity of the words to form typical event types in the field of flight activities. Referring to fig. 2, performing word segmentation, part-of-speech tagging and dependency syntactic analysis on the aerial target flight activity text, extracting a predicate SBV and a verb VOB doublet, and determining words such as a verb "take-off" serving as an important component in a sentence as a flight activity event candidate trigger to form a candidate trigger set. And filtering the trigger by using two filtering rules of keeping general verbs, verb nouns and domain relevance in the candidate trigger for the noise data in the candidate trigger set. The preset filtering rules may include, but are not limited to: general verbs and verb nouns in the set candidate trigger words are reserved, and other types of verbs are filtered out; and/or filtering out trigger words with the correlation degree smaller than a set threshold value according to the correlation degree of the candidate trigger words in the set and the air target flight activity field. For example, the filtering rules for retaining the general verbs and the verb nouns in the candidate triggers are eight classes of dividing the verbs into a series verb VX, an auxiliary verb VZ, a form verb VF, a trend verb VQ, a complement verb VB, a general verb VG, a name verb VN and a side verb VD; the domain relevance filtering rule is the domain relevance to reflect the relevance of the candidate trigger word to the air target flight activity domain, and is calculated by using the frequency of the candidate trigger word in the domain expectation V)/(the frequency of the candidate trigger word in the general domain). For the similarity between the meanings and the usages of the trigger words, Word semantic similarity calculation based on synonym forest (expanded version), Word semantic similarity calculation based on HowNet Word semantic original concept and Word semantic similarity calculation based on Word2Vec are adopted to cluster words with high similarity to complete event trigger Word clustering. Typical event types obtained after triggering word extraction, filtering and clustering may include: take-off, companion flight, warp stop, arrival, normal flight, deployment, hover, and landing.

As the air target intention judging method based on the affair atlas in the embodiment of the invention, further, in the event extraction, the event type is identified by analyzing the text data of the flight activity of the air target, and if the event type is a typical event type, the event extraction is carried out by naming an entity and matching a mode; if the event type is the atypical event type, obtaining the atypical event trigger words and the event elements through the relationship analysis between the syntactic components. Further, aiming at the atypical event type, according to the result of the relationship analysis between the syntactic components, the core word is set as an event trigger word, and the event elements are obtained by combining the subject and the idiom of the object component.

And analyzing the text data of the flight activity of the aerial target, and identifying the event type. Event extraction is carried out on the events which are identified as typical events by adopting a method of named entity identification combined with pattern matching, and event extraction is carried out on the events which are identified as atypical events by adopting a mode based on dependency syntactic analysis. Typical event extraction mainly comprises extraction of flight time, place and entity (model, call sign, number, quantity and the like) of the air target. Referring to fig. 3, for the identification of time and place, using a named entity identification and pattern matching method, an accurate named entity identification task may identify time, perform pattern matching in combination with a common form of time expression, take entities identified as place and organization as places, and extract places in combination with designed patterns such as "take xx as a landing place", "fly over xx", and the like. For the extraction of entities, mainly the model number, call sign, number, quantity, etc., a pattern matching method can be used, the characteristics of various attributes are considered, and letter and number character strings meeting conditions are extracted, for example: the serial number is generally 6 or 7 digits, and a horizontal bar is arranged in the middle of the model. For atypical events, since a fixed pattern is not designed for an event in a text, and it cannot be determined which elements the event should have, a method based on dependency syntax analysis is adopted to obtain the main elements describing the event in the text by analyzing the relationship between syntax components. Extracting trigger words and event elements of atypical events based on dependency syntactic analysis, firstly setting core words as event trigger words according to dependency syntactic analysis results, wherein the common core words are predicates of sentences; and extracting the subject and the object in the dependency syntax analysis result, and combining the fixed languages of the two components as the elements of the event.

As the air target intention judging method based on the case map in the embodiment of the invention, further, in the case map construction, all events are uniquely numbered, the relation among the air target flight activity events is mined, a plurality of events in the same data are marked through the tag attributes, the sequential relation among the events with the same tag attributes is set, and the air target flight activity case map is constructed and stored by using a map database. Further, aiming at the flight activity data of the air target to be processed, firstly extracting an event, and matching the event type through a physical map; and if the event type cannot be matched through the event map, fuzzy matching is carried out by utilizing the keywords to determine the event type membership.

And (4) performing event extraction according to the steps by using the flight activity data of the air target, and extracting to obtain typical events and atypical events of the flight activity of the air target. Referring to fig. 5, all events are uniquely numbered id, so that each event can be represented by the number, the relationship between the flight activity events of the air target is mined in a mode of setting specific attributes, a plurality of events in one piece of data are labeled by using 'tag' attributes, sequential relationship exists among the events with the same 'tag' attributes, and an air target flight activity event map is constructed and stored by using a map database. Determining key elements of various air target flight activity events, and constructing an air target intention event graph by performing operations such as node combination, intention association, confidence degree assignment and the like on the air target flight activity event graph. And judging and verifying the aerial target intention based on an aerial target intention affair atlas and a keyword fuzzy matching method. Firstly, event extraction is carried out, the matching identification is carried out on the factors such as the event type, the model number and the place in the extracted structured event, Cypher query sentences can be written, and the generated Cypher sentences are used for matching intentions in the idea affair map. And when the intention is not matched by using the intention event atlas, carrying out intention matching by using a keyword fuzzy matching mode instead, and determining the membership relation between the keywords, the types and the places of the event elements and the intention.

Further, based on the above method, an embodiment of the present invention further provides an air target intention evidence judgment system based on a case atlas, including: a type acquisition module, an event extraction module, a map construction module and an intention matching module, wherein,

the type acquisition module is used for acquiring typical event types in the activity field of the aerial target according to the flight activity text data of the aerial target;

the event extraction module is used for extracting events by setting event trigger words and event elements corresponding to each type of typical events and analyzing text data of the flight activities of the aerial targets to obtain typical events and atypical events in the flight activities;

the map construction module is used for constructing an air target flight activity event map by labeling label attributes of all events and combining a map database;

and the intention matching module is used for matching the event type according to the flight activity data of the air target to be processed by event extraction and based on the event map and judging the intention of the air target.

To verify the validity of the scheme, the following further explanation is made by combining specific experimental data:

the air target flight activity message data is stored in a txt file, shaped as follows:

a certain air force 1 KC-10A fuel dispenser (MOJO91) with the number of 83-0081 takes off from A11 area base in 2 months, 29 am and flies above B area.

An air force 1 at 1 month and 30 nights has an RC-135U scout (GIVE31) with the number of 64-14847 lifted from the C11 base of the C region and flies above the D region.

The MC-130J special aircraft (RCH1031) numbered 08-6206 on the 27 th special army of a certain air force on 1 month and 27 days takes off from the base E, passes through the parking areas E1, E2 and E3 and finally lands on the base F.

An EP-3E scout (MN806) with rack number 156517 from a certain navy 1 at afternoon of 30 months at 1 takes off from region G and flies above region H.

An air force 1 frame number 64-14848 RC-135V scout (PYTHN54) ascends from base I and flies over area H in the morning of 30.1.1.M.

……

In the first step, a typical event in the field of the flight activities of the air target is found by using a field event word clustering method, and taking 1B-52H bombers to take off from a K city base L in a J area as an example.

1a) Performing word segmentation and part-of-speech tagging on the aerial target flight activity text: 1(m), shelf (q), B-52H (n), bomber (n), slave (p), J area (n), K city (n), base L (n), take-off (v).

The dependency parsing results are as follows:

composition relationship	Sentence component
		ATT	(1，Frame)
ATT	(frame, bomber)
		ATT	(B-52H, bomber)
SBV	(Bomb bomber, take-off)
		ADV	(Slave, take-off)
ATT	(J region, K City)
		ATT	(K City, base L)
POB	(base L, from)
		HEAD	(takeoff)

And extracting a main SBV and a subsidiary VOB binary group from the data, wherein only the (bomber, takeoff) meets the condition, and finally extracting the 'takeoff' with the part of speech being a verb as a candidate trigger word. The method is used for extracting the trigger words from a large amount of text data to obtain a candidate trigger word set.

1b) And filtering the noise data in the candidate trigger word set.

A first filtering rule: and keeping the general verbs and the verb nouns in the extracted candidate trigger words, and discarding other types of verbs.

And a second filtering rule: the correlation degree of the candidate trigger words and the air target flight activity field is reflected by using the field correlation degree, and the calculation formula is as follows:

DR(V)＝(Freq_p(V))/(Freq_G(V))

where dr (V) is the domain relevancy value of the candidate trigger word V, Freq _ p (V) is the frequency of the candidate trigger word V appearing in the domain corpus, and Freq _ g (V) is the frequency of the candidate trigger word appearing in the general domain corpus. After sorting, the top word is selected as the final trigger word.

1c) Event trigger word clustering is carried out by adopting semantic similarity, and words with high semantic similarity such as 'take-off' and 'lift-off' are obtained; "warp stop", "docking" and "stop"; "fly-by", "fly", and "coast"; "deploy", and "deploy" and the like.

And secondly, designing trigger words and event elements of the events corresponding to each type of typical event type, which is specifically as follows.

And thirdly, analyzing the text data of the flight activities of the aerial targets and identifying the event types.

3a) Typical event extraction, mainly includes extraction of air target flight time, place and fighter entity (fighter model number, fighter call sign, fighter number and quantity). KC-10A fuel dispensers (MOJO91) numbered 83-0081 on a certain air force 1 morning in 29 months at 2 am take off from A11 area base and fly over B area. For example, the events extracted are:

event 1: { 'events': takeoff ',' trigger ': takeoff', 'time': 2 month 29 morning 'day,' location ': area' and 'base A11' ], model ': KC-10A' ], call sign ': MOJO91' ], number ': 83-0081' ], quantity ': 1' }

Event 2: { 'events': ordinary flight ',' trigger ': flight', 'time': 2 month 29 morning ',' location ': area B': type ': KC-10A' ], call sign ': MOJO91' ], number ': 83-0081' ], number ': 1' }

3b) And extracting the atypical events, and acquiring main elements describing the events in the text by analyzing the relationship among the syntactic components. The conventional Chinese text dependency parsing analysis has 15 annotation relations, namely, a main meaning relation (SBV), a moving object relation (VOB), an inter-object relation (IOB), a preposed object (FOB), a bilingual (DBL), a centering relation (ATT), an intermediate form structure (ADV), a dynamic complement structure (CMP), a parallel relation (COO), a mediating relation (POB), a left additional relation (LAD), a right additional Relation (RAD), an Independent Structure (IS), a punctuation (WP) and a core relation (HED). For example: "1C-17 transporter has entered the area B1 for empty". The dependency parsing structure is shown in fig. 4:

the core word is "enter"; the subject is "transporter"; the predicate is "leading empty"; the subject and predicate in combination with the predicate element (ATT) result in event elements of "1C-17 transport", "area B1 empty". The final event of extraction is { (1C-17 transporter); (entry); (area B1 empty).

Fourthly, due to the particularity of the flight activity data of the aerial targets, the data describes the flight activity of the aerial targets in rows, and a plurality of flight activities of one aerial target are separated by commas to obtain a field containing a single event. The data of a typical event has 14 attributes, respectively: event type, trigger, time, model, call sign, number, quantity, model 2, call sign 2, number 2, quantity 2, direction, location, and label. The four attributes of the model number 2, the call number 2, the number 2 and the number 2 are stored aiming at the flight accompanying object in the flight accompanying event, the attribute is judged to be null by the non-flight accompanying event, the direction attribute is stored aiming at the flight direction in the common flight event, and the attribute is judged to be null by the non-common flight event. The label attribute has an identification function and is used for marking a plurality of events of one-time flight activity of the aerial target, and the events are marked by numbers. Atypical event data has four attributes, respectively: event type, trigger words, event elements and labels, wherein the labels are also used for labeling a plurality of events of the same flight activity, and 341 typical events and 21 atypical events are finally extracted, and comprise partial abnormal events.

And fifthly, constructing an air target flight activity event map according to the event number id, the label and the relationship, and aggregating the events with the same label attribute according to the front-back sequence to obtain a plurality of (id1, id2) binary groups.

And sixthly, performing node combination, intention association, confidence degree assignment and other operations on the air target flight activity event graph to construct an air target intention event graph.

6a) The merging of the event nodes is mainly aiming at the filtering of the event secondary attributes. The intention of the aerial target is generally related to the model of the fighter plane and the place of the action of the fighter plane, the attributes of time, call letters, numbers, quantity and the like play little role, and only three important attributes of the event type, the model of the fighter plane and the place of the action are reserved when event node combination is carried out.

6b) And establishing an intention node according to the association of the intention, matching the flight activity event node and the intention node according to the existing relationship data of the flight activity event and the intention, and establishing a relationship.

6c) The calculation of confidence includes a compliance relationship confidence calculation between the flight activity event nodes and a causal relationship confidence calculation between the flight activity event nodes and the intent. The sequential bearing relation confidence coefficient calculation method mainly adopts a counting method, the same event node points to other event nodes according to the event id relation binary group obtained in the previous section, and then the occupation ratio is respectively calculated to obtain a weight value, namely the confidence coefficient. Causal confidence is similar.

And seventhly, judging the intention preferentially according to the constructed intention reason graph, but considering that the scale of the reason graph is small, the situation that the matching cannot be achieved is caused. If the intention node can not be matched, the intention is matched in a keyword fuzzy matching mode.

7a) And performing intention judgment based on the air target intention event graph, performing matching identification on the factors such as the event type, the model and the place in the extracted structured event according to a specific regular expression, compiling a Cypher query statement, and matching the intention in the intention event graph by using the generated Cypher statement. The Cypher sentences matched with the event elements extracted in 3a) are as follows:

MATCH (a: takeoff { type: 'KC-10A' }) - [ r1: compliance ] - > (B: normal flight { type: 'KC-10A' }) - [ r2: cause ] - > (c: intent) WHERE (a. location CONTAINS 'region A, base A11' OR 'region A, base A11' CONTAINS a. location) AND (B. location CONTAINS 'region B over' OR 'region B' CONTAINS B. location) RETURN (c. type)

The result of the match in the air target intent event graph is "refuel".

7b) Since the scale of the case map is small, matching may not be intended. The method is characterized in that intention matching is carried out in a keyword fuzzy matching mode, and the key point of the keyword fuzzy matching intention is that the membership relation between certain event element keywords and the intention is determined according to fuzzy statistics carried out by experience and human psychological processes. The membership of event element keywords and intentions as shown in tables 1, 2 and 3 is derived in conjunction with the experience of the relevant person and the study of the text data:

TABLE 1 event type and intent membership Table

Event type keywords	Intention to
		Deploying	Military deployment, support
Spiral	Reconnaissance
		Companion flight	Refueling navigationOil filling

TABLE 2 entity type and intent membership Table

Entity type	Intention to
		Transport plane	Transporting personnel and materials
Oiling machine	Refueling and refueling navigation
		Scout plane	Reconnaissance
Training machine	Training and practicing
		Fighter plane	Deterrence
Special mission aircraft	Reconnaissance

TABLE 3 location and intention membership Table

An RC-135U scout (GIVE31) with the number of 64-14847 is lifted from the C region base C11 by 1-month 30-night certain air force 1, and flies above the region D. "is an example. Assuming that the target flight event in the air can not be matched with the intention on the event graph, extracting key elements of the target flight event: the intention of the scout is scout iran according to the membership table.

For another example, by acquiring text data of target activities of aerial vehicles such as unmanned aerial vehicles, the real intentions (simulation training, performance exhibition or the like) of the aerial vehicles are judged according to the text data. Typical events in the air target flight activity field are found by adopting a field event word clustering method, so that the definition of event types and event argument roles in the air target flight activity field is more definite; and the constructed flight activity map and the intention affair map of the aerial target are utilized to visually and clearly show the flight activity change of the aerial target and the incidence relation between the flight activity and the intention, so that the intention judgment of the aerial target is facilitated, the application of the drilling, the simulation training and the virtual reality technology in games is facilitated, and the real-time interactive perception effect of participants is enhanced.

Unless specifically stated otherwise, the relative steps, numerical expressions, and values of the components and steps set forth in these embodiments do not limit the scope of the present invention.

It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.

Finally, it should be noted that: the above-mentioned embodiments are only specific embodiments of the present invention, which are used for illustrating the technical solutions of the present invention and not for limiting the same, and the protection scope of the present invention is not limited thereto, although the present invention is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the embodiments of the present invention, and they should be construed as being included therein. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims (10)

1. An air target intention evidence judging method based on a case atlas is characterized by comprising the following contents:

obtaining typical event types in the activity field of the aerial target according to the flight activity text data of the aerial target;

setting event trigger words and event elements corresponding to each type of typical events, analyzing text data of the flight activities of the aerial targets to extract the events, and acquiring typical events and atypical events in the flight activities;

marking label attributes on all events and combining a database to construct an air target flight activity event map;

and aiming at the flight activity data of the air target to be processed, judging and verifying the intention of the air target by extracting the event and matching the event type based on the event map.

2. The air target intention evidence judging method based on a matter graph according to claim 1, characterized in that for air target flight activity text data, character replacement and data cleaning are firstly carried out through a regular expression, and then typical events are obtained according to a word clustering method.

3. The air target intention evidence judging method based on the affair map as claimed in claim 1 or 2, characterized in that in the typical event type obtaining, through word segmentation, part of speech tagging and syntax analysis on text data, a two-tuple of a principal and a subordinate relation and an action-guest relation is extracted, a candidate trigger word of a flight activity event is determined, and a candidate trigger word set is formed; and filtering the trigger words in the candidate trigger word set through a preset filtering rule, and clustering words based on the semantic similarity of the words to form typical event types in the field of flight activities.

4. The air target intention adjudication method based on a affairs atlas as claimed in claim 3, wherein the preset filtering rules include but are not limited to: general verbs and verb nouns in the set candidate trigger words are reserved, and other types of verbs are filtered out; and/or filtering out trigger words with the correlation degree smaller than a set threshold value according to the correlation degree of the candidate trigger words in the set and the air target flight activity field.

5. A case atlas-based air target intent identification method according to claim 3, wherein the typical event types obtained comprise: take-off, companion flight, warp stop, arrival, normal flight, deployment, hover, and landing.

6. The air target intention evidence judging method based on a affair atlas as claimed in claim 1, wherein in the event extraction, the event type is identified by analyzing text data of flight activity of the air target, and if the event type is a typical event type, the event extraction is carried out by naming entity and pattern matching; if the event type is the atypical event type, obtaining the atypical event trigger words and the event elements through the relationship analysis between the syntactic components.

7. The method as claimed in claim 6, wherein the core word is set as an event trigger word according to the result of the relationship analysis between the syntactic components for atypical event types, and the event elements are obtained by combining the idioms of the subject and object components.

8. The air target intention evidence judging method based on a case map according to claim 1, characterized in that in the case map construction, all events are uniquely numbered, the relationship between the air target flight activity events is mined, a plurality of events in the same data are labeled through tag attributes, the sequential relationship between the events with the same tag attributes is set, and the air target flight activity case map is constructed and stored by using a map database.

9. The air target intention judging method based on the case atlas as claimed in claim 1 or 8, characterized in that for the air target flight activity data to be processed, first, event extraction is performed, and event types are matched through the case atlas; and if the event type cannot be matched through the event map, fuzzy matching is carried out by utilizing the keywords to determine the event type membership.

10. An air target intention evidence judging system based on a case atlas is characterized by comprising: a type acquisition module, an event extraction module, a map construction module and an intention matching module, wherein,

the type acquisition module is used for acquiring typical event types in the activity field of the aerial target according to the flight activity text data of the aerial target;

the event extraction module is used for extracting events by setting event trigger words and event elements corresponding to each type of typical events and analyzing text data of the flight activities of the aerial targets to obtain typical events and atypical events in the flight activities;

the map construction module is used for constructing an air target flight activity event map by labeling label attributes of all events and combining a map database;

and the intention matching module is used for matching the event type according to the flight activity data of the air target to be processed by event extraction and based on the event map and judging the intention of the air target.

Images