JP2003279647A - Apparatus and method for identifying target group, and program therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a target identifying apparatus that can identify a target objective (the kind and the action object of target), without using an observed value in the past in which the target objective was confirmed, even when information for the identification is not available. <P>SOLUTION: An unconfirmed target record D4a comprising a target behavior characteristic time series 16 for specifying a behavior of the target based on an observed data of items about the behavior of the target and a shape of the target, and comprising a reception shape characteristic integration 17 for specifying a change of a shape characteristic of the target is generated, a case record D4b having a prescribed value or more of similarity among the unconfirmed target records D4a with respect to the identified target is extracted out of the case records D4b provided by observation in the past, the behavior of the target for simulation is simulated according to the target behavior characteristic time series 16 for the identified target, so as to generate a simulated case record D6, and the target objective of which the similarity with respect to the unconfirmed target records D4a comes to the prescribed value or more is determined as an identified result, using the simulated case record D6 and the case record D4b. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a target identifying device, a target identifying method and a program for identifying the type of a target such as an aircraft.

[0002]

2. Description of the Related Art FIG. 9 is a block diagram showing a configuration of a conventional target identifying apparatus, which shows a target identifying apparatus for identifying a target model such as an aircraft using radar waves as an observation means. In the figure, reference numeral 100 denotes a radar wave receiving unit, which receives a radar wave related to a target and generates a reception record in which characteristics such as intensity of a reflected wave and a value for each polarization component are summarized. A model-specific storage unit 101 stores similar candidate records in which reception records acquired from the radar wave reception unit 100 in the past and the corresponding models are associated with each other. 10
A model determination unit 2 determines and presents the model having the largest number of models with the similarity between the received records selected by the model-based storage unit 101 being a predetermined value or more as the estimation result.

Next, the operation will be described. A radar wave transmitted from an unillustrated radar antenna and involved in a target aircraft or the like is received by the radar wave receiving unit 100. At this time, the radar wave receiving unit 100 generates a reception record that summarizes the characteristics of the radar waves involved in the target (for example, the intensity of the reflected wave and the value for each polarization component), and stores it in the model-based storage unit 101 as information d1. Send out.

Here, in the radar device for identifying the model disclosed in, for example, Japanese Patent Laid-Open No. 5-157839, the emission system different from the search / tracking radar wave in the functions of the radar antenna and the radar wave receiving section 100 is used. An identification radar wave having a characteristic value is used. As a result, highly accurate spatial and reflective system characteristic values can be obtained, and the identification rate is improved.

Next, the model-specific accumulation unit 101 compares the reception records obtained from the radar wave reception unit 100 with the similarity between the reception record portions of the similar candidate records accumulated in the past to a predetermined value or more. Search for a set or sets having. For this search, identification information unique to each model, which is preset in the model-based storage unit 101, is used. For example, in the target identification method using the inverse synthetic aperture radar device and the inverse synthetic aperture radar image disclosed in Japanese Patent Laid-Open No. 2001-221857, a target shape model representing a unique shape for each model is stored in advance, Similarity determination is performed between this and the inverse synthetic aperture radar image (for example, paying attention to the size of the inverse synthetic aperture radar image).

Further, for example, in the aircraft identification device disclosed in Japanese Patent Laid-Open No. 60-56276, the jet engine compressor rotation speed peculiar to each target aircraft is held in advance. As a result, the model-specific accumulation unit 101
A similarity search is performed by comparing the jet engine compressor speed calculated from the Doppler frequency, which is one of the reflection system characteristic values, with a preset value.

Explaining the similarity, for example, when the received record is represented by a vector composed of three components of vertical-vertical polarization, vertical-horizontal polarization, and horizontal-horizontal polarization, Euclidean between the vectors The degree of similarity is defined by the distance, and the one having the smaller distance is selected as the higher similarity. Thereby, it is possible to select a predetermined number of similar candidate records based on the ascending order of the distance. When a similar candidate record having a high degree of similarity is searched in this way, the model-specific accumulation unit 101 sends the received record of the search result and the similar candidate record as information d2 to the model determination unit 102.

Subsequently, the model determination unit 102 estimates the model based on the similarity between the similar candidate record and the received record obtained from the model-based storage unit 101 and presents it to the user. In this model estimation, the model estimation is performed with the largest number of models among the similar candidate records having the similarity of the received record obtained from the model-specific accumulation unit 101 of a predetermined value or more.

After this, the user confirms the model (for example,
When the correct answer of the model corresponding to the reception record obtained from the radar wave reception unit 100) is obtained, the model determination unit 102 associates the model confirmation with the reception record, and sets the model as the information d3 that is a new similar candidate record. Separate storage unit 101
Send to.

[0010]

Since the conventional target identifying device is constructed as described above, the information for preset model identification, such as the number of revolutions of the jet engine compressor and the model confirmation, can be obtained in the past. Since the similarity search is performed using the observation information of, there is a problem that the model cannot be identified when the information for identification cannot be obtained.

Further, even if there is observation information whose model has been confirmed, in the situation where the number is small, there is a problem that the type can be identified only when the target takes an action similar to the observation information. It was

As described above, in the conventional target identifying apparatus, since it is necessary to use the information for identifying the model or the observation information for which the model is confirmed, there is a problem that it does not function at all for the unknown model for which the above information is not set. there were.

The present invention has been made in order to solve the above-mentioned problems, and a target purpose (the type of the target aircraft such as an aircraft and its behavioral purpose) simulated from the behavioral characteristics and shape of the target such as an aircraft. ), And the information obtained by observation, the target is confirmed even if the information for identification cannot be obtained without using the actual measured value in the past that confirmed the target. An object of the present invention is to obtain a target identification device, a target identification method, and a program that can identify the purpose.

[0014]

SUMMARY OF THE INVENTION A target identifying apparatus according to the present invention provides time-series action information and a target shape that define a target action from observation data of specifications relating to the target action and shape input in time series. Actual measurement information generating means for generating target actual measurement information including shape integration information that defines time-series changes in characteristics, identification result of target information specifying target type and action purpose, and corresponding target actual measurement information Case information accumulating means for accumulating case information including and extracting case information in which the similarity between the target measurement information and the target to be identified is a predetermined value or more among the case information accumulated in the past observations. And the target information for simulation for specifying the type of the target for simulation and the purpose of action and the information about its shape characteristic are set in advance, and the time-series behavior information of the target to be identified and the simulation according to it Of the case information extracted by the simulated case information and the case information accumulating means, which generates the simulated case information including the shape integrated information obtained by simulating the target behavior and the target information for simulation. From the above, there is provided an identification unit that selects information in which the degree of similarity between the shape integration information of the target actual measurement information and the time-series behavior information is a predetermined value or more and outputs the target information corresponding to the information as an identification result.

The target identifying apparatus according to the present invention provides the information relating to the spatial position of the target obtained from the observation data in the order of the observation time and / or the information relating to the spatial position of the target for every predetermined aggregation time. Information in which the aggregated values are arranged in order of observation time is generated as time-series action information.

The target identifying apparatus according to the present invention generates, as shape integrated information, information that represents information about shape characteristics obtained in time series by integrating the information about an arbitrary position coordinate point of the target.

The target identifying apparatus according to the present invention appropriately changes the time width for extracting the time-series action information and the shape integration information from the observation data.

The target identifying apparatus according to the present invention is provided with hierarchical information in which concepts expressing types of goals and action goals are hierarchized and the goal information is one node.

The target discriminating apparatus according to the present invention stores the case information in which the similarity between the target measurement information and the target to be discriminated is a predetermined value or more among the case information accumulated by the case information accumulating means in the past observation. In extracting, the similar search between the time-series action information of both information and the similar search between the shape integration information are performed by different search methods.

In the target identifying apparatus according to the present invention, a plurality of sets of target information for specifying the type of the target for simulation and the purpose of action and information regarding the shape characteristics thereof are set in the simulation information generating means. Simultaneous and simultaneous simulation of the goal of.

The target identifying apparatus according to the present invention has hierarchical information in which the concept expressing the type of goal and the action purpose is hierarchized and the goal information is one node, and the identifying means is the simulated case information and the case information. Among the case information extracted by the accumulating means, the similar shape information extraction unit that extracts the shape integrated information whose similarity to the shape integrated information of the target actual measurement information is a predetermined value or more, and the similar shape information in the hierarchical information. From the shape system estimation unit that extracts, as the shape system estimation result, the target information corresponding to the type of target concept and / or the action purpose common to the corresponding target information, and the case information extracted by the case information storage unit. , A similar behavior information extraction unit that extracts time-series behavior information whose degree of similarity between the target actual measurement information and the time-series behavior information is a predetermined value or more as similar time-series behavior information; A case system estimator that extracts, as a case system estimation result, target information corresponding to the type of goal and / or action goal of the superordinate concept that is common to the goal information corresponding to similar time-series behavior information, and shape system estimation in hierarchical information An identification processing unit that selects, as an identification result, goal information corresponding to the type of goal and / or action goal of the superordinate concept that is common between the result and the case system estimation result.

In the target identifying device according to the present invention, when the identifying means confirms and / or corrects the identification result, a confirmation result reflecting the identification result is generated, and the case information accumulating means stores the confirmation result in the case information. Is stored in association with.

In the target identifying device according to the present invention, if the identifying means has the target information corresponding to the confirmation result in the case information,
The identification process is executed using the target information, and if there is no target information corresponding to the confirmation result, the identification process is executed using the target information corresponding to the identification result.

The target identifying method according to the present invention detects time-series action information that defines a target action from time-series input observation data of specifications regarding the action and shape of the target and a time-series change of a target shape characteristic. An actual measurement information generation step of generating target actual measurement information including the defined shape integration information,
Among the case information that includes the target information identification results that identify the types of goals and action objectives obtained in past observations and the corresponding target actual measurement information, between the target actual measurement information and the target to be identified. Using the case information extraction step of extracting case information in which the degree of similarity is equal to or greater than a predetermined value, and target information for specifying the type and action purpose of the simulated target that are set in advance and information regarding the shape characteristics thereof, A simulation information generation step for generating simulation case information including time-series behavior information of a target to be identified, shape integration information obtained by simulating the behavior of the simulation target in accordance therewith, and simulation target information; From the simulated case information and the case information extracted in the case information extraction step, information in which the similarity of the shape integration information of the target actual measurement information and the time-series behavior information is equal to or more than a predetermined value is selected, and the relevant information is selected. Identification step using target information corresponding to the identification result as an identification result, and a case storing step of generating a confirmation result reflecting the identification result and confirming and / or correcting the identification result and storing the result in association with the case information And with.

A program according to the present invention defines time-series action information that defines a target action and time-series changes of a target shape characteristic from observation data of specifications regarding a target action and a shape that are input in a time series. Actual measurement information generating means for generating target actual measurement information including shape integration information, case information including the target measurement information corresponding to the identification result of the target information specifying the target type and action purpose, and , Case information accumulating means for extracting case information in which the similarity between the target measurement information and the target to be identified among the case information accumulated in the past observation is equal to or more than a predetermined value, the type of the simulated target, and the action purpose. The target information for simulation to be identified and the information about its shape characteristics are preset, and the time-series behavior information of the target to be identified and the behavior of the target for simulation according to the information are obtained by simulation. The integrated shape information of the target actual measurement information from the case information extracted by the simulated information generation means, the simulated case information and the case information storage means for generating simulated case information including the shape integrated information and the simulated target information. The computer is caused to function as an identification unit that selects information having a similarity of the time-series behavior information equal to or greater than a predetermined value and outputs target information corresponding to the information as an identification result.

[0026]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below. Embodiment 1. 1 is a block diagram showing the configuration of a target identifying apparatus according to Embodiment 1 of the present invention. In the figure, reference numeral 1 denotes a remote monitoring unit (measurement information generating means) which acquires data obtained by observing a target and generates a monitoring record D1 and a simulation record D2 for each observation. The monitoring record D1 is information including a simulation record D2 and a reception shape characteristic indicating a target static characteristic obtained by observation. The simulation record D2 includes an identifier for identifying observation data for each target, a monitoring characteristic indicating the observation condition of the target,
And information that consists of target behavioral characteristics indicating the dynamic characteristics of the target obtained by observation. Reference numeral 2 denotes a target purpose storage unit (actual measurement information generation unit, case information storage unit), which has the same identifier from the monitoring records D1 acquired from the remote monitoring unit 1 and has a monitoring record at an observation time within a predetermined time width. D
1 uncollected target record (target measurement information) D4a
To generate. This unconfirmed target record D4a is subjected to integrated processing of a target behavior characteristic time series (time series behavior information) obtained by time-series processing of a target behavior characteristic representing a target dynamic characteristic, and a reception shape characteristic obtained in time series. This is information including the obtained reception shape characteristic integration (shape integration information).

Further, the target / purpose storage unit 2 generates a set of a single or a plurality of case records D4b obtained by performing a similar search on the case records (case information) D4b accumulated in the past. The case record D4b is information including an unconfirmed target record D4a, a target / purpose estimation result (target information) D5a, and a target / purpose confirmation result (target information) D5b. Here, the remote monitoring unit 1 and the target purpose storage unit 2 are
For example, it can be realized by causing a computer device to execute a program having these functions.
Further, as a medium for storing the data acquired by the target / purpose storage unit 2, a hard disk device or the like mounted in a computer device embodying the present invention may be used.

A monitoring data simulation generator (simulation information generator) 3 generates a target behavior characteristic time series for the target behavior characteristic in the simulation record D2 and uses the target behavior characteristic time series. The simulated reception shape characteristic integration is calculated to generate a simulated case record (simulated case information) D6 for each simulated target purpose. Reference numeral 4 denotes a target / purpose determination unit (identifying means), which is a target / purpose estimation result D in which the target / purpose (target information) of the unconfirmed target record D4a is estimated based on the set of the case records D4b and the simulated case record D6.
5a and a target / purpose confirmation result D5b, which is the confirmation result, are generated. Here, the monitoring data simulation generation unit 3 and the target purpose determination unit 4 can be realized, for example, by causing a computer device to execute a program having these functions.

FIG. 2 is a block diagram showing the configuration of the target / purpose determination unit in FIG. In the figure, 5 is a target shape vector extraction block, which extracts the received shape characteristic integration of the unconfirmed target record D4a as the target shape vector D7. Reference numeral 6 is a case shape vector set extraction block, which extracts each received shape characteristic integration of the set of case records D4b as a case shape vector set D8. Reference numeral 7 is a simulated shape vector set extraction block, which is a simulated case record D.
Each simulated shape characteristic integration of the set of 6 is extracted as a simulated shape vector set D9. Reference numeral 8 is a similar shape vector set generation block (similar shape information extraction unit), which is a shape vector (similar shape information) whose shape similarity with the target shape vector D7 is equal to or greater than the reference from the case shape vector set D8 or the simulated shape vector set D9. ) Is extracted to generate a similar shape vector set. Reference numeral 9 denotes a common target / object search block (shape system estimation unit) which extracts the target object confirmation result D5a corresponding to the similar shape vector set or the common target object in the entire similar shape vector set of the simulated target object to obtain the shape system estimation result. It is sent to the target object estimation block 11 as D10.

Reference numeral 10 denotes a common goal / purpose extraction block (similar behavior information extraction unit), which determines a predetermined behavior similarity from the target behavior characteristic time series of the unconfirmed target record D4a from the set of case records D4b acquired from the target / purpose storage unit 2. The case record D4b which it has is extracted and a similar action case record set (similar time series action information) is generated. Also, the common goal / purpose extraction block 10 extracts the common goal / purpose in the entire similar behavior case record set of the goal / purpose confirmation result D5b of each case record D4b included in the similar behavior case record set, and sets the target as the case system estimation result D11. It is sent to the target estimation block 11. Reference numeral 11 denotes a target / object estimation block (identification processing unit), which searches for a common target / object between the shape system estimation result D10 and the case system estimation result D11 to obtain the target object estimation result D
5a is generated. Further, the target / objective estimation block 11 is
If the user confirms or corrects the target purpose estimation result D5a, the target purpose confirmation result D5b that reflects the confirmation or correction is generated.

Next, the operation will be described. FIG. 3 is a flowchart showing the operation of the target identifying apparatus according to the first embodiment, and the operation will be described in detail with reference to this figure. First, the remote monitoring unit 1 acquires data obtained by observing a target in time series and generates a monitoring record D1 and a simulation record D2 at each observation time point (step ST1). For example, observation data for a far away target is acquired by an optical observation means such as a laser beam or an observation device (not shown) having a radar wave as an observation means. At this time, data obtained by observing the target by differently arranged observation devices may be acquired.

FIG. 4 is a diagram showing the structure of the monitoring record and the simulation record generated by the remote monitoring section in FIG. In the figure, D1 is a monitoring record, which is composed of a simulation record D2 and a reception shape characteristic 15. D2 is a simulation record, which is composed of an identifier 12, a monitoring characteristic 13, and a target behavior characteristic 14. Reference numeral 12 is an identifier, which is information for identifying observation data for each target. Reference numeral 13 is a monitoring characteristic, which is information indicating an observation condition. This monitoring characteristic 13
Is uniquely determined in one monitoring record D1. Reference numeral 14 is a target behavior characteristic, which is information indicating the dynamic characteristic of the target obtained by the observation. Reference numeral 15 is a reception shape characteristic, which is a target static characteristic obtained by observation. For example, there is a target two-dimensional shape when optical observation is performed on the target three-dimensional shape from the observation position direction.

Here, the above operation will be specifically described by taking an observation by a radar pulse as an example. First, the remote monitoring unit 1
When acquiring the data obtained by observing the target in time series from the observation equipment, generates an identifier 12 that identifies the observation data from the target and associates the data with each observation time point. At this time, the remote monitoring unit 1 acquires information about the observation conditions acquired from the observation equipment at each observation time point, for example, the radar carrier frequency, the radar site position coordinates, the radar wave direction,
Intensity, emission time, reception time, etc. are summarized as the monitoring characteristic 13.

Further, the remote monitoring section 1 uses the observation data input in time series to obtain information that guides the dynamic characteristics of the target, such as the radar wave direction and round trip time, the radar wave Doppler frequency, and continuous observation. Extract the time-series transition of the target position. Using these pieces of information, the remote monitoring unit 1 calculates the target behavior characteristic 14 indicating the dynamic characteristic of the target. As the target behavior characteristic 14, for example, the spatial position of the target,
There are velocity vector and acceleration vector.

Subsequently, the remote monitoring unit 1 obtains information for guiding the static characteristics of the target from the observation data input in time series, for example, the two-dimensional distribution of the Doppler frequency of the radar wave, the vertical polarization of the radar wave and the horizontal direction. Extract the vertical polarization component and horizontal polarization component of the reflected wave obtained by emitting polarized waves, the distribution of the target reflection intensity obtained by separating the transmission position and the reception position of the radar wave, etc. . Collecting these pieces of information, the remote monitoring unit 1 collects the target static characteristics (shape characteristics).
The reception shape characteristic 15 indicating

As described above, the remote monitoring unit 1 obtains the monitoring characteristic 13, the target behavior characteristic 14 and the reception shape characteristic 15 for each observation time point of the observation data identified by the identifier 12, and is shown in FIG. Such a monitoring record D1 and a simulation record D2 are generated. Here, the monitoring record D1
Is a dynamic property of the target to be observed,
This is information that summarizes static characteristics and observation conditions. On the other hand, the simulation record D2 does not have information about the static characteristics of the target, but is information that summarizes the dynamic characteristics of the target to be observed and the observation conditions. After that, the monitoring record D1 is sent to the target purpose accumulation unit 2, and the simulation record D2 is sent to the monitoring data simulation generation unit 3.

Next, the target / purpose storage unit 2 is connected to the remote monitoring unit 1.
Of the monitoring records D1 acquired from the above, those having an observation time within a predetermined time width are extracted. After that, the target / purpose storage unit 2 sets the target behavior characteristic 1 for each of the same identifiers 12.
4 is processed in time series, the reception shape characteristic 15 is integrated,
Generate an unconfirmed target record D4a (step ST
2).

FIG. 5 is a diagram showing the structure of the unconfirmed target record and case record generated by the target purpose storage unit in FIG. In the figure, D4b is a case record, which is composed of an unconfirmed target record D4a, a target purpose estimation result D5a, and a target purpose confirmation result D5b. D4a is an unconfirmed target record, which is composed of a target behavior characteristic time series 16 and a reception shape characteristic integration 17. Reference numeral 16 denotes a target behavior characteristic time series (time series behavior information), which is information obtained by performing time series processing on the target behavior characteristic 14. Reference numeral 17 is reception shape characteristic integration (shape integration information), which is information obtained by integrating the reception shape characteristics 15 obtained in time series. D5a is a target purpose estimation result, and D5b is a target purpose confirmation result that reflects the result confirmed or corrected by the user based on the target purpose estimation result D5a.

Here, the time series processing by the target purpose accumulator will be described with reference to FIG. Goal purpose accumulation section 2
When the remote monitoring unit 1 handles observation data from a plurality of observation devices, is not the order in which this device acquires the observation data, but the time series in which the target behavior characteristics 14 in the monitoring record D1 are arranged in the observation time order. A certain target behavior characteristic time series 16 is generated. In the present invention, arranging the target behavior characteristics 14 having the observation times within the predetermined time width in the order of the observation times in this manner is called time series processing.

At this time, a moving average value based on a predetermined singular number or a plurality of time widths for totalization of the target behavior characteristic 14 arranged in the order of observation time is obtained, and a time series in which the moving average value is arranged in the order of observation time is set as the target behavior characteristic. The time series 16 may be used. Specifically, when the monitoring record D1 is obtained every 1 second of observation time, if the predetermined time width is less than 10 seconds, that is, 10 monitoring records D1 and the time width for aggregation are 5 seconds, a total of 5 Individual moving average values are obtained for each target behavior characteristic 14. The time width for aggregation is determined independently of the time width used to generate the unconfirmed target record D4a. In the present invention,
It is also included in the time-series processing to include, in the time-series processing, arranging the partial or entire time series in which the desired behavior characteristics 14 are arranged in the order of the observation time as a new characteristic.

Next, with reference to FIG. 5, the integration processing by the target / purpose storage unit 2 will be described. In the present invention, integrating the information regarding the shape characteristics centering on an arbitrary position coordinate point of the target is called integration processing. For example, consider a case where an image of an aircraft that makes a linear motion is optically captured and obtained as observation data. First, in an aircraft that makes a linear motion, different planes of the aircraft are usually observed over time. In other words, when three images are obtained: an image from the front 0 degree direction at time 0, an image from the right 10 degree elevation angle 0 degree direction after 2 seconds, and an image from the right 20 degree elevation angle 0 degree direction after 3 seconds. When the integration process is performed, there are three images, that is, an image from the front 0 ° direction, an image from the right 10 ° elevation angle 0 ° direction, and an image from the right 20 ° elevation angle 0 ° direction. The integration process will be described in detail below.

FIG. 6 is a diagram showing an example of observation using radar waves by the target discriminating apparatus of the first embodiment, and FIG. 6A shows a case where the targets are observed in time series by two observation devices,
(B) shows the result after the integration process. In the figure,
Reference numeral 18 denotes an aircraft, which is a target, and makes a linear motion along a coordinate axis A defined in the traveling direction. 19a, 19b
Is a coordinate point on the coordinate axis A and corresponds to the time when the observation of the aircraft 18 is performed by the observation devices 20 and 21 using radar waves as observation means. Reference numeral 19c is a coordinate point serving as a reference for the integration processing. Reference numerals 20 and 21 denote observation devices arranged at two different points, and sequentially transmit observation data to the target identifying device of this embodiment. 22a to 22f are round-trip paths for radar waves, and 23 to 26 are paths for radar waves.

Upon receiving the monitoring record D1 generated by the processing in step ST1, the target object storage unit 2
The reception shape characteristic 15 in the monitoring record D1 is extracted. Here, the reception shape characteristic 15 is information for specifying the path of the radar wave as shown in (a), for example. That is, in the observation device 20, the round trip paths 22a, 22 of the radar wave are generated at the coordinate points 19a, 19b corresponding to the two observation points.
c is formed, and the reflected wave of the radar wave emitted by the observation device 21 along the path 23 is captured along the path 24. On the other hand, in the observation equipment 21, a round-trip route 22b of a radar wave is formed with respect to the aircraft 18 at the coordinate point 19a. The observation data that gives the paths of these radar waves are summarized as a reception shape characteristic 15 that represents the static characteristics of the target aircraft 18.

Here, when the integration processing is applied to the above-mentioned two reception shape characteristics 15 with reference to the coordinate point 19c on the coordinate axis A, the result is as shown in (b). That is, the two reception shape characteristics 15 cause two coordinate points 19a, 1
Information regarding the route of the radar wave obtained for the aircraft 18 located at 9b is collected based on the aircraft 18 located at one coordinate point 19c. In other words, in the integration processing, the position coordinates relating to the path of the radar wave at each observation time point described above are converted into those having the coordinate point 19c as the center. In the illustrated example, the round trip path 2
2f corresponds to the round-trip route 22b of the radar wave from the observation device 21, the round-trip route 22d corresponds to the round-trip route 22a of the radar wave from the observation device 20, and the round-trip route 22e returns the round-trip route of the radar wave from the observation device 20. 22c. Further, the routes 25 and 26 correspond to the routes 23 and 24 until the radar wave emitted from the observation device 21 is captured by the observation device 20.

Taking the above-mentioned aircraft that makes a linear motion as an example, an image from the front 0 degree direction at time 0, the right 1 after 2 seconds.
Image from 0 degree elevation angle 0 degree direction, right 20 degrees elevation angle 0 after 3 seconds
When three images from the direction of degrees are obtained, when the integration process is performed on one coordinate point centered on the aircraft, an image from the front 0 degree direction, an image from the right 10 degree elevation angle 0 degree direction, Right 20
An image is obtained from the direction of 0 degree elevation angle.

In this way, the target purpose accumulating section 2 integrates the reception shape characteristics 15 in the monitoring record D1 for each observation time point along the time series of the target behavior characteristics time series 16 to integrate the reception shape characteristics. 17 is generated. That is, the reception shape characteristic integration 17 corresponds to information in which changes in the reception shape characteristic 15 along the time series of the target behavior characteristic time series 16 are summarized centering on the target.

After that, the target purpose accumulating section 2 generates an unconfirmed target record D4a from the target behavior characteristic time series 16 and the reception shape characteristic integration 17, and sends it to the target purpose determining section 4. The unconfirmed target record D4a is composed of time-series data of the target dynamic characteristics and data indicating changes in the target shape at each observation time point. Step S described above
The processing in T1 and step ST2 corresponds to the actual measurement information generation step.

It should be noted that the target / purpose storage unit 2 stores the monitoring record D
For the predetermined time width for extracting 1, the unconfirmed target record D4a can be generated periodically using a single time width. Further, it is also possible to generate an unconfirmed target record D4a having different time widths at the same time using a plurality of time widths.

Subsequently, the target / purpose accumulating section 2 performs a similarity search of the unconfirmed target records D4a obtained in step ST2 from the case records D4b accumulated in the past to obtain a set of case records D4b having a predetermined degree of similarity or higher. Generate (step ST3, case information extraction step). More specifically, the target / purpose storage unit 2 compares the target behavior characteristic time series 16 of the case record D4b accumulated in the past with the target behavior characteristic time series 16 of the unconfirmed target record D4a obtained in step ST2, for example. Then, the target movement characteristic represented by the time series 16 of target behavior characteristics is selected such that the transition of the target movement does not reach a predetermined difference. The movement of the target may be a change in the position of the target or a change in the velocity vector. The set of case records D4b generated in this manner is sent to the target / purpose determination unit 4.

In addition to this, for example, the reception shape characteristic integration 17 of the case record D4b is compared with the reception shape characteristic integration 17 of the unconfirmed target record D4a obtained in step ST2, and the reception shape characteristic integration 17 is expressed. You may make it select according to whether shape characteristics satisfy | fill a predetermined reference | standard. For example, a high-dimensional space in which the shape characteristic in the similar direction within the predetermined angular width on the relative coordinates centered on the target position, which is represented by the reception shape characteristic integration 17, has the dimension number as the number of the characteristic values. It is regarded as a point inside and the point whose Euclidean distance is below a predetermined standard is selected.

Next, the generation of the simulated case record by the monitoring data simulation generator will be described. First, the purpose of the present invention will be described. In the present invention, the goal / objective is expressed as one node of hierarchical information (hereinafter, referred to as goal / objective hierarchy) in which the goal / objective type and the action purpose thereof are classified.

FIG. 7 is a diagram showing an example of a goal / objective hierarchy composed of goal goals. In the figure, the target / target layer is composed of a first layer 27a, 27b, a second layer 28 and a third layer 29. In the first layer 27a, major classification nodes of target types such as "aircraft" and "helicopter" are defined. Further, in the first hierarchy 27b, classification nodes for target action objectives such as "reconnaissance" and "transportation" are defined. In the second layer 28, nodes are defined which further classify the types of targets such as "fighters", "transport planes", "passenger planes", "reconnaissance helicopters", and the like. A model node that specifies a target model (type) such as “F15” or “B747” is defined in the third layer 29.

FIG. 8 is a diagram showing the structure of a simulated case record generated by the monitoring data simulation generator in FIG. In the figure, D6 is a simulated case record, which is composed of a target behavior characteristic time series 16a, a simulated reception shape characteristic integration 17a, and a simulated target purpose 30. Reference numeral 16a denotes a target behavior characteristic time series, which is generated by time-series processing of the target behavior characteristic 14 in the simulation record D2. Reference numeral 17a denotes a simulated reception shape characteristic integration, which is generated from the simulated target purpose 30, the monitoring characteristic 13 in the simulation record D2, and the desired behavior characteristic time series 16a. 30 is a simulation target purpose, and the monitoring data simulation generation unit 3
It is a target goal for simulation that is set in advance. here,
It is assumed that simulated target objectives 30 relating to a plurality of targets are preset in the monitoring data simulation generator 3. The simulated target purpose 30 is represented by, for example, the third layer 29 in FIG. That is, the simulated target purpose 30 is, for example, “F15” or “B747” in the third hierarchy 29, and has a combination of parameter values that specify a model that expresses its static characteristics (shape characteristics).

More specifically, the model according to the simulated target purpose 30 is expressed as, for example, an aggregate of planes having a small surface shape, and the radio wave characteristic is expressed as an attribute of the plane. In addition, the parameters that specify the model include, for example, options of posture and shape change. Here, the posture is defined in the nose direction in the three-dimensional coordinates with the horizontal plane and the vertical direction as references. Further, when a part of the body is deformed, a plurality of surface shape aggregates are defined in one model as the simulated target purpose 30, and one of them is selected according to the above parameters.

When the simulation record D2 generated in step ST1 is received, the monitoring data simulation generator 3 selects the monitoring record D1 of the simulation record D2 which is the processing target of the target purpose accumulator 2 in step ST2. A simulation record D2 having an observation time within the same time width as is extracted. Then, the monitoring data simulation generator 3 performs step S
At T2, the target behavior characteristic 14 in the simulation record D2 is time-sequentially processed for each identifier 12 that is the same as the data that the target purpose accumulating unit 2 has processed, and the target behavior characteristic time series 16a.
Is generated (step ST4). That is, the target behavior characteristic time series 16a has the same content as the target behavior characteristic time series 16 generated by the target purpose accumulating unit 2.

Subsequently, the monitoring data simulation generation unit 3 takes the behavior represented by the target behavior characteristic time series 16a for the simulated target purpose 30, and according to the monitoring characteristic 13 of the simulation record D2 acquired from the remote monitoring unit 1. Assuming that the observation has been performed, the same processing as step ST2 is performed to calculate the simulated reception shape characteristic integration 17a. After that, the monitoring data simulation generator 3 associates each simulated target purpose 30 with the target behavior characteristic time series 16 a and the simulated reception shape characteristic integration 17.
A simulated case record D6 including a is generated (step ST5). The processing in steps ST4 and ST5 corresponds to the simulated information generation step.

Explaining with a concrete example, when simulating the observation by the radar pulse, the monitoring data simulation generator 3 is based on the attitude parameter of the simulated target object 30, centered on the simulated target object 30, and in the nose direction. And polar coordinates with respect to the horizontal plane in the design of the airframe are defined. Thus, the angle at which the radar pulse arrives and the direction in which the radar pulse is reflected are observed, and the emission position and the observation position of the radar pulse are converted to the polar coordinates.

Next, using the model of the target based on the simulated target purpose 30, the radar pulse arrival angle and the observation angle obtained as described above, the transmission polarization characteristic and the reception polarization characteristic of the radar pulse, Simulate the radar pulse reflected wave characteristics for each polarization characteristic in. The posture parameter of the simulated target purpose 30 can also be estimated from the acceleration vector of the target behavior characteristic time series 16a, for example. Alternatively, the horizontal attitude can be simply determined by a random number having a predetermined variance. In this way, the radar pulse reflected wave characteristic for each polarization characteristic at the predetermined observation position, that is, the simulated reception shape characteristic integration 17a is obtained.

A plurality of sets of simulated reception shape characteristic integration 17a may be generated by setting a plurality of model parameters for one simulated target purpose 30.

The operation of calculating the simulated reception shape characteristic integration 17a for one set of model parameters of a single simulated target object 30 is one unit of processing, and each model parameter of all the preset simulated target objects 30 is set. May be executed in parallel by a plurality of processors.
As a result, the simulation reception shape characteristic integration 17a and the simulation case record D6 can be calculated at higher speed.

In the conventional target identifying device, the received record d
As information to be compared with 1, it is necessary to preset information that specifies the target dynamic characteristic and static characteristic as identification information. On the other hand, in the present invention, only the simulated target purpose 30 is preset. The simulated case record D6 to be compared with the data actually measured in the present invention is the target behavior characteristic time series 16a of the data representing the static characteristic uniquely determined by the simulated target purpose 30 and the simulated reception shape characteristic integration 17a. It is information indicating a time-series change in the operation defined by. By comparing the simulated case record D6 with the case record D4b which is the observation data to identify the target, the target purpose can be estimated even in the situation where the information for identification is not available. In addition, it is possible to estimate the closest target to the unknown target.

When the monitoring data simulation generator 3 generates a set of simulated case records D6 relating to all the preset simulated target purposes 30 as described above, it sets the simulated case record D6 to the target purpose determination unit. Send to 4.

In the target / purpose determination unit 4, the case record D4
Based on the set b and the simulated case record D6, the target purpose regarding the unconfirmed target record D4a is estimated (step ST6, identification step). This will be described in detail with reference to FIG. First, the set of the case records D4b from the target purpose accumulating unit 2 is input to the target shape vector extracting block 5, the case shape vector set extracting block 6 and the common target object extracting block 10 in the target purpose determining unit 4, respectively.

Here, the target shape vector extraction block 5
Then, the unconfirmed target record D4a acquired from the target purpose accumulating unit 2 is extracted, and the information collected in the reception shape characteristic integration 17 is extracted as the target shape vector D7. This target shape vector D7 is the received shape characteristic integration 1
This information is composed of combinations of parameter values that specify the model expressing the static characteristics of the target summarized as 7. For example, a shape vector that specifies the direction of the target on the relative coordinates centered on the target position may be used.

On the other hand, in the case shape vector set extraction block 6, the unconfirmed target records D4a are extracted from the set of case records D4b, and the information gathered in each reception shape characteristic integration 17 is used as the case shape vector set D.
Extract as 8. This case shape vector set D8
Is information indicating a time series change of the target shape vector D7.

The set of simulated case records D6 from the monitoring data simulation generator 3 are sequentially input to the simulated shape vector set extraction block 7 in the target purpose determination unit 4.
Here, the simulated shape vector set extraction block 7 extracts, as a simulated shape vector set D9, the information collected in the simulated reception shape characteristic integration 17a extracted from each set of the simulated case records D6. The simulated shape vector set D9 is information that represents a time-series change of information that is a combination of parameter values that specify a model that expresses the static characteristics of the simulated target purpose 30 that are summarized as the simulated reception shape characteristic integration 17a.

The target shape vector D7, the case shape vector set D8 and the simulated shape vector set D9 respectively extracted by the target shape vector extraction block 5, the case shape vector set extraction block 6 and the simulated shape vector set extraction block 7 are similar shapes. It is sent to the vector set generation block 8.

Similar shape vector set generation block 8
Is a shape vector whose shape similarity with the target shape vector D7 is greater than or equal to the shape similarity criterion obtained from the common target object search block 9 from the case shape vector set D8 and the simulated shape vector set D9. , And a similar shape vector set is generated by summing them up. Here, the shape vector refers to the element vector of the case shape vector set D8 or the simulated shape vector set D9, or the target shape vector D7.

Further, the shape similarity is, for example, in a high dimensional space where the number of dimensions is the number of the characteristic values of the shape vector in the direction within a predetermined angular width on the relative coordinates centering on the target position. It is indexed by the Euclidean distance between the shape vectors when it is regarded as the point of. In this case, the similar shape vector set generation block 8 selects, from the case shape vector set D8 and the simulated shape vector set D9, a shape vector in which the Euclidean distance between the shape vectors is smaller than a predetermined reference, and selects the similar shape vector. Generate a set. The similar shape vector set thus generated is sent to the common goal search block 9.

In defining the distance, the Manhattan distance may be used to make it robust against noise in the case. Furthermore, the Mahalanobis distance may be defined for each of the case shape vector set D8 and the simulated shape vector set D9 to improve the accuracy of similarity determination.

The target / object determination section 4 is preset with target / objective hierarchy as shown in FIG. 7, that is, hierarchical information in which the target / objective is classified as one node with respect to the type of the target and its action / objective. Further, each similar shape vector in the similar shape vector set is associated with the target purpose confirmation result D5b via the case record D4b or the simulated target purpose 30 via the simulated case record D6.
Therefore, the common target / purpose search block 9 uses the above-mentioned target / target layer to calculate the target / target confirmation result D5b of the case record D4b or the simulated target / purpose 30 of the simulated case record D6 corresponding to each similar shape vector. Search for common goals and objectives in the entire vector set.

That is, the common goal search block 9 is
The case record D4b corresponding to the case shape vector set D8 in which the shape vector is extracted as the similar shape vector
Alternatively, each of the simulated case records D6 corresponding to the simulated shape vector set D9 is selected. Then, by using the target purpose confirmation result D5b or the simulated target purpose 30 corresponding to these, the search for the common target purpose between the similar shape vectors is sequentially executed to obtain the common target purpose in the entire similar shape vector set. Note that case record D4
When there is no target purpose confirmation result D5b in b, the common target purpose search block 9 substitutes the target purpose estimation result D5a in the case record D4b.

Here, the common goal purpose refers to a common parent of the goal goal nodes in the goal goal hierarchy. For example, the common target purpose of “F15” on the third hierarchy 29 and “transport aircraft” on the second hierarchy 28 in FIG. 7 is “aircraft” on the first hierarchy 27a.

At this time, if there is a common goal / object, the common goal / object search block 9 sends the common goal / object as a shape system estimation result D10 to the target / object estimation block 11, respectively. When there is no common goal / object, the common goal / object search block 9 sends a similarity criterion having a lower criterion value than the previous one to the similar shape vector set generation block 8. As a result, the similar shape vector set generation block 8 is caused to generate a similar shape vector set consisting of shape vectors selected by the similarity criterion having a low determination reference value, and the search is repeated until the common target object is found.

On the other hand, in the common goal / purpose extraction block 10, the target behavior characteristic time series 16 is extracted from the set of the unconfirmed target record D4a and the case record D4b acquired from the target / purpose storage unit 2, and the target of the unconfirmed target record D4a is extracted. A case record D4b having a predetermined behavior similarity to the behavior characteristic time series 16 is extracted, and these are combined to generate a similar behavior case record set.

Thereafter, the common goal / purpose extraction block 10
Is the target / purpose confirmation result D5b of each case record D4b included in the similar action case record set with respect to the target / purpose layer processed in the common target / purpose search block 9.
The common goal and objective of the entire set of similar behavior example records of is extracted. That is, the common goal / purpose extraction block 10
Sequentially searches for the common goal / object between the goal / object confirmation results D5b of the case record D4b included in the similar action case record set to obtain the common goal / object in the entire similar action case record set.

At this time, if a common target purpose exists, this is sent to the target purpose estimation block 11 as a case system estimation result D11. If there is no common target purpose, the null value is sent to the target purpose estimation block 11.

Here, the behavior similarity can be indexed by the minimum value of the Euclidean distance obtained as follows, for example. First, a plurality of subsequences having a predetermined length are extracted from the target behavior characteristic time series 16 of the two targets A and B whose behavior similarities are to be obtained, and the average speed, the average horizontal position, and the average altitude in the subsequences are obtained. , The average behavior vector. Next, the Euclidean distance is calculated for all combinations of the average action vector of the target A and the average action vector of the target B. The minimum value of this distance is an index of action similarity.

Subsequently, the target / target estimation block 11 determines the shape system estimation result D10 obtained from the common target / target search block 9 and the common target for the target / target layer processed by the common target / target search block 9. Objective extraction block 10
The common goal and purpose with the case system estimation result D11 obtained from are searched. At this time, the common target purpose extracted as the search result is set as the target purpose estimation result D5a.

Further, the target / target estimation block 11 presents the target / target estimation result D5a obtained as described above to the user. Specifically, it is conceivable to display the target / target estimation result D5a on the display unit of the computer device that embodies the target identifying apparatus of the present invention.

After that, when confirmation or correction is received from the user for the target / object estimation result D5a, the target / object estimation block 11 reflects the confirmation or correction in the target / object estimation result D5a. To generate. The target purpose estimation result D5a and the target purpose confirmation result D5b generated in this way are stored in the target purpose storage unit 2
Sent to.

In the target purpose accumulating unit 2, a new case record D4b is generated and accumulated by combining the target purpose estimation result D5a and the target purpose confirmation result D5b from the target purpose estimation block 11 and the unconfirmed target record D4a. (Step ST7, case storing step).

As described above, according to the first embodiment, the target action time series 16 that defines the target action from the observation data of the specifications regarding the target action and the shape that are input in time series, and the target action time series 16. An unconfirmed target record D4a including a reception shape characteristic integration 17 that defines a time-series change of shape characteristics
Is generated, and the target purpose estimation result D obtained in the past observation
5a, a target / purpose confirmation result D5b, and a case record D4b including an unconfirmed target record D4a corresponding thereto.
Of the unidentified target record D4a with the target to be identified
While extracting the case record D4b in which the similarity between two or more is a predetermined value or more, the target of the target to be identified by using the simulated target purpose 30 that specifies the type of the simulated target and the action purpose that are set in advance. Generate a simulated case record D6 including a simulated reception shape characteristic integration 17a obtained by simulating a target behavior for simulation according to a target behavior characteristic time series 16a that is the same information as the behavior characteristic time series 16; From the simulated case record D6 and the case record D4b, information whose similarity to the unconfirmed target record D4a is a predetermined value or more is selected, and the target purpose corresponding to the information is output as an identification result. The target purpose can be identified even in the situation where the identification information cannot be obtained without using the past measured values of the distorted shape characteristics.

[0084]

As described above, according to the present invention, the time-series action information for defining the target action and the target shape characteristic from the observation data of the specifications regarding the action and the shape of the target input in time series. Target measurement information including shape integration information that defines time-series changes is generated, and the target information identification results that identify the types of goals and action objectives accumulated in past observations and the corresponding target measurement information Of the case information including the target information to be identified and the similarity between the target measurement information and the target actual measurement information are equal to or more than a predetermined value, and the type of the target for simulation and the target for simulation to specify the action purpose Information and information about its shape characteristics are set in advance, and time-series behavior information of a target to be identified and shape integration information and simulation target information obtained by simulating the behavior of the simulation target according to the information are obtained. Including Then, from the simulated case information and the case information, the information in which the similarity of the shape integration information of the target actual measurement information and the time-series behavior information is equal to or more than a predetermined value is selected, and the corresponding information is selected. Since the target information is output as the identification result, there is an effect that the target information can be identified even if the information for identification cannot be obtained without using the past measured value for which the target information has been confirmed. .

According to the present invention, the information obtained by associating the information relating to the spatial position of the target obtained from the observation data in the order of the observation time and / or the information relating to the spatial position of the target is aggregated at every predetermined aggregation time. Since the information in which the values are arranged in the order of the observation time is generated as the time-series action information, there is an effect that the observation data from a plurality of observation devices can be efficiently collected. In addition, since the simulation result is expressed in the same manner, it is possible to generate simulated case information that matches the case information and to easily identify the target information.

According to the present invention, the information about the shape characteristics obtained in time series is integrated and expressed around the arbitrary arbitrary position coordinate point of the target as the shape integrated information. This has the effect of efficiently collecting observation data. In addition, since the simulation result is expressed in the same manner, it is possible to generate simulated case information that matches the case information and to easily identify the target information.

According to the present invention, the time width for extracting the time-series behavior information and the shape integration information from the observation data is appropriately changed, so that, for example, an easily identifiable target is processed by a short time width and processed at high speed. For a difficult target, there is an effect that the convenience of the identification processing can be improved, for example, by increasing the time width and processing with high accuracy.

According to the present invention, since the concept expressing the type of goal and the action purpose is hierarchized and the hierarchical information having the goal information as one node is provided, even in the observation data identified as different types This has the effect of identifying the type of goal and the purpose of action of the concept.

According to the present invention, in extracting the case information in which the similarity between the target measurement information and the target to be identified is equal to or more than a predetermined value from the case information accumulated in the past observations, the case Since the similarity search between the series behavior information and the similarity search between the shape integration information are performed by different search methods, there is an effect that the accuracy of the similarity search can be improved.

According to the present invention, a plurality of sets of target information for specifying the type of the target for simulation and the purpose of action and information regarding its shape characteristic are set in the simulated information generating means, and the plurality of targets for simulation are related. Since the simulations are simultaneously executed in parallel, the simulations can be performed at high speed, and the simulations and identifications can be easily performed in real time.

According to the present invention, the concept expressing the type of goal and the action purpose is hierarchized, and the hierarchical information having the goal information as one node is provided, and the target actual measurement information is selected from the simulated case information and the case information. The shape integration information having a similarity with the shape integration information equal to or more than a predetermined value is extracted as the similar shape information, and the target type and / or the action of the superordinate concept common to the target information corresponding to the similar shape information in the hierarchical information. The target information corresponding to the purpose is extracted as the shape system estimation result, and the time-series behavior information whose similarity with the time-series behavior information of the target actual measurement information is a predetermined value or more from the case information is used as the similar time-series behavior information. Example of target information corresponding to the type of target concept and / or action purpose common to the target information corresponding to the similar time-series action information in the hierarchical information It is extracted as an estimation result, and the target information corresponding to the target type and / or the action purpose of the superordinate concept common between the shape system estimation result and the case system estimation result in the hierarchical information is extracted as the identification result. Even if the type or the action purpose cannot be uniquely identified, it is possible to identify either the type of the target or the action purpose.

According to the present invention, when the identification result is confirmed and / or corrected, the confirmation result reflecting this is generated and the confirmation result is stored in association with the case information. It is possible to increase the number, and it is possible to increase the range in which the target information can be estimated in the next identification.

According to the present invention, if the target information corresponding to the confirmation result is included in the case information, the identification processing is executed using the target information, and if the target information corresponding to the confirmation result is not present, the identification result is determined. Since the identification process is executed using the corresponding target information, there is an effect that the range in which the target information can be estimated can be widened in the next identification.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a target identifying device according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a target / purpose determination unit in FIG.

FIG. 3 is a flowchart showing the operation of the target identifying device according to the first embodiment.

FIG. 4 is a diagram showing a configuration of a monitoring record and a simulation record generated by a remote monitoring unit in FIG.

5 is a diagram showing the configuration of an unconfirmed target record and a case record generated by the target / purpose storage unit in FIG.

FIG. 6 is a diagram showing an example of observation using radar waves by the target identifying apparatus according to the first embodiment.

FIG. 7 is a diagram showing an example of a target / purpose hierarchy composed of target purposes.

8 is a diagram showing a configuration of a simulation case record generated by a monitoring data simulation generation unit in FIG.

FIG. 9 is a block diagram showing a configuration of a conventional target identifying device.

[Explanation of symbols]

1 remote monitoring unit (measurement information generation unit), 2 target purpose storage unit (measurement information generation unit, case information storage unit) 3, monitoring data simulation generation unit (simulation information generation unit), 4 target purpose determination unit (identification unit) 5, target shape vector extraction block, 6 case shape vector set extraction block, 7 simulated shape vector set extraction block, 8 similar shape vector set generation block (similar shape information extraction unit), 9
Common goal / object search block (shape system estimation unit), 10 common goal / object extraction block (similar behavior information extraction unit), 11
Target purpose estimation block (identification processing unit), 12 identifier, 13 monitoring characteristic, 14 target behavior characteristic, 15 reception shape characteristic, 16, 16a target behavior characteristic time series (time series behavior information), 17, 17a reception shape characteristic integration ( Shape integration information), 18 aircraft (target), 19a-19c coordinate points, 20,21 observation equipment, 22a-22f round-trip route,
23-26 route, 27a, 27b first hierarchy, 28
2nd layer, 29 3rd layer, 30 Simulated target purpose, D1
Monitoring record, D2 simulation record, D4a unconfirmed target record (target actual measurement information), D4b case record (case information), D5a target purpose estimation result (target information), D5b target purpose confirmation result (target information), D6
Mock case record (mock case information).

Claims (12)

[Claims] 1. Time-series action information that defines the action of the target from observation data of specifications relating to the action and shape of the target that are input in time series, and shape integration information that defines time-series change of the shape characteristic of the target. The actual measurement information generating means for generating the target actual measurement information including the, and the case information including the target measurement information corresponding to the identification result of the target information specifying the type of the target and the action purpose and the corresponding target actual measurement information are accumulated. , Case information accumulating means for extracting case information in which the similarity between the target measurement information and the target to be identified among the case information accumulated in the past observation is a predetermined value or more, the type of the simulated target, and the action purpose The target information for simulation and the information about its shape characteristics for identifying the target are set in advance, and are obtained by simulating the time-series behavior information of the target to be identified and the behavior of the target for simulation according to the time-series behavior information. Of the target actual measurement information from the simulated information generation means for generating simulated case information including the situation integrated information and the simulated target information, and the case information extracted by the simulated case information and the case information storage means. A target identification device comprising: an identification unit that selects information whose similarity between the shape integration information and the time-series behavior information is a predetermined value or more and outputs target information corresponding to the information as an identification result. 2. The information obtained by associating the information relating to the spatial position of the target obtained from the observation data in the order of observation time and / or the aggregated value obtained by aggregating the information relating to the spatial position of the target at a predetermined aggregation time. The target identifying apparatus according to claim 1, wherein the information arranged in chronological order is generated as time-series action information. 3. The target according to claim 1, wherein the information about the shape characteristic obtained in time series is integrated and expressed centered on an arbitrary position coordinate point of the target to generate as shape integrated information. Identification device. 4. The target identifying device according to claim 1, wherein the time width for extracting the time-series action information and the shape integration information from the observation data is changed as appropriate. 5. The goal identifying apparatus according to claim 1, further comprising hierarchical information in which concepts expressing goal types and action goals are hierarchized and goal information is one node. 6. The case information storage means extracts, from the case information accumulated in the past observations, case information in which the similarity between the target actual measurement information and the target to be identified is a predetermined value or more. The target identifying apparatus according to claim 1, wherein the similarity search between the time-series action information and the similarity search between the shape integration information are performed by different search methods. 7. The simulation information generation means is set with a plurality of sets of target information for specifying the type of the goal for simulation and the purpose of action and information regarding the shape characteristics thereof, and simultaneously performs simulation for the plurality of simulation targets. The target identifying device according to claim 1, wherein the target identifying devices are executed in parallel. 8. Hierarchical information in which the concept expressing the type of goal and the action purpose is hierarchized and the goal information is one node is provided, and the identification means is the case information extracted by the simulated case information and the case information storage means. Between the target shape information corresponding to the similar shape information in the hierarchical information, and the similar shape information extraction unit that extracts the shape integrated information whose similarity to the shape integrated information of the target actual measurement information is a predetermined value or more as the similar shape information. In the shape information estimating unit that extracts the target information corresponding to the type of target and / or the action purpose of the common concept as the shape estimation result, and the case information extracted by the case information accumulating unit,
A similar behavior information extraction unit that extracts time-series behavior information whose similarity to the time-series behavior information of the target actual measurement information is a predetermined value or more as similar time-series behavior information, and the similar time-series behavior information in the hierarchical information. A case system estimation unit that extracts, as a case system estimation result, target information corresponding to the type of target concept and / or action purpose common to corresponding target information, and the shape system estimation result and the case in the hierarchical information. Type of target of superordinate concept common with system estimation result and /
The target identifying apparatus according to claim 1, further comprising an identification processing unit that selects target information corresponding to an action purpose as an identification result. 9. The identification means generates a confirmation result reflecting the confirmation and / or correction of the identification result, and the case information accumulating means associates the confirmation result with the case information. The target identification device according to claim 1, wherein the target identification device stores the information. 10. The identifying means executes the identifying process using the target information if the target information corresponding to the confirmation result is included in the case information, and the identification result if there is no target information corresponding to the confirmation result. 10. The target identifying device according to claim 9, wherein the identifying process is executed by using the target information corresponding to. 11. Time-series action information that defines the action of the target from observation data of specifications relating to the action and shape of the target that are input in time series, and shape integration information that defines time-series changes of the shape characteristics of the target. And a target measurement information corresponding to the target measurement information generation step that generates target measurement information including The case information extraction step of extracting case information in which the similarity between the target measurement information and the target actual measurement information is equal to or greater than a predetermined value among the case information consisting of Using the target information for specifying the action purpose and the information on its shape characteristics, the time-series action information of the target to be identified and the shape integrated information obtained by simulating the action of the above-mentioned simulated target according to the information The shape integration of the target actual measurement information from the simulation information generation step of generating simulation case information including the target information for simulation and the case information extracted in the simulation case information and the case information extraction step. An identification step of selecting information having a similarity between the information and the time-series behavior information equal to or greater than a predetermined value, and setting target information corresponding to the information as an identification result, and confirming and / or correcting the identification result. And a case storing step of storing a check result reflecting the generated check result in association with the case information. 12. Time-series action information that defines the action of the target from observation data of specifications relating to the action and shape of the target that are input in time series, and shape integration information that defines time-series change of the shape characteristic of the target. Measured information generating means for generating target measured information including and accumulating case information including target measured information corresponding to the identification result of the target information specifying the type of goal and action purpose, and Of case information accumulated in past observations, case information storage means that extracts case information in which the similarity between the target measurement information and the target to be identified exceeds a predetermined value, the type of goal for simulation, and the purpose of action are specified The target information for simulation and the information regarding its shape characteristics are preset, and the time-series behavior information of the target to be identified and the shape obtained by simulating the behavior of the above-mentioned target for simulation Shape integration of the target actual measurement information from the simulated information generation means for generating the simulated case information including the integrated information and the simulated target information, the case information extracted by the simulated case information and the case information storage means. A program that causes a computer to function as an identification unit that selects information whose similarity between information and time-series behavior information is a predetermined value or more and outputs target information corresponding to the information as an identification result.