Method for generating control command and electronic apparatus therefor

The method using a humanoid robot with AI modules for generating and correcting control commands addresses aviation automation challenges, enhancing safety and reliability by adapting to aircraft state changes and ensuring compliance with safety conditions.

KR103044971B1Active Publication Date: 2026-11-06REPUBLIC OF KOREADEFENSE ACQUISITION PROGRAM ADMINISTATION
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
KR1020260039110
Authority / Receiving Office
KR · KR
Patent Type
Patents
Current Assignee / Owner
Filing Date
2026-03-04
Publication Date
2026-11-06
Estimated Expiration
2046-03-04

AI Technical Summary

Technical Problem

Existing automation technologies in aviation struggle to handle exceptional situations, complex abnormal conditions, and high-level decisions while ensuring low-latency safety control, particularly in aircraft with diverse types and changing operating conditions, necessitating immediate responses to maintain stability and safety.

Method used

A method using a humanoid robot with an artificial intelligence system, including a VCAL module for updating flight policies, an HL-PSR module for generating initial control commands, and an SD/SC-ICR module for safety verification, to generate and correct control commands based on aircraft state vectors and safety flags, ensuring safety and adaptability.

Benefits of technology

Enhances safety and reliability by reducing unsafe control instructions and improving response to abnormal situations, maintaining stability during flight, and simplifying integration with internal controllers.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 112026026370662-PAT00236_ABST
    Figure 112026026370662-PAT00236_ABST
Patent Text Reader

Abstract

A method is provided for an electronic device linked to a humanoid robot to generate a control command for an aircraft. The method for generating a control command may include the steps of: obtaining a state vector of the aircraft measured from a sensor of the humanoid robot and verifying a plurality of safety flags pre-set for the aircraft; generating structured aircraft state information based on the state vector of the aircraft and the plurality of safety flags; and inputting the structured aircraft state information into an artificial intelligence system embedded in the humanoid robot to generate a control command for changing the state of the aircraft according to changes in the environment surrounding the aircraft.
Need to check novelty before this filing date? Find Prior Art

Description

Technology Field

[0001] The technical concept of the present disclosure relates to a method for generating control commands and an electronic device for doing so, and more specifically, to a method for generating control commands for an aircraft and an electronic device for performing the same. Background Technology

[0002] Automation technology in the aviation sector has evolved into autopilots, flight management systems, and fault prediction and health management technologies, which can maintain flight status or execute operational procedures based on pre-set rules or control logic. However, automation technology relying on rule-based logic may find it difficult to adequately account for exceptional situations, complex abnormal conditions, or new operational procedures in situations where aircraft types diversify or operating conditions change. Furthermore, as high-level decisions such as air traffic control communication, procedural judgment, and regulation interpretation may be required during operations, and low-latency safety control linked to attitude, speed, angular velocity, and maneuver rate may be required simultaneously, it can become complex to consistently process functions of different time scales within a single system.

[0003] Furthermore, while aircraft failure prediction and durability management technologies can detect anomalies and provide alerts, an immediate closed-loop response system may also be required to restore a safe flight state in situations involving sudden environmental changes, engine status changes, or reduced sensor reliability during flight. In particular, an operational mode may be required to maintain a stable state by immediately adjusting control inputs in response to anomalies occurring during flight, while satisfying constraints limited by safety invariants such as speed, load, and maneuverability. The problem to be solved

[0004] The technical concept of the present disclosure provides a structure for generating control commands that allows control instructions involving high-level judgment and low-latency safety control to be operated simultaneously, while reflecting the state and safety conditions of an aircraft observed during flight together in the control judgment, and provides a method for generating control commands and an electronic device for performing the same, which allows operation to continue by verifying the safety of the control commands even in situations where abnormal signs are observed.

[0005] The technical problems to be solved by the present disclosure are not limited to the technical problems described above, and other technical problems not mentioned will be clearly understood by those skilled in the art to which the present disclosure belongs from the description below. means of solving the problem

[0006] A method for generating a control command for an aircraft performed by an electronic device linked to a humanoid robot may include: obtaining a state vector of the aircraft measured from a sensor of the humanoid robot and verifying a plurality of safety flags pre-set for the aircraft; generating structured aircraft state information based on the state vector of the aircraft and the plurality of safety flags; and inputting the structured aircraft state information into an artificial intelligence system embedded in the humanoid robot to generate the control command for changing the state of the aircraft according to changes in the environment surrounding the aircraft, and the humanoid robot may control the aircraft based on the control command.

[0007] In a method for generating control commands performed by an electronic device according to one embodiment, the artificial intelligence system may include a Verified Constraint Autonomous Learning (VCAL) module that updates flight policy parameters of the aircraft based on the structured aircraft state information.

[0008] In a method for generating a control command performed by an electronic device according to one embodiment, the step of generating the control command may include: a step of generating an initial control command by processing the structured aircraft status information based on an HL-PSR (High / Low-rate Perception Safety Reasoning) module; and a step of generating a final control command that corrects the initial control command by determining whether the initial control command is safe based on an SD / SC-ICR (Self-Debug / Self-Coding In-flight Code Repair) module.

[0009] In a method for generating a control command performed by an electronic device according to one embodiment, the step of generating the initial control command may include: generating a micro-intention to control the aircraft from the structured state information through a Large Language Model (LM) included in the artificial intelligence system; determining the validity of the micro-intention based on the operation of a gating function that takes the micro-intention and the plurality of safety flags as inputs; and generating the initial control command based on the validity of the micro-intention.

[0010] In a method for generating a control command performed by an electronic device according to one embodiment, the step of generating the initial control command based on whether the micro-intention is valid may include, if the micro-intention is determined to be valid, the step of converting the micro-intention to generate the initial control command.

[0011] In a method for generating a control command performed by an electronic device according to one embodiment, the step of generating the initial control command based on whether the micro-intention is valid may include the step of nullifying the micro-intention by determining that the micro-intention is unsafe when the micro-intention is determined to be invalid.

[0012] In a method for generating a control command performed by an electronic device according to one embodiment, the step of generating the final control command may include: a step of calculating a comprehensive anomaly corresponding to an error in the state vector of the aircraft when the state of the aircraft changes according to the initial control command; and a step of generating the final control command based on whether the comprehensive anomaly exceeds a preset threshold value.

[0013] In a method for generating a control command performed by an electronic device according to one embodiment, the step of generating the final control command may include: generating a candidate patch for correcting the initial control command based on a Domain Specific Language (DSL) when the overall anomaly exceeds a threshold value; determining whether the candidate patch is applicable by applying the candidate patch to a sandbox on an onboard Digital Twin (DT) and performing a stress test; and generating the final control command by correcting the initial control command based on the candidate patch when it is determined that the candidate patch is applicable.

[0014] In a method for generating a control command performed by an electronic device according to one embodiment, the step of generating the final control command may include, when it is determined that the candidate patch is not applicable, generating a command identical to the initial control command as the final control command.

[0015] In a method for generating a control command performed by an electronic device according to one embodiment, the step of generating the final control command may include, when the total abnormality does not exceed the threshold value, generating a command identical to the initial control command as the final control command.

[0016] A non-transient computer-readable storage medium capable of reading a computer and storing a program for executing a method for generating a control command for an aircraft according to one embodiment, wherein the method for generating the control command comprises: a step of obtaining a state vector of the aircraft measured from a sensor of a humanoid robot and verifying a plurality of safety flags pre-set for the aircraft; a step of generating structured aircraft state information based on the state vector of the aircraft and the plurality of safety flags; and a step of inputting the structured aircraft state information into an artificial intelligence system embedded in the humanoid robot to generate the control command for changing the state of the aircraft according to changes in the environment surrounding the aircraft, and wherein the humanoid robot controls the aircraft based on the control command.

[0017] An electronic device linked to a humanoid robot, which performs a method for generating a control command for an aircraft according to one embodiment, may be provided, comprising: a processor; and one or more memories for storing one or more instructions, wherein, when the one or more instructions are executed, the processor controls the processor to generate a control command for changing the state of the aircraft according to changes in the environment surrounding the aircraft by acquiring a state vector of the aircraft measured from a sensor of the humanoid robot, checking a plurality of safety flags pre-set for the aircraft, generating structured aircraft state information based on the state vector of the aircraft and the plurality of safety flags, and inputting the structured aircraft state information into an artificial intelligence system embedded in the humanoid robot, and the humanoid robot controls the aircraft based on the control command. Effects of the invention

[0018] According to the method and electronic device of the exemplary embodiment of the present disclosure, since state information and safety conditions observed during flight can be reflected together in the generation of control commands, the frequency of outputting control instructions that do not consider safety conditions can be reduced, and accordingly, safety during operation can be improved.

[0019] In addition, according to an exemplary embodiment of the present disclosure, control commands or control parameters may be applied after undergoing a verification procedure in situations where abnormal signs are observed, thereby improving the reliability of the response to exceptional situations during flight, and consequently reducing unnecessary waiting, operational delays, or inspection burdens.

[0020] Furthermore, according to an exemplary embodiment of the present disclosure, since the output of a control command can be standardized into a normalized control axis command form, linkage with an internal robot controller can be simplified, and the convenience of application can be improved even when the model or platform changes.

[0021] The effects obtainable from the present disclosure are not limited to those described above, and other unmentioned effects will be clearly understood by those skilled in the art from the description below to which the present disclosure pertains. Brief explanation of the drawing

[0022] FIG. 1 is a block diagram showing the configuration of an electronic device according to an exemplary embodiment of the present disclosure. FIG. 2 is a flowchart illustrating a method for generating control commands for an aircraft according to an exemplary embodiment of the present disclosure. FIG. 3 is a flowchart illustrating a method for generating control commands by inputting structured aircraft status information into an artificial intelligence system according to an exemplary embodiment of the present disclosure. FIG. 4 is a flowchart illustrating a method for generating an initial control command according to an exemplary embodiment of the present disclosure. FIG. 5 is a flowchart illustrating a method for generating a final control command according to an exemplary embodiment of the present disclosure. FIG. 6 is a block diagram showing an artificial intelligence system according to an exemplary embodiment of the present disclosure. FIG. 7 is a block diagram illustrating an HL-PSR module according to an exemplary embodiment of the present disclosure. FIG. 8 is a block diagram illustrating an SD / SC-ICR module according to an exemplary embodiment of the present disclosure. Specific details for implementing the invention

[0023] Hereinafter, various embodiments of the present disclosure are described in detail with reference to the accompanying drawings so that those skilled in the art can easily practice them. The detailed description disclosed below is intended to explain exemplary embodiments of various examples and is not intended to limit the scope to specific embodiments.

[0024] The terms used in the embodiments have been selected to be as widely used as possible, taking into account their functions in the present disclosure; however, these may vary depending on the intent of those skilled in the art, case law, the emergence of new technologies, etc. Additionally, in specific cases, terms have been arbitrarily selected by the applicant, and in such cases, their meanings will be described in detail in the relevant explanatory section. Therefore, terms used in the present disclosure should be defined not merely by their names, but based on their meanings and the overall content of the present disclosure.

[0025] The following embodiments are combinations of the components and features of various embodiments in a predetermined form. Each component or feature may be considered optional unless otherwise explicitly stated. Each component or feature may be implemented in a form not combined with other components or features. Additionally, various embodiments may be constructed by combining some components and features. The order of operations described in various embodiments may be changed. Some components or features of one embodiment may be included in another embodiment, or may be replaced with corresponding components or features of another embodiment.

[0026] In the description of the drawings, procedures or steps that could obscure the essence of the various embodiments were not described, nor were procedures or steps that could be understood by a person with ordinary knowledge in the relevant technical field described.

[0027] Throughout the specification, when a part is described as "comprising" or "including" a component, it means that, unless specifically stated otherwise, it does not exclude other components but may include additional components. In this document, the singular form of a noun corresponding to an item may be used to include both singular and plural forms, unless otherwise indicated in the specification or clearly contradicted by the context. In this document, phrases such as "A or B," "at least one of A and B," "at least one of A or B," "A, B or C," "at least one of A, B and C," and "at least one of A, B, or C" may each include any one of the items listed together in the corresponding phrase, or any combination thereof. Terms such as "first," "second," or "first" or "second" may be used simply to distinguish a component from other corresponding components and do not limit the components in any other aspect (e.g., importance or order).

[0028] Each component (e.g., module or program) of the components described in this document may include a singular or multiple entities. According to various embodiments, one or more of the components or operations may be omitted, or one or more other components or operations may be added. Additionally, multiple components (e.g., module or program) may be integrated into a single component. In this case, the integrated component may perform one or more functions of each of the components of the multiple components in the same or similar manner as they were performed by each of the components of the multiple components prior to integration. According to various embodiments, operations performed by a module, program, or other component may be executed sequentially, in parallel, or repeatedly; one or more of the operations may be executed in a different order; may be omitted; or one or more other operations may be added.

[0029] As used in this document, the terms "module" or "...part" refer to a unit that processes at least one function or operation, and may include a unit implemented by hardware, software, firmware, or a combination thereof.

[0030] Various embodiments of this document may be implemented as software (e.g., a program or application) comprising one or more instructions stored in a storage medium (e.g., memory) readable by a machine. For example, the processor of the machine may call at least one of the one or more instructions stored from the storage medium and execute it. This enables the machine to operate to perform at least one function according to at least one called instruction. One or more instructions may include code generated by a compiler or code that can be executed by an interpreter. The storage medium readable by a machine may be provided in the form of a non-transitory storage medium. Here, 'non-transitory' simply means that the storage medium is a tangible device and does not contain a signal, and does not distinguish between cases where data is stored semi-permanently and cases where it is stored temporarily in the storage medium.

[0031] In addition, specific terms used in various embodiments are provided to aid in understanding the various embodiments, and the use of such specific terms may be modified in other forms within the scope of not departing from the technical concept of the various embodiments.

[0032] The method for generating control commands for an aircraft and the electronic device for doing so proposed in the present disclosure may correspond to a method and an electronic device for performing the same, which secure the safety and reliability of generating control commands even in situations where environmental changes and state uncertainty exist during flight by acquiring a state vector of an aircraft measured from a sensor of a humanoid robot, verifying a plurality of pre-set safety flags for the aircraft, generating structured aircraft state information based on the state vector and safety flags, inputting the structured aircraft state information into an artificial intelligence system to generate control commands for changing the state of the aircraft according to changes in the environment surrounding the aircraft, and having the humanoid robot control the aircraft based on the generated control commands. In particular, since an initial control command can be generated based on an HL-PSR module, a final control command can be generated by determining the safety status of the initial control command based on an SD / SC-ICR module, and flight policy parameters can be updated based on a VCAL module, the generation, verification, and correction of control commands can be performed in conjunction with safety conditions.

[0033] In particular, the control command generation method of the present disclosure can be effectively applied in defense sectors where flight safety and mission execution reliability are directly linked, such as military aircraft, unmanned aerial vehicles (UAVs), and hybrid manned-unmanned operational systems. In battlefield environments, sudden weather changes, electromagnetic interference, reduced sensor reliability, communication delays, and changes in maneuver requirements during a mission may occur simultaneously; consequently, there is a risk that automatically generated control instructions may conflict with safety conditions or result in unexpected state changes. Therefore, a control command generation structure that adaptively responds to environmental changes while reflecting safety conditions during flight may be required.

[0034] In response to the characteristics of the defense sector, the present disclosure generates structured state information based on state vectors and safety flags, generates micro-intentions through an artificial intelligence system embedded in a humanoid robot, and determines the validity of the micro-intentions using gating operations with safety flags as input, thereby selectively activating only safe control paths. Additionally, after the application of an initial control command, a comprehensive anomaly corresponding to the state vector error is calculated, and if the comprehensive anomaly exceeds a threshold value, a candidate patch is generated to determine applicability through a stress test in an onboard digital twin sandbox, after which the control command can be corrected based on the applicable candidate patch. Therefore, the present disclosure may constitute a technical concept that can contribute to improving the stability and reliability of aircraft control even in battlefield environments and supporting the safe operation of military aviation platforms and unmanned aerial systems.

[0035] FIG. 1 is a block diagram showing the configuration of an electronic device (100) according to an exemplary embodiment of the present disclosure.

[0036] Referring to FIG. 1, the electronic device (100) may include a processor (110) and a memory (120) according to one embodiment. The electronic device (100) illustrated in FIG. 1 is illustrated only with components related to the present embodiment, and it will be understood by those skilled in the art related to the present embodiment that other general-purpose components may be included in addition to the components illustrated in FIG. 1.

[0037] For example, an electronic device (100) may include a communication device comprising one or more transceivers, an input unit, and an output unit. The communication device is a device for performing wired / wireless communication and can communicate with an external electronic device. The external electronic device may be a terminal or a server. In addition, communication technologies used by the communication device may include GSM (Global System for Mobile communication), CDMA (Code Division Multi Access), LTE (Long Term Evolution), 5G, WLAN (Wireless LAN), Wi-Fi (Wireless-Fidelity), Bluetooth, RFID (Radio Frequency Identification), Infrared Data Association (IrDA), ZigBee, NFC (Near Field Communication), etc. The input unit may be, for example, a traditional type of keypad or keyboard, a mouse, a microphone for receiving voice signals, a camera, and various other types of input means for detecting or receiving various forms of user input. The output unit may be, for example, a display that outputs images, a speaker that outputs sound, a haptic device that generates vibrations, and various other forms of output means.

[0038] Additionally, at least some of the components within the electronic device (100) may be implemented as an integrated unit or as a single or multiple unit. At least some of the components within the electronic device (100) may be connected to each other via a bus, GPIO (general purpose input / output), SPI (serial peripheral interface) or MIPI (mobile industry processor interface), etc., to exchange data and / or signals.

[0039] The electronic device (100) of FIG. 1 can perform a method for generating control commands. Specifically, the electronic device (100) can obtain a state vector of an aircraft measured from a sensor of a humanoid robot, check a plurality of safety flags pre-set for the aircraft, generate structured aircraft state information based on the state vector of the aircraft and the plurality of safety flags, and input the structured aircraft state information into an artificial intelligence system embedded in the humanoid robot to generate a control command for changing the state of the aircraft according to changes in the environment around the aircraft.

[0040] A processor (110) is a component capable of performing operations or data processing regarding the control and / or communication of each component of an electronic device (100). For example, the processor (110) can control the electronic device (100) overall by executing programs stored in memory (120) within the electronic device (100). The processor (110) may be implemented as a CPU (central processing unit), GPU (graphics processing unit), AP (application processor), etc., provided within the electronic device (100), but is not limited thereto. Unless there are special circumstances, in this disclosure, the processor (110) may refer to a set of one or more processors (110).

[0041] The processor (110) may be implemented as a computer or a similar device according to hardware, software, or a combination thereof. In terms of hardware, the processor (110) may be implemented in the form of an electronic circuit that processes electrical signals to perform control functions, and in terms of software, it may be implemented in the form of a program that drives the hardware processor (110). Meanwhile, unless otherwise specifically mentioned in the description below, the operation of the electronic device (100) may be interpreted as being performed by the control of the processor (110). That is, when modules implemented in the electronic device (100) are executed to perform a control command generation method, it may be interpreted that the processor (110) controls the modules to perform the following operations of the electronic device (100).

[0042] Memory (120) is hardware that stores various data processed within the electronic device (100), and memory (120) can temporarily or semi-permanently store data processed and data to be processed in the electronic device (100). For example, data regarding an operating system (OS) for operating the electronic device (100) can be stored in the memory (120) of the electronic device (100). Memory (120) may include non-volatile memory such as EEPROM (electrically erasable programmable read-only memory), CD-ROM (compact disc read-only memory), Blu-ray disc, HDD (hard disk drive), SSD (solid state drive), etc. Such memory (120) may be provided as an embedded type or a removable type. In addition, memory (120) may store one or more instructions for the operation of the processor (110). Unless there are special circumstances, in the present disclosure, memory (120) may mean a set of one or more memories (120).

[0043] The method for generating control commands according to the present disclosure, performed by the electronic device (100) of FIG. 1, may also be implemented by a non-transient computer-readable storage medium (or non-transient recording medium) for the operation. The method for generating control commands may be implemented as a software module or algorithm and may be stored on a computer-readable recording medium as computer-readable codes or program instructions executable on a processor (110). Here, computer-readable recording media include magnetic storage media (e.g., ROM (read-only memory), floppy disk, hard disk, etc.) and optical reading media (e.g., CD-ROM, DVD: Digital Versatile Disc). Computer-readable recording media may be distributed across networked computer systems, so that computer-readable code can be stored and executed in a distributed manner. The medium may be computer-readable, stored in memory (120), and executed on a processor (110).

[0044] An electronic device (100) according to one embodiment may further include a display (not shown). The display may visually provide various information to the outside under the control of a processor (110). The display may include a touch circuit configured to detect a user's touch or a sensor circuit configured to measure the intensity of the force generated by the touch.

[0045] FIG. 2 is a flowchart illustrating a method for generating a control command for an aircraft according to an exemplary embodiment of the present disclosure. Referring to FIG. 2, the method of FIG. 2, performed by an electronic device (100), may include steps S210 to S230.

[0046] In step S210, the electronic device (100) is the aircraft's state vector measured from the humanoid robot's sensor ( Can obtain ). Aircraft state vector( ) may include current aircraft status information data such as position, speed, attitude, engine, and environment. In addition, the electronic device (100) may include information corresponding to the aircraft's flight plan and current flight phase ( ), context related to aircraft weather, communication, constraints, etc. ( ), aircraft's past event log( You can obtain ).

[0047] In one embodiment, the state vector of the aforementioned aircraft (from the sensor of the humanoid robot) Acquiring information can be performed through multiple perception paths. For example, an electronic device (100) can receive a data stream output from a glass cockpit such as a Garmin G1000, and the electronic device (100) can parse the received data stream to acquire information including speed, altitude, heading, attitude, and engine-related parameters. Additionally, a camera mounted on a humanoid robot can photograph an analog instrument panel and warning lights, and the electronic device (100) can acquire relevant information by recognizing the instrument indicator values ​​and the lighting status of the warning lights from the captured images. Furthermore, a microphone mounted on a humanoid robot can receive communication voice, and the electronic device (100) can convert the received communication voice into text using a Speech-To-Text (STT) method to acquire information associated with communication events or operation procedures. An electronic device (100) according to one embodiment can time-align information obtained from the aforementioned plurality of perception paths based on timestamps, and fuses the time-aligned state items to form a state vector ( You can ultimately obtain ).

[0048] In one embodiment, the electronic device (100) has a plurality of safety flags (pre-set for the aircraft) ) can be checked. Here, multiple safety flags ( ) is, state vector( Based on ), safety invariant ( F It may be a set of binary safety flags pre-set for an aircraft by being calculated based on ), where safety invariant ( F ) is a formula containing minimum safety conditions to ensure safety, and may be a set of formulas including, for example, limits related to the aircraft's stall speed, limits related to pitch, roll, load, and maneuvering rate, validity of command format and command range, and data reliability conditions. In one embodiment, safety invariant ( F As the satisfaction status of each condition of ) is mapped to each flag, multiple safety flags including each flag ( ) can be calculated and pre-set for the aircraft, and the electronic device (100) has a plurality of safety flags ( ) can be checked. Multiple safety flags ( ) can be expressed as in [Mathematical Formula 1].

[0049] [Mathematical Formula 1]

[0050]

[0051]

[0052] In [Equation 1], each of the multiple safety flags may be a flag for determining whether some elements of the aircraft are within a reference range for the safe flight of the aircraft, for example, It may be a flag for determining whether the aircraft is above a standard speed for safe flight.

[0053] In step S220, the electronic device (100) has a state vector ( ) and multiple safety flags ( Structured aircraft status information based on ) Can generate ). Structured aircraft status information ( ) can be configured in a form that can be processed as input by an artificial intelligence system according to one embodiment, for example, structured aircraft state information ( ) may be in JSON file format. For example, the electronic device (100) is a state vector ( ), multiple safety flags ( ), information corresponding to the aircraft's flight plan and current flight phase ( ), context related to aircraft weather, communication, constraints, etc. ( ), and aircraft's past event logs ( Structured aircraft state information based on a function (f) that takes ) as input ( It can generate ) which can be expressed as [Equation 2].

[0054] [Mathematical Formula 2]

[0055]

[0056] In step S230, the electronic device (100) has aircraft state information structured in an artificial intelligence system embedded in a humanoid robot ( By inputting ), control commands can be generated to change the state of the aircraft according to changes in the environment surrounding the aircraft. The control commands can be output as normalized control axis commands that can be converted into physical manipulation of the humanoid robot, and the generated control commands and the basis for their generation can be recorded as logs in the form of JSON files. The control commands may include initial control commands and final control commands, and the process of generating initial control commands and final control commands will be described in more detail later with reference to FIG. 3.

[0057] FIG. 3 is a flowchart illustrating a method for generating a control command by inputting structured aircraft status information into an artificial intelligence system according to an exemplary embodiment of the present disclosure. Referring to FIG. 3, the method of FIG. 3, performed by an electronic device (100), may include steps S310 and S320.

[0058] In one embodiment, an artificial intelligence system embedded in a humanoid robot, structured aircraft state information ( By processing ) safety invariant( F ) and / or multiple safety flags( It may include an HL-PSR (High / Low-rate Perception Safety Reasoning) module for generating an initial control command based on ), an SD / SC-ICR (Self-Debug / Self-Coding In-flight Code Repair) module for determining the safety of the initial control command based on changes in the aircraft state after the application of the initial control command and, if necessary, generating a final control command through candidate patch-based correction, and a VCAL (Verified Constraint Autonomous Learning) module for updating flight policy parameters within the scope of maintaining safety constraints based on operational logs.

[0059] In step S310, the electronic device (100) can generate an initial control command by processing structured aircraft state information based on an HL-PSR module included in an artificial intelligence system. In one embodiment, the HL-PSR module may include a high-period safety / control layer corresponding to an H-PSR module and a low-period inference layer corresponding to an L-PSR module, for example, the high-period safety / control layer may have a period of 10 Hz or more, and the low-period inference layer may have a period of 1 Hz or less. In one embodiment, the electronic device (100) through the low-period inference layer, a micro-intention ( Can generate ).

[0060] According to one embodiment, the electronic device (100) has a plurality of safety flags ( ) and safety invariant( F Based on ), micro-intention( It can determine whether ) can be converted into a control command for controlling the aircraft. Accordingly, the initial control command can be generated in a form that reflects the safety conditions included in the structured aircraft state information, which will be explained in detail with reference to FIG. 4 and / or FIG. 7 described later.

[0061] In step S320, the electronic device (100) can generate a final control command that corrects the initial control command by determining whether the initial control command is safe based on the SD / SC-ICR module included in the artificial intelligence system. For example, assuming a case where the state of the aircraft changes according to the initial control command generated in step S310, if it is determined that the changed state of the aircraft is unsafe, a final control command that corrects the initial control command within a safe range can be generated. The process regarding step S320 will be described in detail with reference to FIG. 5 and / or FIG. 8, which will be described later.

[0062] FIG. 4 is a flowchart illustrating a method for generating an initial control command according to an exemplary embodiment of the present disclosure. Referring to FIG. 4, the method of FIG. 4 may include steps S410 to S450.

[0063] In step S410, the electronic device (100) receives structured aircraft state information ( Micro-intention to control the aircraft from ) Can generate ). Micro-intention( ) can be defined as an immediately actionable goal instruction, for example, a goal heading( ), target speed( ), target altitude( ), and rate of decline or rise ( It may include ).

[0064] According to one embodiment, micro-intention ( ) is structured aircraft status information ( Context related to aircraft weather, communication, constraints, etc., which is information included in ) ), and aircraft's past event logs ( It can be generated based on at least one of ), and thus traceability and auditability can be ensured.

[0065] In step S420, the electronic device (100) has a micro-intention ( ), multiple safety flags ( ) and safety invariant( F Based on the operation of a gating function (Gate) with ) as input, micro-intention ( The validity of ) can be determined, and this may be equal to [Equation 3].

[0066] [Mathematical Formula 3]

[0067]

[0068] In [Equation 3], when the micro-intention is determined to be valid, that is, the micro-intention ( ) multiple safety flags( ) and / or safety invariants( F If it is determined that there is no violation of safety conditions according to ), micro-intention ( ) can be output based on a gating function (Gate), and micro-intention( If ) is determined to be invalid, that is, micro-intention( ) multiple safety flags( ) and / or safety invariants( FIf it is determined that there is a possibility of violating safety conditions in accordance with ), a command (FAILSAFE) to nullify the micro-intention and switch the aircraft to a safe state may be issued.

[0069] In step S430, the electronic device (100) can determine whether the micro-intention is valid based on [Equation 3], and if the micro-intention is determined to be valid, step S440 can be performed, and if the micro-intention is determined to be invalid, step S450 can be performed.

[0070] In step S440, the micro-intention (by the gating function (gate) If ) is determined to be valid, the electronic device (100) is a micro-intention ( ) can be converted to generate an initial control command. In one embodiment, the electronic device (100) has a micro-intention ( Target headings included in ) ), target speed( ), target altitude( ), and rate of decline or rise ( Using multiple reference values ​​such as ), they can be mapped to a low-latency control command that is updated at a high frequency, for example, the low-latency control command can be updated at a frequency set to 10 Hz or higher. In one embodiment, the electronic device (100) can derive reference values ​​for each of the Roll, Pitch, Rudder, and Throttle axes by calculating the difference between the multiple reference values ​​and the current state measurement value, and can generate an initial control command by converting the derived reference values ​​into normalized control axis commands. For example, the initial control command is a target roll command for controlling the roll axis ( ), target pitch command corresponding to vertical tilt command based on speed / altitude error( ), normalized roll command( ), normalized pitch command( ), normalized rudder pedal axis command( ), binary commands for left / right brakes( ), and normalized throttle axis command( It may include ).

[0071] In one embodiment, the electronic device (100) has a target heading ( ) and current heading( The difference of ) Angle calculated It can be normalized to the range and reflected in roll axis control, and side slip Target roll command including as a correction term ( ) can be calculated, which can be equal to [Equation 4].

[0072] [Mathematical Formula 4]

[0073]

[0074] In [Equation 4], the sat() function may be a function for limiting commands to a safe range, and and Each Current heading used in calculations ( ) and side slip It can be a pre-set weight corresponding to each.

[0075] In one embodiment, the electronic device (100) has a target speed ( ) and target altitude( ) and the target pitch command based on the difference between the current speed (v) and the current altitude (h) ) can be calculated, which can be equal to [Equation 5].

[0076] [Mathematical Formula 5]

[0077]

[0078] In [Mathematical Formula 5], and Each may be a preset weight corresponding to the current speed (v) and current altitude (h), respectively.

[0079] In one embodiment, the electronic device (100) and cast and Normalized roll command by normalizing based on and then restricting to a safe range using the sat() function ( ), and normalized pitch command( ) can be calculated, which can be expressed as [Equation 6] and / or [Equation 7].

[0080] [Mathematical Formula 6]

[0081] ,

[0082] [Mathematical Formula 7]

[0083] ,

[0084] In one embodiment, the electronic device (100) is a normalized rudder pedal axis command based on at least some of the aforementioned state information and weights ( It can produce ) and, if necessary, binary commands for left / right brakes ( ) can be calculated, and this can be equal to [Equation 8].

[0085] [Mathematical Formula 8]

[0086]

[0087]

[0088] In one embodiment, the electronic device (100) is a normalized throttle axis command ( ) as the feedforward term( ) and speed difference( It can be calculated based on ), and this can be as [Equation 9].

[0089] [Mathematical Formula 9]

[0090] ,

[0091] In one embodiment, the electronic device (100) has a calculated target roll command ( ), Target Pitch Command( ), normalized roll command( ), normalized pitch command( ), normalized rudder pedal axis command( ), binary commands for left / right brakes( ), and normalized throttle axis command( ) can be output as an initial control command, and the humanoid pilot robot can physically operate the joystick, rudder, brake, and throttle based on the initial control command. In addition, the electronic device (100) can have a micro-intention ( The correspondence relationship and application time information of axis commands mapped to ) can be recorded in the log.

[0092] In step S450, the micro-intention (by the gating function (gate) If ) is determined to be invalid, the electronic device (100) is a micro-intention ( By judging ) as unsafe, micro-intention( ) can be nullified. In one embodiment, the electronic device (100) has a micro-intention ( ) multiple safety flags( ) and safety invariant( F If it is determined that there is a possibility of violating safety conditions pursuant to ), micro-intention ( It can output a command (FAILSAFE) to nullify ) and switch the aircraft to a safe state.

[0093] In one embodiment, the command (FAILSAFE) may be defined as a standardized avoidance or stop procedure. The standardized avoidance or stop procedure may be managed as a predefined procedure for each aircraft type, and a plurality of safety flags ( It can be selected depending on the type of occurrence among ).

[0094] In one embodiment, the electronic device (100) can generate and output a normalized control axis command corresponding to a command (FAILSAFE), and the axis command output therein is a micro-intention ( Rather than directly tracking the target value derived by ), multiple safety flags ( ) and safety invariant( F It can be set in a direction that stabilizes the state of the aircraft within a safety range defined by ). In addition, the electronic device (100) has a plurality of safety flags ( ) and safety invariant( F ) Information regarding potential violations, identification information of selected standard procedures, and output axis command information can be recorded in the log.

[0095] FIG. 5 is a flowchart illustrating a method for generating a final control command according to an exemplary embodiment of the present disclosure. Referring to FIG. 5, the method of FIG. 5, performed by an electronic device (100), may include steps S510 to S570.

[0096] In one embodiment, the electronic device (100) can propose, verify, and apply a self-healing procedure through lightweight modifications, such as parameter, filter, and threshold correction, for anomalies during aircraft flight by utilizing a restricted Domain-Specific Language (DSL), an onboard digital twin, and a hot-swap application flow based on an SD / SC-ICR module. In this case, the modification range may be set as Tier-1, focusing on gain, threshold, control command scale, and filter adjustment, and as Tier-2, corrections at the level of short formulas may be allowed if necessary.

[0097] In step S510, the electronic device (100) has a composite anomaly corresponding to the error in the aircraft's state vector when the aircraft's state changes according to the initial control command ( ) can be calculated. In one embodiment, the electronic device (100) receives the initial control command generated in step S440 and estimated disturbances such as wind direction and wind speed as inputs based on the onboard digital twin, and the predicted value that the aircraft will exhibit in a fault-free normal state ( ) can be calculated in real time. The electronic device (100) can calculate the actual value ( ) and predicted values( The difference of ) is the residual( It can be calculated as ), and the residual( ) can be expressed as in [Mathematical Formula 10]. The electronic device (100) is the residual ( If the value increases, it can be determined that physical / soft abnormalities have occurred.

[0098] [Mathematical Formula 10]

[0099]

[0100] In addition, the electronic device (100) has a mission performance degradation amount ( ), residual deterioration amount( ), and safety flag occurrence frequency ( Comprehensive anomaly ( based on ) Can calculate ) and overall outlier( ) can be expressed as in [Equation 11]

[0101] [Mathematical Formula 11]

[0102]

[0103] Here, , can mean the weight for each item, and the frequency of safe flag occurrence ( ) may include the frequency of events such as stall warnings, overspeeding, and exceeding the handling rate, and the mission performance degradation amount ( ) may include values ​​corresponding to events such as an increase in tracking error, and the amount of residual deterioration ( ) is the residual calculated based on [Equation 10] ( It may include a value corresponding to an event where the average or maximum value of ) increases.

[0104] In step S520, the electronic device (100) has a total anomaly ( ) is a pre-set threshold value( It can determine whether it exceeds ). In one embodiment, the total anomaly ( ) Pre-set threshold value( If it exceeds ) ), the electronic device (100) can perform step S530 based on a self-coding pipeline. In one embodiment, the total anomaly ( ) is a pre-set threshold value( If ) or less, the electronic device (100) can perform step S570.

[0105] In step S530, the overall anomaly ( ) is a pre-set threshold value( If it is determined that it exceeds ), the electronic device (100) provides a candidate patch ( ) can be generated. In one embodiment, the electronic device (100) is a candidate patch ( It can be restricted to being expressed only by a limited DSL. The DSL can allow arithmetic operations, clamp, first-order filters, low-pass filters, moving average filters, and table lookups. Additionally, the DSL can restrict operations related to loops, file I / O, network I / O, random imports, and parallel threads.

[0106] In step S540, the electronic device (100) is a candidate patch ( By applying ) to the onboard DT sandbox and performing stress tests, the candidate patch ( It can determine whether ) is applicable. In one embodiment, the stress test may include strong winds, turbulence, sensor noise, and near-limit maneuver conditions. Additionally, the electronic device (100) may check mission performance indicators and whether safety conditions have been violated together to evaluate the results of the stress test.

[0107] In step S550, the electronic device (100) is a candidate patch ( based on the stress test results) It can determine whether ) is applicable. In one embodiment, a candidate patch ( The applicability conditions of ) are mission performance indicators ( Whether ) is improved, safety risk( Whether ) does not deteriorate, and safety invariant( F It may include whether there is no violation of the constraints defined in ), which can be expressed as [Equation 12] to [Equation 14].

[0108] [Mathematical Formula 12]

[0109]

[0110]

[0111] [Mathematical Equation 12] is a formula for quantitatively verifying whether mission performance has been improved, where can be a function for calculating the expected value, and ... may be at least one error among the aforementioned control axis commands with reference to [Equation 6] to [Equation 9], and It may be a pre-set mission performance standard value.

[0112] [Mathematical Formula 13]

[0113]

[0114]

[0115] [Mathematical Formula 13] is safety risk( It can be understood as a formula to check whether ) is increasing, and is a function that defines the safety set used in the CBF (Control Barrier Function), Is The rate of change of time, is a function representing safety used in the CLF (Control Lyapunov Function), Is It can represent the time rate of change of, and can mean weight.

[0116] [Mathematical Formula 14]

[0117]

[0118] [Equation 14] is a candidate patch ( ) and safety invariant( F It can be understood as an expression for checking whether there are any safety violations from the Inv function, which takes ) as input. For example, the Inv function is a safety invariant ( F It is possible to determine whether the previously set safety conditions are violated in response to ).

[0119] Based on the aforementioned S550 step, candidate patch ( If it is determined that ) is applicable, step S560 may be performed, and the candidate patch ( If it is determined that ) is not applicable, step S570 may be performed.

[0120] In step S560, candidate patch ( If it is determined that ) is applicable, the electronic device (100) is a candidate patch ( A final control command can be generated by correcting the initial control command based on ). For example, the electronic device (100) has a candidate patch ( ) can be applied to the adapter layer of a humanoid robot in a hot-swap manner. In one embodiment, a candidate patch ( When applied, the control axis commands may be updated, and the updated control axis commands may include Yoke-Roll, Yoke-Pitch, Rudder-Brake, Rudder-Pedal, and Throttle. Additionally, the electronic device (100) may record the reasons for application and rejection and related indicators in a log.

[0121] In step S570, candidate patch ( If it is determined that ) is not applicable, the electronic device (100) may generate a command identical to the initial control command as the final control command. In addition, candidate patch ( If side effects are observed after ) has been applied as a hot swap, the electronic device (100) can immediately perform a rollback, and indicators related to the rollback can be recorded in the log.

[0122] FIG. 6 is a block diagram illustrating an artificial intelligence system according to an exemplary embodiment of the present disclosure. Referring to FIG. 6, the artificial intelligence system embedded in a humanoid robot may include an HL-PSR module (610), an SD / SC-ICR module (620), a log storage unit (Logs) (630), and a VCAL module (640).

[0123] As described above with reference to FIGS. 2 to 5, the electronic device (100) is structured aircraft state information ( By inputting ) into the artificial intelligence system, an initial control command and / or a final control command can be generated through the HL-PSR module (610) and / or SD / SC-ICR module (620) included in the artificial intelligence system. The HL-PSR module (610) will be described in more detail with reference to FIG. 7, and the SD / SC-ICR module (620) with reference to FIG. 8.

[0124] In one embodiment, the log storage unit (630) may be a storage unit for recording the input and output of the artificial intelligence system and evaluation indicators generated during operation. The electronic device (100) is structured aircraft status information ( ), micro-intention( ), multiple safety flags ( ), safety invariant( F ), initial control command, final control command, overall anomaly ( ), candidate patch( The reasons for the application and rejection of the candidate patch, and the candidate patch, can be recorded in a log. The log can be stored with time information and can be stored in a JSON file format. Additionally, the electronic device (100) can record the application time, applied parameters or rules, and observation indicators after application in a log when the candidate patch is applied in a hot-swap manner, and can record the rollback time and the reason for the rollback in a log when a rollback is performed.

[0125] In one embodiment, the VCAL module (640) maintains a flight control policy ( It may be a module for updating ). The electronic device (100) has a flight control policy ( based on the VCAL module (640). The parameter (θ) of ) can be updated, and the updated parameter (θ) can be reflected in the subsequent control command generation process. For example, flight control policy ( ) is structured aircraft status information ( ) or state vector( It may be a policy for producing normalized control axis commands based on ), and parameter (θ) may be a set of parameters defining the policy.

[0126] In one embodiment, the loss function used by the VCAL module (640) can be set in a combined form of reinforcement learning loss and safety constraint loss, and the VCAL loss can be expressed as [Equation 15].

[0127] [Mathematical Formula 15]

[0128]

[0129] In [Mathematical Equation 15], can refer to general reinforcement learning loss, and and Each can represent an integrated loss combining CBF / CLF constraints, which can be expressed in more detail by [Equation 16]. and Each may be a weight corresponding to the combined loss of the CBF / CLF constraints.

[0130] [Mathematical Formula 16]

[0131]

[0132] The parameters expressed in [Equation 16] can correspond to the parameters expressed in [Equation 13] mentioned above.

[0133] In one embodiment, the electronic device (100) can perform learning to satisfy safety constraints during an update process based on the VCAL module (640), and a flight control policy reflecting the learning result ( ) and parameter (θ) can be set to be reflected in the operation of the HL-PSR module (610) and SD / SC-ICR module (620).

[0134] FIG. 7 is a block diagram illustrating an HL-PSR module (700) according to an exemplary embodiment of the present disclosure. Referring to FIG. 7, the HL-PSR module (700) may include an H-PSR module (710), an L-PSR module (720), and a Safe Gating module (730). The electronic device (100) provides structured aircraft state information ( An initial control command can be generated by processing ), and this operation can correspond to step S310 of FIG. 3 and steps S410 to S450 of FIG. 4.

[0135] In one embodiment, the electronic device (100) provides structured aircraft state information ( Micro-intention based on ) ) can be generated, which can correspond to step S410 of FIG. 4. In addition, the electronic device (100) can generate a plurality of safety flags ( ) and safety invariant( F ) can be calculated or updated, which is a plurality of safety flags ( in step S210 of FIG. 2) ) and safety invariant( F It can respond to verifying ).

[0136] In one embodiment, the electronic device (100) is based on the Safe Gating module (730) and has a micro-intention ( ), multiple safety flags ( ), and safety invariant( F Using ) as input, micro-intention( The validity of ) can be determined, which corresponds to steps S420 and S430 of FIG. 4, and determining the validity can correspond to [Equation 3].

[0137] In addition, the electronic device (100) has a micro-intention ( The initial control command can be generated by converting the ) into a normalized control axis command, which corresponds to step S440 of FIG. 4. At this time, the low-latency control conversion can correspond to generating the commands described in [Equation 4] through [Equation 9]. Additionally, if the gating result is derived as FAILSAFE, the electronic device (100) can select a standardized avoidance or stop procedure corresponding to step S450 of FIG. 4.

[0138] FIG. 8 is a block diagram illustrating an SD / SC-ICR module (800) according to an exemplary embodiment of the present disclosure. Referring to FIG. 8, the SD / SC-ICR module (800) may include an onboard digital twin (Onboard DT) (810), a determination unit (820), an onboard DT sandbox (830), and a hot-swap (840).

[0139] The electronic device (100) can determine whether the initial control command is safe based on the SD / SC-ICR module (800) and, if necessary, perform candidate patch-based correction to generate the final control command, and this operation may correspond to step S320 of FIG. 3 and steps S510 to S570 of FIG. 5.

[0140] In one embodiment, the electronic device (100) has a steady-state prediction value ( based on an onboard digital twin (810) It can calculate ) and the actual value( ) and predicted values( Residuals based on the difference of ) ) can be calculated, which corresponds to step S510 of FIG. 5 and [Equation 10]. In addition, the electronic device (100) can calculate the mission performance degradation amount ( ), residual deterioration amount( ), and safety flag occurrence frequency ( Comprehensive anomaly ( based on ) ) can be calculated, which corresponds to step S510 of FIG. 5 and [Equation 11].

[0141] The electronic device (100) determines the overall abnormality based on the judgment unit (820) ) is a pre-set threshold value( It can determine whether it exceeds ) and the overall anomaly ( ) is a pre-set threshold value( If determined to exceed ), candidate patch( ) can be generated, which can correspond to steps S520 and S530 of FIG. 5. Candidate patch ( ) can be expressed as a restricted DSL, and candidate patch( ) can be verified by a stress test in an onboard DT sandbox (830), which corresponds to step S540 of FIG. 5.

[0142] According to one embodiment, the electronic device (100) is a candidate patch ( The applicability of ) can be determined based on step S550 of FIG. 5 and [Equation 12] to [Equation 14].

[0143] The electronic device (100) is a candidate patch ( If it is determined that ) is applicable, the candidate patch can be applied to the adapter layer based on hot swap (840), which may correspond to step S560 of FIG. 5. Candidate patch ( If it is determined that ) is not applicable, the electronic device (100) may generate a final control command identical to the initial control command, which may correspond to step S570 of FIG. 5. In addition, candidate patch ( If side effects are observed after application, the electronic device (100) can perform a rollback.

[0144] In the foregoing, although all components constituting the embodiments disclosed in this document have been described as being combined or operating in combination, the embodiments disclosed in this document are not necessarily limited to such embodiments. That is, within the scope of the purposes of the embodiments disclosed in this document, all components may be selectively combined in one or more ways to operate.

[0145] Furthermore, terms such as "include," "compose," or "have" as described above, unless specifically stated otherwise, mean that the relevant component may be inherent; thus, they should be interpreted as allowing for the inclusion of additional components rather than excluding them. All terms, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the embodiments disclosed in this document pertain, unless otherwise defined. Commonly used terms, such as those defined in advance, should be interpreted in accordance with their meaning in the context of the relevant technology and, unless explicitly defined in this document, should not be interpreted in an ideal or overly formal sense.

[0146] The foregoing description is merely an illustrative explanation of the technical concept disclosed in this document, and a person skilled in the art to which the embodiments disclosed in this document pertain can make various modifications and variations within the scope of the essential characteristics of the embodiments disclosed in this document. Accordingly, the embodiments disclosed in this document are intended to explain, not limit, the technical concept of the embodiments disclosed in this document, and the scope of the technical concept disclosed in this document is not limited by these embodiments. The scope of protection of the technical concept disclosed in this document shall be interpreted by the claims below, and all technical concepts within an equivalent scope shall be interpreted as being included within the scope of rights of this document.

Claims

Claim 1 A method for generating control commands for an aircraft using an electronic device linked to a humanoid robot, comprising the steps of: obtaining a state vector of the aircraft measured from a sensor of the humanoid robot and verifying a plurality of pre-set safety flags calculated based on a safety invariant for the aircraft; and generating structured aircraft state information based on the state vector of the aircraft and the plurality of safety flags. A method for generating a control command, comprising the step of inputting the structured aircraft state information into an artificial intelligence system embedded in the humanoid robot to generate the control command for changing the state of the aircraft within a safety range defined by at least one of the safety invariant and the plurality of safety flags according to changes in the environment surrounding the aircraft, wherein the artificial intelligence system includes a Verified Constraint Autonomous Learning (VCAL) module that updates the flight policy parameters of the aircraft by performing reinforcement learning based on the structured aircraft state information, and the electronic device updates the flight policy parameters of the aircraft to satisfy safety constraints by performing reinforcement learning through the VCAL module using the VCAL loss calculated as the sum of a reinforcement learning loss, a loss based on a Control Barrier Function (CBF) constraint, and a loss based on a Control Lyapunov Function (CLF) constraint, and reflecting the updated flight policy parameters in the generation of the control command, and wherein the humanoid robot controls the aircraft based on the control command. Claim 2 delete Claim 3 A method for generating a control command according to claim 1, wherein the step of generating the control command comprises: a step of generating an initial control command by processing the structured aircraft status information based on an HL-PSR (High / Low-rate Perception Safety Reasoning) module; and a step of generating a final control command corrected by determining the safety status of the initial control command based on an SD / SC-ICR (Self-Debug / Self-Coding In-flight Code Repair) module. Claim 4 A method for generating a control command according to claim 3, wherein the step of generating the initial control command comprises: generating a micro-intention to control the aircraft from the structured state information through a Large Language Model (LM) included in the artificial intelligence system; determining the validity of the micro-intention based on the operation of a gating function that takes the micro-intention and the plurality of safety flags as inputs; and generating the initial control command based on the validity of the micro-intention. Claim 5 A method for generating a control command according to claim 4, wherein the step of generating the initial control command based on the validity of the micro-intention includes, when the micro-intention is determined to be valid, the step of converting the micro-intention to generate the initial control command. Claim 6 A method for generating a control command according to claim 4, wherein the step of generating the initial control command based on whether the micro-intention is valid includes the step of invalidating the micro-intention by determining that the micro-intention is unsafe when the micro-intention is determined to be invalid. Claim 7 A method for generating a control command according to paragraph 3, wherein the step of generating the final control command comprises: a step of calculating a composite anomaly corresponding to an error in the state vector of the aircraft when the state of the aircraft changes according to the initial control command; and a step of generating the final control command based on whether the composite anomaly exceeds a preset threshold value. Claim 8 A method for generating a control command according to claim 7, wherein the step of generating the final control command comprises: generating a candidate patch for correcting the initial control command based on a Domain Specific Language (DSL) when the overall anomaly exceeds the threshold value; determining whether the candidate patch is applicable by applying the candidate patch to a sandbox of an onboard Digital Twin (DT) and performing a stress test; and generating the final control command by correcting the initial control command based on the candidate patch when it is determined that the candidate patch is applicable. Claim 9 A method for generating a control command according to claim 8, wherein the step of generating the final control command includes, when it is determined that the candidate patch is not applicable, generating a command identical to the initial control command as the final control command. Claim 10 A method for generating a control command according to claim 7, wherein the step of generating the final control command includes the step of generating a command identical to the initial control command as the final control command when the overall abnormality does not exceed the threshold value. Claim 11 In a computer-readable, non-transient computer-readable storage medium storing a program for executing on a computer a method for generating control commands for an aircraft, the method for generating control commands comprises: a step of obtaining a state vector of the aircraft measured from a sensor of a humanoid robot and verifying a plurality of pre-set safety flags calculated based on a safety invariant for the aircraft; and a step of generating structured aircraft state information based on the state vector of the aircraft and the plurality of safety flags. A non-transient computer-readable storage medium comprising the step of inputting the structured aircraft state information into an artificial intelligence system embedded in the humanoid robot to generate the control command for changing the state of the aircraft within a safety range defined by at least one of the safety invariant and the plurality of safety flags according to changes in the environment surrounding the aircraft, wherein the artificial intelligence system includes a VCAL (Verified Constraint Autonomous Learning) module that updates the flight policy parameters of the aircraft by performing reinforcement learning based on the structured aircraft state information, and through the VCAL module, the reinforcement learning is performed using the VCAL loss calculated as the sum of the reinforcement learning loss, the loss based on the CBF (Control Barrier Function) constraint, and the loss based on the CLF (Control Lyapunov Function) constraint, thereby updating the flight policy parameters of the aircraft so as to satisfy the safety constraint, the updated flight policy parameters are reflected in the generation of the control command, and the humanoid robot controls the aircraft based on the control command. Claim 12 In an electronic device associated with a humanoid robot that performs a method of generating control commands for an aircraft, a processor; and includes one or more memories storing one or more instructions, wherein, when the one or more instructions are executed, the processor controls the processor to: acquire a state vector of the aircraft measured from a sensor of the humanoid robot, check a plurality of pre-set safety flags calculated based on a safety invariant for the aircraft, generate structured aircraft state information based on the state vector of the aircraft and the plurality of safety flags, and input the structured aircraft state information into an artificial intelligence system embedded in the humanoid robot to generate a control command for changing the state of the aircraft within a safety range defined by at least one of the safety invariant and the plurality of safety flags according to changes in the environment surrounding the aircraft, and the artificial intelligence system includes a Verified Constraint Autonomous Learning (VCAL) module that updates flight policy parameters of the aircraft by performing reinforcement learning based on the structured aircraft state information, and the electronic device, through the VCAL module, a reinforcement learning loss, a loss based on Control Barrier Function (CBF) constraints, and a Control An electronic device that performs reinforcement learning using the VCAL loss calculated as the sum of losses based on Lyapunov Function constraints, thereby updating the flight policy parameters of the aircraft to satisfy safety constraints, reflecting the updated flight policy parameters in the generation of the control command, and having the humanoid robot control the aircraft based on the control command.

Citation Information

Patent Citations

  • Airplane flying robot

    JP2006193034A

  • Training system and training method for operation inference learning model that controls automatically maneuvering robot

    JP2022048416A

  • Humanoid robot pilot system for autonomous flight method

    KR1020250051842A

  • JP03134577U