A text instruction generation method and device for air traffic control simulation training

Abstract

The present invention discloses a text instruction generation method and device for air traffic control simulation training, which belongs to the technical field of civil aviation control simulation training, comprising: inputting the instruction text of the controller of the air traffic control dialogue, and preprocessing the instruction text; inputting the preprocessed instruction text into an air traffic control instruction parsing module, extracting the semantic label sequence and intention of the preprocessed instruction text; classifying the control instructions in the preprocessed instruction text according to the intention; based on an instruction repetition generation module, generating corresponding instruction parameters and instruction repetition text for operating simulated flight from the semantic label sequence and the classified control instructions, thereby realizing automatic generation of dialogues and instructions in the control simulation training process, reducing training costs, and improving the efficiency of the control simulation training.

Description

Technical Field

[0001] The present invention relates to the technical field of civil air traffic control simulation training, and in particular to a method and device for generating text instructions for air traffic control simulation training. Background Art

[0002] In civil aviation control scenarios, air traffic controllers communicate with pilots via radio to direct aircraft. The air traffic controller issues control instructions, and the captain repeats them back to confirm their accuracy before maneuvering the aircraft accordingly. Similarly, in air traffic controller simulation training, a pilot role is required to conduct conversations and execute control instructions, simulating real-world air traffic control scenarios. Existing air traffic control simulation training systems rely on humans to act as the pilot role, manually inputting control instructions to control the aircraft in the flight simulator. This results in high air traffic controller training costs, and manual input of control instructions is inefficient and prone to errors.

[0003] Therefore, a method and device for generating text instructions for air traffic control simulation training are needed to solve the problems of manual dialogue and manual input of instructions, thereby improving the efficiency of air traffic control simulation training. Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the existing technology that the control simulation training system relies on manual means to act as the captain and manually input control instructions to control the flight of the aircraft in the simulated flight software during the training process, resulting in high costs for controller training, and the manual input of control instructions is inefficient and prone to errors. A text instruction generation method and device for air traffic control simulation training are provided.

[0005] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:

[0006] A method for generating text instructions for air traffic control simulation training comprises the following steps:

[0007] S1: inputting the command text of the air traffic controller in the air traffic control dialogue and preprocessing the command text;

[0008] S2: Inputting the pre-processed instruction text into the air traffic control instruction parsing module to extract the semantic label sequence and intent of the pre-processed instruction text;

[0009] The air traffic control instruction parsing module includes: a word segmentation model, an intent recognition model, and an entity naming model;

[0010] The word segmentation model consists of a bidirectional LSTM layer and a CRF layer, the bidirectional LSTM layer is used to extract features, and the CRF layer is used to output word segmentation labels;

[0011] The intent recognition model consists of a bidirectional LSTM layer and two fully connected layers. The mean calculation of the time dimension of the text sequence is added between the two fully connected layers to unify the feature dimension.

[0012] The entity naming model consists of a bidirectional LSTM layer and a CRF layer, wherein the bidirectional LSTM layer is used to extract features and the CRF layer is used to output semantic labels;

[0013] S3: classifying the control instructions in the preprocessed instruction text according to the intention;

[0014] S4: Based on the instruction repetition generation module, the semantic label sequence and the classified control instructions are respectively generated into corresponding instruction parameters and instruction repetition texts for operating simulated flight.

[0015] By adopting the above technical solution, the input instruction text is analyzed for intent and annotated with word segmentation tags through the air traffic control instruction parsing module, and then different types of control instructions are processed according to the air traffic control repetition rules and air traffic control simulation training instruction rules to generate corresponding intent parameters and repetition instruction texts. At the same time, the special situation handling module is used to output the captain's text instructions when the aircraft encounters special circumstances, thereby realizing dialogue generation and aircraft flight control, and realizing the automatic generation of dialogues and instructions during the control simulation training process, reducing training costs, improving the efficiency of control simulation training, and also improving the richness of control simulation training.

[0016] As a preferred solution of the present invention, the step S1 of preprocessing the instruction text includes the following steps:

[0017] S11: removing punctuation marks and special characters from the instruction text, and filtering meaningless words from the instruction text, wherein the meaningless words include modal particles and repeated words;

[0018] S12: Merging the specialized vocabulary for the combination types appearing in the instruction text.

[0019] As a preferred embodiment of the present invention, step S2 includes:

[0020] S21: Input the pre-processed instruction text and first classify it into Chinese and English;

[0021] S22: Inputting Chinese characters into the word segmentation model to perform Chinese instruction word segmentation and outputting word segmentation labels;

[0022] S23: comparing the phrases in the segmentation label with an air traffic control special vocabulary, correcting any erroneous special vocabulary segmentations in the segmentation label to obtain a corrected segmentation label, wherein the data in the air traffic control special vocabulary includes at least location, instruction, airline name, and place name;

[0023] S24: Input the English text or the modified word segmentation label into the intent recognition model, output a probability vector of all intent categories, the length of the vector being the same as the number of intent categories, the value of each position in the vector being a predicted probability value of containing the intent, and at the same time, set a threshold, filter the predicted probability values, filter out predicted probability values with a credibility higher than the threshold, check whether the keywords corresponding to the predicted probability values are contained in the preprocessed instruction text, and if so, obtain the final intent;

[0024] S25: Input the English text or the corrected word segmentation label into the entity naming model, and output the semantic label sequence corresponding to the preprocessed instruction text. At the same time, for the instruction text containing the correction intention in the final intention, remove the instruction semantics that need to be corrected in its semantic label sequence, and only retain the corrected instruction semantics as the final semantic label sequence.

[0025] As a preferred solution of the present invention, the step S2 further includes: performing corresponding post-processing on the intent and the semantic tag sequence, replacing fixed professional vocabulary with English abbreviations, and replacing Chinese and English numbers with Arabic numerals.

[0026] As a preferred embodiment of the present invention, the intention includes: declarative control intention, control identification intention, and inquiry control intention.

[0027] As a preferred solution of the present invention, in step S3, the control instructions are divided into: declarative control instructions, control identification instructions, and inquiry instructions;

[0028] The declarative control instruction is a traffic intervention instruction issued to the pilot through a declarative statement, and the instruction format is AB type with the flight number plus the instruction;

[0029] The control identification instruction is the controller's instruction after the aircraft actively contacts the controller for the first time when entering the control sector;

[0030] The inquiry instructions are divided into: repeated confirmation instructions and negotiation instructions. The repeated instructions are the controller's repeated confirmation of the pilot's instruction request, and the negotiation instructions are to inquire the pilot about the feasibility of issuing control instructions.

[0031] As a preferred embodiment of the present invention, step S4 includes: receiving the declarative control instruction; adjusting and supplementing the flight number and the instruction in sequence according to the air traffic control readback rule, converting the declarative control instruction into a BA-type instruction consisting of the instruction plus the flight number, wherein the BA-type instruction is the readback instruction text of the declarative control instruction; extracting declarative control intention parameters from the declarative control intent corresponding to the declarative control instruction; and converting the declarative control intent and the declarative control intention parameters into instruction parameters recognizable by the simulator;

[0032] receiving the control identification instruction, having the pilot repeat the aircraft call sign according to the air traffic control readback rule, parsing the semantic tag sequence corresponding to the control identification instruction, extracting the aircraft call sign, extracting control identification intent parameters from the control identification intent corresponding to the control identification instruction, and converting the control identification intent and the control identification intent parameters into instruction parameters recognizable by the simulator;

[0033] Receive the inquiry-type instruction, determine the type of the inquiry-type instruction; if it is determined to be a re-confirmation-type instruction, then according to the air traffic control repetition rule, the pilot repeats the statement instruction, segment the aircraft call sign and other instructions in the re-confirmation-type instruction according to the corresponding semantic tag sequence, reorder them, and end with the aircraft call sign to obtain the instruction repetition text of the inquiry-type instruction; if it is determined to be a negotiation instruction, then according to the air traffic control repetition rule, extract the inquiry control intention parameters from the inquiry control intention corresponding to the negotiation instruction, the captain judges the inquiry control intention and the inquiry control intention parameters according to the flight situation, and repeats the instruction if it is legal, and at the same time converts the control identification intention and the control identification intention parameters into instruction parameters that can be recognized by the simulator; if it is illegal, start the next round of negotiation.

[0034] As a preferred solution of the present invention, step S1 further includes: constructing a special service handling module to handle special situation scenarios encountered during the flight, and generate and output text instructions from the captain.

[0035] As a preferred solution of the present invention, the special situation scenarios in step S1 include: aircraft failure, crew emergency, and external environmental interference.

[0036] On the other hand, a text instruction generation device for air traffic control simulation training is provided, comprising at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any of the methods described above.

[0037] Compared with the prior art, the beneficial effects of the present invention are: the input instruction text is analyzed for intent and labeled with word segmentation tags through the air traffic control instruction parsing module, and then different types of control instructions are processed according to the air traffic control repetition rules and air traffic control simulation training instruction rules to generate corresponding intent parameters and repetition instruction texts. At the same time, the special situation handling module is used to output the captain's text instructions when the aircraft encounters special circumstances, which improves the accuracy of instruction text parsing and has strong applicability, making the instruction text parsing more intelligent, thereby realizing dialogue generation and aircraft flight control, and realizing the automatic generation of dialogues and instructions during control simulation training, reducing training costs, improving the efficiency of control simulation training, and also improving the richness of control simulation training. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is a flowchart of a method for generating text instructions for air traffic control simulation training according to Example 1 of the present invention;

[0039] Figure 2 This is a flowchart of instruction parsing for a method for generating text instructions for air traffic control simulation training according to Example 1 of the present invention;

[0040] Figure 3 This is a structural diagram of a word segmentation model of a method for generating text instructions for air traffic control simulation training according to Example 1 of the present invention;

[0041] Figure 4 This is the intent prediction and word segmentation label diagram of the method for generating text instructions for air traffic control simulation training described in Example 1 of the present invention;

[0042] Figure 5 This is a structural diagram of a text instruction generation device for air traffic control simulation training according to embodiment 2 of the present invention. DETAILED DESCRIPTION

[0043] The present invention will be further described in detail below in conjunction with test examples and specific embodiments. However, this should not be understood as limiting the scope of the present invention to the following embodiments, and all technologies implemented based on the present invention fall within the scope of the present invention.

[0044] Example 1

[0045] A method for generating text instructions for air traffic control simulation training, such as Figure 1 As shown, the following steps are included:

[0046] S1: inputting the command text of the air traffic controller in the air traffic control dialogue and preprocessing the command text;

[0047] The step S1 of pre-processing the instruction text comprises the following steps:

[0048] S11: removing punctuation marks and special characters from the instruction text, and filtering meaningless words from the instruction text, wherein the meaningless words include modal particles and repeated words;

[0049] Specifically, punctuation marks and special characters in the instruction text are removed, as shown in Table 1:

[0050]

[0051]

[0052] Table 1

[0053] Specifically, the meaningless words in the instruction text are filtered, including modal particles and repeated words, to avoid negative interference with subsequent semantic analysis, as shown in Table 2:

[0054]

[0055] Table 2

[0056] S12: Merging the combination-type proprietary words appearing in the instruction text, and merging the combination-type proprietary words appearing in the water sentence, such as English airline names;

[0057] Step S1 further includes: constructing a special service handling module to handle special situation scenarios encountered during flight, generating and outputting a captain's text instruction, wherein the captain's text instruction is in the form of an aircraft call sign plus a special situation status, and can include captain-requested operations such as requesting an emergency landing, requesting a fly-around, and altitude adjustment. The special situation scenarios include aircraft failure, crew emergency, and external environmental interference;

[0058] Specifically, aircraft failure refers to a mechanical failure of the aircraft itself, such as failure or damage to the engine or other aircraft equipment. In this scenario, the captain proactively reports the situation to the controller, requests certain operations, or waits for instructions from the controller, as shown in Table 3:

[0059]

[0060]

[0061] Table 3

[0062] Crew emergency: The crew encounters an abnormal situation, such as personnel injury, hijacking, attack, or loss of direction. In this scenario, the captain proactively reports the situation to the controller, requests certain operations, or waits for the controller's instructions, as shown in Table 4:

[0063]

[0064] Table 4

[0065] External environmental interference: When encountering extreme weather conditions, such as cumulonimbus clouds, thunderstorms, hail, or other environmental factors that affect flight, the captain proactively reports the situation to the controller, requests certain maneuvers, or waits for instructions from the controller, as shown in Table 5:

[0066]

[0067] Table 5

[0068] S2: Inputting the pre-processed instruction text into the air traffic control instruction parsing module to extract the semantic label sequence and intent of the pre-processed instruction text;

[0069] like Figure 2 As shown, step S2 includes:

[0070] S21: Input the pre-processed instruction text, first classify it into Chinese and English, and detect whether it contains Chinese characters. If it does, it is Chinese; if not, it is English;

[0071] S22: Figure 3 As shown, Chinese characters are input into the word segmentation model for Chinese instruction word segmentation. The word segmentation model consists of a bidirectional LSTM layer and a CRF layer. The bidirectional LSTM layer is used to extract features, and the CRF layer is used to output word segmentation labels.

[0072] S23: Comparing the phrases in the segmentation label with the air traffic control special vocabulary, correcting any erroneous special vocabulary in the segmentation label to obtain a corrected segmentation label to ensure accuracy of the segmentation. The data in the air traffic control special vocabulary includes a vocabulary consisting of fixed phrases such as location, instruction, airline name, and place name;

[0073] S24: Build an intent recognition model, such as Figure 4 As shown, the intent recognition model consists of a bidirectional LSTM layer and two fully connected layers. The mean value calculation on the time dimension of the text sequence is added between the two fully connected layers to unify the feature dimension. English text or the corrected word segmentation label is input, and the probability vector of all intent classifications is output. The vector length is the same as the number of intent categories. The value of each position in the vector is a predicted probability value for judging whether the intent is contained. At the same time, a threshold is set to filter the predicted probability values, and the predicted probability values with a credibility higher than the threshold are filtered out. It is checked whether the keywords corresponding to the predicted probability values are contained in the preprocessed instruction text. If so, the final intent is obtained.

[0074] S25: Construct entity naming model, such as Figure 4As shown, the entity naming model consists of a bidirectional LSTM layer and a CRF layer. The English text or the corrected word segmentation tags are input, and the semantic tag sequence corresponding to each value in the preprocessed instruction text is output. At the same time, for the instruction text whose final intention contains a correction intention, the instruction semantics to be corrected in its semantic tag sequence are removed, and only the corrected instruction semantics are retained as the final semantic tag sequence.

[0075] The step S2 further includes: performing corresponding post-processing on the intention and the semantic tag sequence, replacing fixed professional vocabulary with English abbreviations, and replacing Chinese and English numbers with Arabic numerals. For example, the flight number "Dongfang 2307" is replaced with "CES2307"; the numerical values in the instruction are normalized to aviation standard values. For example, the flight altitude "10700" becomes the standard value "10700".

[0076] The intention includes: declarative control intention, control identification intention, and inquiry control intention.

[0077] S3: Classify the control instructions in the preprocessed instruction text according to the intention.

[0078] In the step S3, the control instructions are divided into: declarative control instructions, control identification instructions, and inquiry instructions.

[0079] The declarative control instruction is a traffic intervention instruction issued to the pilot through a declarative statement. The instruction form is AB type of flight number plus instruction, and its intention contains a declarative control intention.

[0080] Specifically, as shown in Table 6:

[0081] Control order: Sichuan Babaguai rises to Yaodongguai [[ID=2,2]]

[0082] Table 6

[0083] The control identification instruction is the control instruction of the controller after the aircraft takes the initiative to contact the controller for the first time when entering the control sector, and its intention contains a control identification intention.

[0084] Specifically, as shown in Table 7:

[0085] Control order: Dongfang Liangsandong Turn Chengdu radar has been identified

[0086] Table 7

[0087] The inquiry type instructions are divided into: whether the intention contains the query control intention to determine whether it is an inquiry type instruction; if it does, it is an inquiry type instruction; and then, based on the keywords in the instruction text, the inquiry type instructions are divided into reconfirmation type instructions and negotiation type instructions. The reconfirmation type instruction is the controller's reconfirmation of the pilot's instruction request, and the instruction contains the keyword "confirm";

[0088] S4: Based on the instruction recitation generation module, the semantic label sequence and the classified control instructions are respectively generated into corresponding instruction parameters and instruction recitation texts for operating the simulated flight;

[0089] Specifically, the command type is derived by analyzing a large amount of real air traffic control data, and a system configuration file is provided for adding or modifying the command type. It has the characteristics of comprehensive command processing and flexible and configurable system.

[0090] The step S4 includes: receiving the declarative control instruction, adjusting and supplementing the flight number and the instruction according to the air traffic control readback rule, converting the declarative control instruction into a BA-type instruction consisting of the instruction plus the flight number, the BA-type instruction being the readback instruction text of the declarative control instruction, extracting declarative control intention parameters from the declarative control intent corresponding to the declarative control instruction, and converting the declarative control intent and the declarative control intention parameters into instruction parameters recognizable by the simulator;

[0091] Specifically, the instruction in Table 6 is input into the air traffic control instruction parsing module to obtain the semantic parsing result of the instruction, including: word segmentation label, declarative control intent and declarative control intent parameters, as shown in Table 8:

[0092]

[0093] Table 8

[0094] Based on the semantic analysis results in Table 8 and the air traffic control replay rules, the command replay generation module adjusts the order of the flight number and the key elements of the command and completes them. The command format is changed to a BA type command with the command plus the flight number, as shown in Table 9:

[0095] Repeat the command: rise to Yaodong turn Sichuan 88 turn 8

[0096] Table 9

[0097] At the same time, the conventional declarative control intention and the conventional declarative control intention parameters are used to operate the specified aircraft to perform the command operation. According to the air traffic control simulation training command rules, the conventional declarative control intention and the conventional declarative control intention parameters in the command instance are converted as shown in Table 10:

[0098] ACID: CSC8878; CMD: LVL1070

[0099] Table 10

[0100] receiving the control identification instruction, having the pilot repeat the aircraft call sign according to the air traffic control readback rule, parsing the semantic tag sequence corresponding to the control identification instruction, extracting the aircraft call sign, extracting control identification intent parameters from the control identification intent corresponding to the control identification instruction, and converting the control identification intent and the control identification intent parameters into instruction parameters recognizable by the simulator;

[0101] The inquiry command is received and its type is determined. If it is a re-confirmation command, the pilot repeats the statement command according to the air traffic control readback rule. The aircraft call sign and other commands in the re-confirmation command are segmented according to the corresponding semantic tag sequence and reordered, ending with the aircraft call sign. The command readback text of the inquiry command is obtained, as shown in Table 11:

[0102] Controller: Is it confirmed that the turn at Dongfang Liangsandong is heading to Chengdu? The captain repeated: Fly to Chengdu East, turn two or three holes

[0103] Table 11

[0104] If the command is determined to be a negotiation command, the captain extracts the inquiry control intent parameters from the inquiry control intent corresponding to the negotiation command according to the ATC readback rules. The captain then determines the inquiry control intent and the inquiry control intent parameters based on the flight situation. If the command is legal, the captain reads back the command. If not, the next round of negotiation begins. The legality of the negotiation command readback text is shown in Table 12:

[0105] Controller: Can I turn east two or three holes and head towards 200? The captain repeated: Heading 200, east, turn 23.

[0106] Table 12

[0107] By adopting the above technical solution, the input instruction text is analyzed for intent and annotated with word segmentation tags through the air traffic control instruction parsing module, and then different types of control instructions are processed according to the air traffic control repetition rules and air traffic control simulation training instruction rules to generate corresponding intent parameters and repetition instruction texts. At the same time, the special situation handling module is used to output the captain's text instructions when the aircraft encounters special circumstances, which improves the accuracy of instruction text parsing and has strong applicability, making the instruction text parsing more intelligent, thereby realizing dialogue generation and aircraft flight control, and realizing the automatic generation of dialogues and instructions during control simulation training, reducing training costs, improving the efficiency of control simulation training, and also improving the richness of control simulation training.

[0108] Example 2

[0109] like Figure 5As shown, a device for generating text instructions for air traffic control simulation training includes at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method and apparatus for generating text instructions for air traffic control simulation training described in the aforementioned embodiment. The input and output interfaces may include a display, a keyboard, a mouse, and a USB interface for inputting and outputting data; the power supply is used to provide power to the electronic device.

[0110] Those skilled in the art will understand that all or part of the steps of implementing the above-mentioned method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above-mentioned method embodiment; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM), magnetic disks or optical disks, and other media that can store program codes.

[0111] When the above-mentioned integrated unit of the present invention is implemented in the form of a software functional unit and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0112] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for generating text instructions for air traffic control simulation training, characterized in that: The following steps are involved: S1: inputting the command text of the air traffic controller in the air traffic control dialogue and preprocessing the command text; S2: Inputting the pre-processed instruction text into the air traffic control instruction parsing module to extract the semantic label sequence and intent of the pre-processed instruction text; The air traffic control instruction parsing module includes: a word segmentation model, an intent recognition model, and an entity naming model; The word segmentation model consists of a bidirectional LSTM layer and a CRF layer, the bidirectional LSTM layer is used to extract features, and the CRF layer is used to output word segmentation labels; The intent recognition model consists of a bidirectional LSTM layer and two fully connected layers. The mean calculation of the time dimension of the text sequence is added between the two fully connected layers to unify the feature dimension. The entity naming model consists of a bidirectional LSTM layer and a CRF layer, wherein the bidirectional LSTM layer is used to extract features and the CRF layer is used to output semantic labels; S3: classifying the control instructions in the preprocessed instruction text according to the intention; S4: Based on the instruction recitation generation module, the semantic label sequence and the classified control instructions are respectively generated into corresponding instruction parameters and instruction recitation texts for operating the simulated flight; The step S2 comprises: S21: Input the pre-processed instruction text and first classify it into Chinese and English; S22: Inputting Chinese characters into the word segmentation model to perform Chinese instruction word segmentation and outputting word segmentation labels; S23: comparing the phrases in the segmentation label with an air traffic control special vocabulary, correcting any erroneous special vocabulary segmentations in the segmentation label to obtain a corrected segmentation label, wherein the data in the air traffic control special vocabulary includes at least location, instruction, airline name, and place name; S24: Input the English text or the modified word segmentation label into the intent recognition model, output a probability vector of all intent categories, the length of the vector being the same as the number of intent categories, the value of each position in the vector being a predicted probability value of containing the intent, and at the same time, set a threshold, filter the predicted probability values, filter out predicted probability values with a credibility higher than the threshold, check whether the keywords corresponding to the predicted probability values are contained in the preprocessed instruction text, and if so, obtain the final intent; S25: Input the English text or the corrected word segmentation label into the entity naming model, and output the semantic label sequence corresponding to the preprocessed instruction text. At the same time, for the instruction text containing the correction intention in the final intention, remove the instruction semantics that need to be corrected in its semantic label sequence, and only retain the corrected instruction semantics as the final semantic label sequence.

2. The method for generating text instructions for air traffic control simulation training according to claim 1, characterized in that: The step S1 of pre-processing the instruction text comprises the following steps: S11: removing punctuation marks and special characters from the instruction text, and filtering meaningless words from the instruction text, wherein the meaningless words include modal particles and repeated words; S12: Merging the combined proprietary words appearing in the instruction text.

3. The method for generating text instructions for air traffic control simulation training according to claim 1, characterized in that: The step S2 further includes: performing corresponding post-processing on the intent and the semantic tag sequence, replacing fixed professional vocabulary with English abbreviations, and replacing Chinese and English numbers with Arabic numerals.

4. The method for generating text instructions for air traffic control simulation training according to claim 3, characterized in that: The intentions include: declarative regulatory intentions, regulatory identification intentions, and inquiring regulatory intentions.

5. The method for generating text instructions for air traffic control simulation training according to claim 4, characterized in that: In step S3, the control instructions are divided into: declarative control instructions, control identification instructions, and inquiry instructions; The declarative control instruction is a traffic intervention instruction issued to the pilot through a declarative statement, and the instruction format is AB type with the flight number plus the instruction; The control identification instruction is the controller's instruction after the aircraft actively contacts the controller for the first time when entering the control sector; The inquiry instructions are divided into: repeated confirmation instructions and negotiation instructions. The repeated confirmation instructions are the controller's repeated confirmation of the pilot's instruction request, and the negotiation instructions are to inquire the pilot about the feasibility of issuing control instructions.

6. The method for generating text instructions for air traffic control simulation training according to claim 5, characterized in that: The step S4 includes: receiving the declarative control instruction, adjusting and supplementing the flight number and the instruction in sequence according to the air traffic control readback rule, converting the declarative control instruction into a BA-type instruction with the instruction plus the flight number, the BA-type instruction being the readback instruction text of the declarative control instruction, extracting declarative control intention parameters from the declarative control intent corresponding to the declarative control instruction, and converting the declarative control intent and the declarative control intention parameters into instruction parameters recognizable by the simulator; receiving the control identification instruction, having the pilot repeat the aircraft call sign according to the air traffic control readback rule, parsing the semantic tag sequence corresponding to the control identification instruction, extracting the aircraft call sign, extracting control identification intent parameters from the control identification intent corresponding to the control identification instruction, and converting the control identification intent and the control identification intent parameters into instruction parameters recognizable by the simulator; Receive the inquiry-type instruction, determine the type of the inquiry-type instruction; if it is determined to be a re-confirmation-type instruction, then according to the air traffic control repetition rule, the pilot repeats the statement instruction, segment the aircraft call sign and other instructions in the re-confirmation-type instruction according to the corresponding semantic tag sequence, reorder them, and end with the aircraft call sign to obtain the instruction repetition text of the inquiry-type instruction; if it is determined to be a negotiation instruction, then according to the air traffic control repetition rule, extract the inquiry control intention parameters from the inquiry control intention corresponding to the negotiation instruction, the captain judges the inquiry control intention and the inquiry control intention parameters according to the flight situation, and repeats the instruction if it is legal, and at the same time converts the control identification intention and the control identification intention parameters into instruction parameters that can be recognized by the simulator; if it is illegal, start the next round of negotiation.

7. A method for generating text instructions for air traffic control simulation training according to any one of claims 1 to 6, characterized in that: The step S1 also includes: constructing a special service handling module to handle special situation scenarios encountered during the flight, and generate and output captain's text instructions.

8. The method for generating text instructions for air traffic control simulation training according to claim 7, characterized in that: The special situation scenarios in step S1 include: aircraft failure, crew emergency, and external environmental interference.

9. A text instruction generation device for air traffic control simulation training, characterized in that: The invention comprises at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method according to any one of claims 1 to 8.

Images