JP2006202046A - Control instruction creation apparatus and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a control instruction creation apparatus or the like capable of obtaining proper control instructions even without presetting creation requirements. <P>SOLUTION: In a learning phase, a process control part 25 obtains control instruction data indicating previously transmitted control instructions, and corresponding track data and flight environment data, from a control instruction DB 23, a track DB 21, and a flight environment DB 22, respectively. An index calculating part 26 calculates index values on the basis of the track data and the flight environment data obtained, and stores the index values and the control instruction data in an index DB 24 in association with each other. A model creating part 27 reads the associated index values and flight instruction data from the index DB 24 and creates a statistical model using learning data for which the index values and the flight instruction data are inputted and outputted, respectively. In an inference phase, the index calculating part 26 calculates index values on the basis of the track data and the flight environment data to be determined. A control inference part 28 inputs the index values to the statistical model to obtain output data. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a control instruction generating device for generating a control instruction for air traffic control.

Conventionally, air route control work is performed by the Air Traffic Control Department (ACC). The controller in the Air Traffic Control Department will fly based on various information such as flight information such as the position, speed, altitude, etc. of the aircraft in flight and flight environment information such as wind direction, wind speed, temperature, etc. at each altitude. The necessary control instructions were judged and sent to the aircraft inside. The operation of judging the control instructions is very complicated and requires a high degree of accuracy, which is a heavy burden on the controller.
In order to reduce such a burden, development of a system that automatically generates a control instruction using a computer is also underway. For example, a predetermined control instruction is automatically generated by setting a generation condition of the control instruction in advance. A system that can be generated has also been devised (Patent Document 1).
Japanese Patent Laid-Open No. 11-14745 (page 7, FIG. 1)

  However, in the above-described system, only a control instruction in which generation conditions are set in advance can be generated. Further, in the above-described system, the generation target control instructions are limited to those whose generation conditions are relatively simple (for example, those that can be determined by one parameter), and are determined based on a plurality of factors. Such control instructions could not be automatically generated.

The present invention has been made in view of the above circumstances, and an object thereof is to provide a control instruction generation device and the like that can acquire an appropriate control instruction without setting a generation condition in advance.
Another object of the present invention is to provide a control instruction generating device or the like that can acquire a control instruction based on more complicated conditions.

In order to achieve the above object, a control instruction generation device according to the first aspect of the present invention provides:
Means for acquiring control instruction data indicating control instructions transmitted in the past, wake data relating to past wakes and aviation environment data relating to aviation environment corresponding to the control instruction data;
First index value acquisition means for obtaining an index value based on the acquired track data and aviation environment data;
Means for associating the index value obtained by the first index value obtaining means with the obtained control instruction data in a database;
Means for reading an associated index value and aviation instruction data from the database, and generating a statistical model using learning data having the index value as input data and the aviation instruction data as output data;
Means for acquiring track data and aviation environment data to be judged;
Second index value acquisition means for obtaining an index value based on the track data to be determined and the aviation environment data;
An output means for inputting the index value obtained by the second index value acquisition means to the statistical model, acquiring and outputting output data for the input;
It is characterized by providing.

The output data by the statistical model may indicate an estimation result regarding the appropriate rate of each control instruction,
The output means may specify predetermined control instruction data based on an appropriate rate of each control instruction indicated by the acquired output data, and perform output related to the specified control instruction data.

  Means for adjusting the statistical model by inputting the index value obtained by the second index value obtaining means and the output data obtained by the output means to the statistical model as learning data; You may prepare.

A program according to the second aspect of the present invention is
Computer
Means for obtaining control instruction data indicating control instructions transmitted in the past, wake data relating to past wakes and aviation environment data relating to aviation environment corresponding to the control instruction data;
First index value acquisition means for determining an index value based on the acquired track data and aviation environment data;
Means for associating the index value obtained by the first index value acquiring means with the acquired control instruction data in a database;
Means for reading an associated index value and aviation instruction data from the database, and generating a statistical model using learning data having the index value as input data and the aviation instruction data as output data;
Means for obtaining track data and aviation environment data to be judged,
Second index value obtaining means for obtaining an index value based on the track data to be determined and the aviation environment data;
An output means for inputting the index value obtained by the second index value acquiring means to the statistical model, and acquiring and outputting output data for the input;
To function as.

  According to the present invention, an appropriate control instruction can be acquired without setting a generation condition in advance.

Hereinafter, a control instruction generation system according to an embodiment of the present invention will be described with reference to the drawings.
An example of the physical configuration of the control instruction generation system 1 is shown in FIG. As illustrated, the control instruction generation system 1 includes a control unit 11, a main storage unit 12, an external storage unit 13, an input unit 14, a display unit 15, and a communication unit 16.
The control unit 11 is configured by a CPU (Central Processing Unit) or the like, reads a program stored in the external storage unit 13 and data necessary for processing, and executes processing to be described later.
The main storage unit 12 includes a RAM (Random Access Memory) or the like, and is used as a work area for the control unit 11.
The external storage unit 13 includes a hard disk device and the like, and stores programs executed by the control unit 11 and various data. The external storage unit 13 includes a wake DB (database), an aviation environment DB, a control instruction DB, and an index DB, which will be described later.
The input unit 14 includes an input device such as a keyboard and a mouse, and supplies input data to the control unit 11.
The display unit 15 includes a display device, and displays data related to the control instruction in accordance with an instruction from the control unit 11.
The communication unit 16 communicates with other devices according to instructions from the control unit 11.

  An example of the logical configuration of the control instruction generation system 1 is shown in FIG. As shown in the figure, the control instruction generation system 1 includes a wake DB 21, an aviation environment DB 22, a control instruction DB 23, an index DB 24, a processing control unit 25, an index calculation unit 26, a model generation unit 27, A control reasoning unit 28. Among these, the index calculation unit 26, the model generation unit 27, and the control reasoning unit 28 are realized by the control unit 11 executing a program stored in the external storage unit 13.

  The wake DB 21 stores data (wake data) relating to the wake of each aircraft acquired in the past. The wake data is acquired every predetermined time (for example, every minute) for the aircraft in flight. For example, as shown in FIG. 3, the data acquisition time, the aircraft type, the X coordinate, the Y coordinate, Includes data such as altitude, speed, heading, destination airport, and aircraft type.

  The aviation environment DB 22 stores data related to the aviation environment (aviation environment data). The aviation environment data includes first aviation environment data and second aviation environment data. For example, as shown in FIG. 4A, the first aviation environment data includes data such as wind direction, wind speed, and temperature at each altitude. The second aviation environment data relates to the aircraft capability and the like, and includes data such as the turning rate, the rising rate, and the cruise speed of each aircraft type as shown in FIG. 4B, for example.

  In the control instruction DB 23, data (control instruction data) indicating a control instruction actually transmitted in the past is stored in association with data of an instruction time when the control instruction is transmitted to each aircraft.

  In the index DB 24, index value data relating to air traffic and control instruction data are stored in association with each other.

  The various data stored in the wake DB 21, the aviation environment DB 22, and the control instruction DB 23 are received by the control unit 11 from other devices via the communication unit 16 and stored in the corresponding DBs, for example. Alternatively, the control unit 11 may read various data from a recording medium in which these data are recorded via a recording medium reading device (not shown) and store them in the corresponding DBs.

The process control unit 25 controls the entire system. The processing executed by this system mainly consists of a learning phase and an inference phase.
In the learning phase, the processing control unit 25 extracts the track data, the aviation environment data, and the control instruction data that are in a correspondence relationship from the wake DB 21, the aviation environment DB 22, and the control instruction DB 23, and passes the extracted data to the index calculation unit 26. At the same time, the index calculation unit 26 is requested to calculate the index value and associate the index value with the control instruction data and store them in the index DB 24. Further, the processing control unit 25 requests the model generation unit 27 to generate a statistical model (knowledge data). Note that the extraction method of the wake data, the aviation environment data, and the control instruction data in the correspondence relation is arbitrary. For example, with respect to certain control instruction data, the instruction time data associated therewith is acquired from the control instruction DB 23, and the time before the instruction time and closest to the instruction time is set as the “acquisition time”. The wake data may be searched from the wake DB 21 and the corresponding wake data may be specified as the wake data corresponding to the control instruction data to be processed. The wake data and the first aviation environment data are associated with each other by a data item such as “altitude”, and the wake data and the second aviation environment data are associated with each other by a data item such as “aircraft model”. Therefore, based on these data items, the aviation environment data corresponding to the track data specified earlier may be specified.
In the inference phase, the process control unit 25 passes the track data to be determined and the aviation environment data to the index calculation unit 26 and requests calculation of the index value. Further, the processing control unit 25 passes the index value based on the track data to be determined and the aviation environment data calculated by the index calculation unit 26 to the control reasoning unit 28 and requests the control instruction estimation.

The index calculation unit 26 calculates an index value related to air traffic from the track data and the aviation environment data received from the processing control unit 25 in response to the calculation and registration request of the index value from the processing control unit 25. As the calculation formula of the index value, a mathematical formula using the value of one or more data items in at least one of the wake data and the aviation environment data as a variable is set in advance for each index value. The content of the calculated index value is arbitrary. For example, “Elapsed time since last control instruction”, “Agility of aircraft performance”, “Wind intensity with respect to traveling direction”, “Margin time for collision avoidance” Good. Then, the index calculation unit 26 associates the calculated index value with the control instruction data received from the processing control unit 25 (that is, control instruction data corresponding to the track data and the aviation environment data used for calculating the index value). The index calculation unit 26 calculates an index value from the determination target track data and the aviation environment data received from the processing control unit 25 in response to an index calculation request from the processing control unit 25. .

  In response to the statistical model generation request from the processing control unit 25, the model generation unit 27 reads the associated index value data and control instruction data from the index DB 24, and uses the index value data as input data. Is generated as learning data, the learning data is input to a predetermined statistical model, and learning processing for learning the statistical causal relationship between the input data and the output data is executed. This statistical model outputs statistical data related to control instructions (how appropriate each control instruction is) (appropriate rate of each control instruction) for each index value input, for example, fuzzy It may be a collection of functions or a learning network model. For the generation of such a statistical model, for example, a known technique such as β-RNA of Add-in Laboratory Co., Ltd. may be used. FIG. 5A shows an example of learning data input to the statistical model in the learning phase, and FIG. 5B shows an example of input / output data of the statistical model in the inference phase. In the learning phase, for example, learning data as shown in FIG. 5A is input to the statistical model to learn the relationship between input and output. In the inference phase, input data as shown in FIG. Is input to the statistical model, output data (inference result) for this input data is output from the statistical model. The estimation result data includes, for example, control instruction data and appropriate rate data as shown in FIG.

  In response to the index value calculation request from the process control unit 25, the control reasoning unit 28 inputs each index value received from the process control unit 25 to the statistical model generated by the model generation unit 27, and this input Output data is acquired, and the acquired output data is output to the display unit 15 or the like. Note that the method for inputting the track data and the aviation environment data to be determined to the system is arbitrary. For example, the processing control unit 25 uses a recording medium reading device (not shown) for the determination target data recorded on the recording medium. Or may be received from another computer.

Next, the processing operation of the control instruction generation system 1 according to the present embodiment will be described. As described above, the main processing by the control instruction generation system 1 includes a learning phase and an inference phase.
First, the processing operation in the learning phase will be described with reference to FIG.
First, the process control unit 25 extracts the track data, the aviation environment data, and the control instruction data that are in a correspondence relationship from the wake DB 21, the aviation environment DB 22, and the control instruction DB 23, and extracts the extracted data together with the calculation of the index value and the registration request. It passes to the index calculation unit 26 (step S1).
In response to this, the index calculation unit 26 calculates the index value based on the track data and the aviation environment data received together with the calculation and registration request of the index value (step S2), and the calculated index value is used as the index value. The data is stored in the index DB 24 in association with the control instruction data received together with the calculation and registration request (step S3).

  Next, the process control unit 25 determines whether the above process (steps S1 to S3) has been completed for all the control instruction data registered in the control instruction DB 23 (step S4). If the processing has not been completed for all the control instruction data (step S4: NO), the process returns to step S1. Further, when the above processing is completed for all the control instruction data (step S4: YES), the processing control unit 25 requests the model generation unit 27 to generate a statistical model, and in response to this, the model The generation unit 27 reads the associated index value data and control instruction data stored in the index DB 24, generates learning data that uses the index value data as input data and the control instruction data as output data, It inputs into a statistical model and performs a learning process (step S5). Thereby, a statistical model adjusted by the learning process is generated. The generated statistical model data is stored in the external storage unit 13.

Next, the processing operation in the inference phase will be described with reference to FIG.
The processing control unit 25 acquires the track data and the aviation environment data to be determined by, for example, reading the data stored in the recording medium through a recording medium reader, and requesting calculation of the index value. At the same time, it is passed to the index calculator 26 (step S11).
The index calculation unit 26 calculates an index value based on the track data and the aviation environment data received together with the index value calculation request (step S12).
Next, when the processing control unit 25 requests the control reasoning unit 28 to estimate the control instruction based on the calculated index value, the control reasoning unit 28 responds to the request by the index value calculated in step S12. Then, the statistical model generated in the inference phase is input, and output data corresponding to this input is obtained (step S13). And the control reasoning part 28 displays output data (inference result) on the display part 15 (step S14). Thereby, the estimation result about the control instruction | indication based on track data or aviation environment data is acquirable. For example, the controller can see the estimation result, select a control instruction with the highest appropriate rate, and transmit the control instruction to the aircraft.

  As described above, according to the present invention, the system learns the relationship between the control instruction actually made in the past and the data (wake data and aviation environment data) of the situation in which the control instruction is performed. Since a correct control instruction is estimated based on the result, the control instruction can be estimated without setting a generation condition for each control instruction as in the past. In addition, since a plurality of indicators are used as the determination elements of the control instruction, it is possible to acquire not only the control instruction obtained by a simple determination but also a control instruction based on a complicated determination taking various situations into consideration.

The present invention can be variously modified and applied.
For example, in the inference phase, the control instruction data corresponding to the predetermined condition based on the appropriate rate of each control instruction indicated by the output data of the statistical model may be specified, and the specified control instruction data may be displayed. For example, the control instruction having an appropriate rate of more than a predetermined value (for example, 75%) may be extracted, or the control instruction data indicated by the inference result having the highest appropriate rate may be extracted. May be. Moreover, you may perform the process for transmitting the one control instruction data extracted from the inference result with respect to an aircraft. In this case, for example, when a message prompting confirmation of transmission is displayed on the display unit and an input indicating confirmation is made, a control including the aircraft type, control instruction data, etc. is sent to the computer that transmits the control instruction to the aircraft. An instruction transmission request may be transmitted.

  In addition, in the inference phase, the index value input to the statistical model and the inference result data output from the statistical model are fed back to the statistical model as learning data and the learning process is executed. You may make it adjust further. This feedback may be performed every time the inference result data for the determination target track data and the aeronautical environment data is acquired, and the learning data (the determination target data and the inference result data for this) is stored in the external storage unit 13. It may be accumulated and stored and performed every predetermined period.

The system of the present invention can be realized using a normal computer system, not a dedicated system. For example, a program for executing the above operation is stored in a computer-readable recording medium (FD, CD-ROM, DVD, etc.) and distributed, and the program is installed in the computer to execute the above processing. A control instruction generation system may be configured. Alternatively, it may be stored in a disk device of a server device on a network such as the Internet and downloaded to a computer to which the audio input device and the audio output device are connected.
In addition, when the OS realizes the above functions by sharing the OS or jointly with the OS and the application, etc., only the part other than the OS may be stored and distributed in the medium, or may be downloaded to the computer. Good.

It is a figure which illustrates the physical structure of the control instruction | indication production | generation system which concerns on embodiment of this invention. It is a figure which illustrates the logical structure of the control instruction | indication production | generation system of FIG. It is a figure which shows an example of wake data. It is a figure which shows an example of aviation environment data. It is a figure for demonstrating specifically the input / output data of a statistical model. It is a flowchart for demonstrating the processing operation in the learning phase of the control instruction | indication production | generation system of FIG. It is a flowchart for demonstrating the processing operation in the inference phase of the control instruction | indication production | generation system of FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Control instruction generation system 11 Control part 12 Main memory part 13 External storage part 14 Input part 15 Display part 16 Communication part 21 Track DB
22 Aviation Environment DB
23 Control instruction DB
24 Indicator DB
25 Processing Control Unit 26 Index Calculation Unit 27 Model Generation Unit 28 Control Inference Unit

Claims (4)

Means for acquiring control instruction data indicating control instructions transmitted in the past, wake data relating to past wakes and aviation environment data relating to aviation environment corresponding to the control instruction data;
First index value acquisition means for obtaining an index value based on the acquired track data and aviation environment data;
Means for associating the index value obtained by the first index value obtaining means with the obtained control instruction data in a database;
Means for reading an associated index value and aviation instruction data from the database, and generating a statistical model using learning data having the index value as input data and the aviation instruction data as output data;
Means for acquiring track data and aviation environment data to be judged;
Second index value acquisition means for obtaining an index value based on the track data to be determined and the aviation environment data;
An output means for inputting the index value obtained by the second index value acquisition means to the statistical model, acquiring and outputting output data for the input;
A control instruction generating device comprising: The output data by the statistical model shows the estimation result regarding the appropriate rate of each control instruction,
The output means specifies predetermined control instruction data based on an appropriate rate of each control instruction indicated by the acquired output data, and performs output related to the specified control instruction data.
The control instruction generation device according to claim 1, wherein Means for adjusting the statistical model by inputting the index value obtained by the second index value obtaining means and the output data obtained by the output means to the statistical model as learning data; Prepare
The control instruction generation device according to claim 1 or 2, wherein Computer
Means for obtaining control instruction data indicating control instructions transmitted in the past, wake data relating to past wakes and aviation environment data relating to aviation environment corresponding to the control instruction data;
First index value acquisition means for determining an index value based on the acquired track data and aviation environment data;
Means for associating the index value obtained by the first index value acquiring means with the acquired control instruction data in a database;
Means for reading an associated index value and aviation instruction data from the database, and generating a statistical model using learning data having the index value as input data and the aviation instruction data as output data;
Means for obtaining track data and aviation environment data to be judged,
Second index value obtaining means for obtaining an index value based on the track data to be determined and the aviation environment data;
An output means for inputting the index value obtained by the second index value acquiring means to the statistical model, and acquiring and outputting output data for the input;
Program to function as.

Images