A method for wind tunnel assisted training of tiltrotor aircraft flight control law

Abstract

The present invention relates to a method for wind tunnel-assisted training of a tiltrotor flight control law, comprising initializing a wind tunnel simulation device for self-checking, setting the tiltrotor flight operation and corresponding expected instructions, flight mode and corresponding flight control law; continuously obtaining the current flight state parameters of the tiltrotor and calculating the error value with the expected flight state parameters, judging whether the current error value is less than a preset value; and adjusting the flight state parameters to the expected instructions based on the error value each time, and finally obtaining the first data link under different flight operations, different flight modes, different flight control laws and different flight control law parameters. The present invention continuously adjusts the flight state until the expected instruction is reached based on the error determined by the current actual flight state of the tiltrotor, solves the existing problem of error accumulation, improves the accuracy of flight control law training, and facilitates the execution of subsequent flight missions based on these first data links.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence / data-driven technology, and in particular to a method for wind tunnel-assisted training of tiltrotor aircraft flight control laws. Background Art

[0002] The existing technical means to verify the control law design principle is to use MATLAB / Simulink software for numerical simulation, such as Figure 1 As shown in the figure, the mathematical model is built using MATLAB / Simulink and then the parameters are adjusted and the simulation is verified. However, there is an error between the tiltrotor aircraft flight dynamics model built by MATLAB / Simulink and the actual model. There is an error between the flight state parameters output by the model and the actual flight state parameters. When the output flight state parameters are used as feedback parameters to form a closed-loop control, there is error accumulation in the iterative process of design, parameter adjustment and verification simulation. This error accumulation will cause the flight control law with good parameter adjustment to fail to achieve the expected control effect in actual flight, or even the error is too large, which poses a certain risk when performing autonomous flight missions.

[0003] In summary, the existing technology for designing the flight control law of a tiltrotor aircraft uses numerical simulation using MATLAB / Simulink software, which results in error accumulation, thereby affecting the accuracy of flight control law training. In view of this, the present invention provides a method for training the flight control law of a tiltrotor aircraft with the assistance of a wind tunnel. Summary of the invention

[0004] The purpose of the present invention is to provide a method for training the flight control law of a tiltrotor aircraft with the aid of a wind tunnel. The method uses a wind tunnel simulation device to train the flight control law of a tiltrotor aircraft, thereby solving the problem of error accumulation and improving the accuracy of flight control law training.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A method for wind tunnel-assisted training of a tiltrotor aircraft flight control law, comprising:

[0007] Performing an initialization self-check on the wind tunnel simulation device, setting the flight operation of the tiltrotor aircraft and the corresponding expected flight state parameters, flight mode, and flight control law under the current flight mode;

[0008] Using the inertial measurement unit in the wind tunnel simulation device to obtain the initial flight state parameters of the tiltrotor aircraft; setting the initial flight state parameters as the current actual flight state parameters;

[0009] Calculating an error value between the expected flight state parameter and the current actual flight state parameter, and determining whether the current error value is less than a preset value;

[0010] If not, driving the autopilot to calculate the control value update amount according to the current error value and the current flight control law;

[0011] Adjusting the flight state parameter of the tiltrotor aircraft to a target flight state parameter corresponding to the current manipulation value update amount according to the current manipulation value update amount;

[0012] Using the inertial measurement unit in the wind tunnel simulation device to obtain the current actual flight state parameter of the tiltrotor aircraft, and returning to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter, and determining whether the current error value is less than a preset value", until the current error value is less than the preset value;

[0013] If yes, determining the control effect score of the tiltrotor aircraft to achieve the current actual flight state parameter, and constructing a first data link including the expected flight state parameter, the flight mode, the flight control law, the flight control law parameter and the control effect score;

[0014] Randomly changing the parameter value of the flight control law, and making the flight control law after the parameter change the current flight control law, and returning to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter" until the number of parameter value changes reaches a preset number;

[0015] The type of the current flight control law is changed, and the flight control law after the type change is set as the flight control law under the current target flight mode, and the process returns to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter, and determining whether the current error value is less than a preset value", until all types of the flight control laws under the current target flight mode are traversed;

[0016] Changing the type of the current flight mode, making the changed flight mode the current target flight mode, and returning to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter, and determining whether the current error value is less than a preset value", until all the flight control laws under all types of the target flight modes are traversed; until all flight mode types are traversed;

[0017] Change the current flight operation and the current expected flight state parameter, set the changed flight operation as the current flight operation, and the changed expected flight state parameter as the current expected flight state parameter, and return to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter, and determining whether the current error value is less than a preset value", until all flight operation types are traversed;

[0018] storing each of the first data links under all the flight operations, all the flight modes, all the flight control laws and all the parameters of the flight control laws in a first database;

[0019] A flight strategy of the tiltrotor aircraft when performing a flight mission is determined according to each of the first data links stored in the first database.

[0020] Optionally, the initializing self-check of the wind tunnel simulation device, setting the flight operation of the tiltrotor aircraft and the corresponding expected flight state parameters, flight mode, and flight control law under the current flight mode specifically includes:

[0021] Driving the nine-channel servo steering engine in the wind tunnel simulation device to adjust the tilt-rotor aircraft to the current flight mode;

[0022] Driving the autopilot in the tiltrotor aircraft to switch to the flight control law in the current flight mode;

[0023] The current flight operation of the tiltrotor aircraft is set, and the expected flight state parameters are obtained according to the current flight operation.

[0024] Optionally, adjusting the flight state parameter of the tiltrotor aircraft to a target flight state parameter corresponding to the current manipulation value update amount according to the current manipulation value update amount specifically includes:

[0025] According to the current control value update amount, a nine-channel servo steering engine in the wind tunnel simulation device is driven to change the current control amount of the tilt-rotor aircraft;

[0026] The corresponding target flight state parameters are calculated according to the manipulation value update amount, and the six-degree-of-freedom motion mechanism in the wind tunnel simulation device is driven according to the calculation result, and the flight state parameters of the tilt-rotor aircraft are adjusted to the current target flight state parameters in cooperation with the wind tunnel laboratory blowing.

[0027] Optionally, determining the control effect score of the tiltrotor aircraft reaching the current actual flight state parameter specifically includes:

[0028] Outputting loop response curves of the stabilization loop, attitude angle loop, altitude control loop, and speed control loop when the tiltrotor aircraft reaches the current actual flight state parameters;

[0029] A control effect score for the tiltrotor aircraft to achieve the current actual flight state parameters is given based on each of the loop response curves in combination with the performance index requirements when designing the tiltrotor aircraft or the ADS-33EPRF requirements.

[0030] Optionally, after the parameter value is changed a preset number of times, the method further includes:

[0031] Determine whether all the control effect scores are equal to the target scores; if so, execute the step of “changing the type of the current flight control law”;

[0032] If not, after deleting all the first data links corresponding to the control effect scores that do not meet the standards, execute the step of "changing the type of the current flight control law".

[0033] Optionally, determining the flight strategy of the tiltrotor aircraft when performing a flight mission according to each of the first data links stored in the first database specifically includes:

[0034] The path between the starting point and the end point of the flight mission is divided into a plurality of sub-paths; each sub-path has two track points at both ends; all the track points are numbered according to the direction of the starting point and the end point of the flight mission;

[0035] According to the second expected flight state parameter corresponding to the i-th track point; i=1, 2, ..., n; n is the total number of track points;

[0036] randomly selecting the first data link from the first database according to the second expected flight state parameter, and controlling the tiltrotor aircraft to fly from a current position to the next track point in combination with the flight performance constraint information and the track point information;

[0037] Let i=i+1, and return to step "according to the second expected flight state parameter corresponding to the i-th track point" until the last track point is reached;

[0038] Scoring the completion of the flight mission according to the time cost, safety cost and fuel consumption cost of the flight mission, and constructing a second data link including the flight mission, data link 1, data link 2, ..., data link i, ..., and evaluation scores and storing the second data link in a second database;

[0039] Determine whether the number of the second data links is equal to a preset value, if not, return to the step of "according to the second expected flight state parameter corresponding to the i-th track point" until the number of the second data links is equal to the preset value;

[0040] If so, autonomous optimization is performed on all the second data links in the second database to obtain the optimal flight strategy for completing the flight mission.

[0041] Optionally, the flight modes include helicopter mode, tilt transition mode and airplane mode.

[0042] Optionally, the flight control law includes PID control law, fuzzy control, neural network control, optimal control of transition section, adaptive control based on model inversion, H ∞ Loop shaping control and model following control.

[0043] Optionally, the flight operations include hovering, hovering turns, centripetal turns, vertical maneuvers, acceleration and deceleration, and transient turns.

[0044] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0045] The present invention provides a method for training the flight control law of a tiltrotor aircraft in a wind tunnel, comprising initializing and self-checking a wind tunnel simulation device, setting the flight operation of the tiltrotor aircraft and the corresponding expected flight state parameters, flight mode, and flight control law under the current flight mode; using an inertial measurement unit to obtain the initial flight state parameters of the tiltrotor aircraft and calculate the error value with the expected flight state parameters, and judging whether the current error value is less than a preset value; calculating the control value update amount according to the error value and the current flight control law to adjust the flight state parameters to the target flight state parameters; calculating the error value again based on the target flight state parameters and adjusting the current flight state parameters until the current error value is less than the preset value; constructing a first data link including the expected flight state parameters, flight mode, flight control law, flight control law parameters, and control effect scores, and then obtaining the first data links under different flight operations, different flight modes, different flight control laws, and different flight control law parameters, and based on these first data links, it is possible to facilitate the execution of subsequent flight tasks. The present invention continuously adjusts the flight state until the expected instruction is reached based on the error determined based on the current actual flight state of the tiltrotor aircraft, solves the existing problem of error accumulation, and improves the accuracy of flight control law training. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 A schematic diagram of numerical simulation using MATLAB / Simulink software provided by the present invention;

[0048] Figure 2 A schematic diagram of a training working mode flow chart for wind tunnel-assisted training of tiltrotor aircraft flight control laws provided in Embodiment 1 of the present invention;

[0049] Figure 3 A schematic diagram of an iteration of a training working mode for wind tunnel-assisted training of a tiltrotor aircraft flight control law provided in Embodiment 1 of the present invention;

[0050] Figure 4 A system structure diagram of a wind tunnel assisted training flight control law provided in Example 1 of the present invention;

[0051] Figure 5 A schematic diagram of the flight mission track provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] The flight control law design process includes preliminary design, detailed design and simulation tests. The preliminary design is to determine the structure of the control law, the required feedback signal, etc.; the detailed design is to determine the pending parameters in the control law; the simulation test is to determine whether the design effect is met, otherwise repeat the above process. The existing technical means in the detailed design stage are to manually adjust the pending parameters of the control law according to the response dynamics. The existing technical means in the simulation verification stage are to use MATLAB / Simulink software for numerical simulation, and then output a large amount of data through wind tunnel tests, which are compared and verified with the corresponding theoretical data output by the simulation. If the design effect is not met, the numerical simulation and wind tunnel verification process are repeated, which will consume a lot of financial resources, time and manpower. The unique and complex operation of the tilt-rotor aircraft makes the entire cycle longer and manual adjustment more difficult;

[0054] There are three flight modes for tiltrotor aircraft. Different flight modes have different flight modes. Different flight modes require the design of different flight control laws for numerical simulation. The existing technical means for numerical simulation of the flight control law library built into the tiltrotor flight control system are to independently build mathematical models and perform numerical simulation under a given expected instruction. This is cumbersome and does not conform to the actual flight control system. Moreover, the flight control law after parameter adjustment is limited to achieving the ideal control effect for the expected instruction. There is a lack of integrated simulation verification of different modes of tiltrotor aircraft, a lack of flight control law parameter adjustment under the expected instructions of real flight missions, and a lack of simulation verification that is closer to the flight missions that the actual flight control system will face, which will lead to low confidence, poor control effect and poor maneuverability during actual flight.

[0055] Problems with existing technical means for designing flight control laws for tilt-rotor aircraft: manually adjusting control law parameters according to response curves, which results in a waste of manpower and time; independently performing numerical simulations in a MATLAB / Simulink environment to verify the stability of the tilt-rotor helicopter mode, fixed-wing aircraft mode and tilt-transition mode under different controllers under specific expected instructions, especially performing simulation analysis and optimization on the tilt-transition process to achieve better distribution of rotor and rudder control during the transition process, but lacking theoretical simulation of flight missions to be completed by actual flight control systems; after the simulation verification is completed, a verification aircraft is manufactured, the flight conditions are tested through a ground measurement and control system, and the flight quality is debugged and verified based on the collected data. If the design effect is not met, the above process is repeated, resulting in long cycles and high losses.

[0056] The present invention relates to the fields of wind tunnel test technology for tiltrotor aircraft, flight control law design for tiltrotor aircraft and artificial intelligence / data-driven technology, and in particular to a method for wind tunnel-assisted training of flight control laws for tiltrotor aircraft: for a tiltrotor aircraft flight control law library that has been preliminarily designed, when a wind tunnel simulation device simulates the actual flight state of the tiltrotor aircraft, a computer device has a built-in intelligent algorithm that replaces / assists the human brain in learning wind tunnel dynamic test data, extracts characteristic parameters of the flight control law and performs correction and optimization according to the control effect, and at the same time trains the tiltrotor aircraft autopilot flight strategy autonomous optimization function.

[0057] The purpose of the present invention is to provide a method for training the flight control law of a tiltrotor aircraft with the aid of a wind tunnel. The method uses a wind tunnel simulation device to train the flight control law of a tiltrotor aircraft, which can solve the problem of error accumulation, improve the accuracy of flight control law training, and provide a more accurate flight control law for the flight mission, thereby ensuring the efficiency of the flight mission. At the same time, it can also solve the problems of manpower and time consumption and the problems of long cycle and high loss, and improve the efficiency of flight control law training.

[0058] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0059] Example 1

[0060] This embodiment provides a method for wind tunnel-assisted training of a tiltrotor aircraft flight control law, comprising:

[0061] (I) First, the training process of the flight control law of this embodiment is briefly described in conjunction with the wind tunnel simulation device:

[0062] like Figure 2 and 3 As shown, the specific training process is as follows:

[0063] Step 1: Initialize self-check of wind tunnel simulation device.

[0064] like Figure 4 As shown, the wind tunnel simulation device includes: a wind tunnel laboratory, a six-degree-of-freedom motion mechanism, a tilt-rotor aircraft (including an autopilot), and a model support system (including a nine-channel servo actuator and an inertial measurement unit).

[0065] The tilt-rotor aircraft is located above the model support system, the six-degree-of-freedom motion mechanism is located below the model support system, and the semi-physical simulation device composed of the three is located in the wind tunnel laboratory.

[0066] The wind tunnel laboratory blows air in coordination with the flight status to assist the tilt-rotor aircraft in reaching the target flight state.

[0067] The six-degree-of-freedom motion mechanism cooperates with the flight state to drive the model support system to achieve three-axis angular motion and linear motion.

[0068] The autopilot built into the tiltrotor solves the control values ​​of the tiltrotor;

[0069] The model support system is equipped with a nine-channel servo steering engine to change the control value of the tilt-rotor aircraft, and the model support system is equipped with an inertial measurement unit to measure the actual flight state parameters of the tilt-rotor aircraft.

[0070] The wind tunnel simulation device can simulate the tiltrotor aircraft's take-off and landing, hovering, autonomous flight in a small range, as well as low-speed flight, tilt transition, and high-speed flight processes in all states.

[0071] The computer equipment includes a tiltrotor aircraft flight dynamics simulation system and a data processing system. The computer equipment and the wind tunnel simulation device constitute a system for wind tunnel-assisted training of flight control laws.

[0072] Among them, the tiltrotor flight dynamics simulation system calculates the flight state of the tiltrotor according to the control values ​​calculated by the tiltrotor's built-in autopilot in the wind tunnel simulation device.

[0073] Data processing system: classifies, stores and transmits the electrical signals of the tiltrotor aircraft's flight status to the six-degree-of-freedom motion mechanism and wind tunnel laboratory;

[0074] Classify and store the expected command signal given by the computer device and the actual flight status parameter signal of the tiltrotor aircraft, and transmit the error value between the two to the autopilot built in the tiltrotor aircraft;

[0075] The control values ​​of the tiltrotor aircraft calculated by the autopilot are classified and stored, and the control value electrical signals are transmitted to the nine-channel servo steering gear in the model support system;

[0076] The computer equipment and the wind tunnel simulation device can be connected via a data cable / radio frequency identification system (RFID), and the two form a method for wind tunnel-assisted training of tiltrotor aircraft flight control laws through data tagging, identification, interaction, operation and control.

[0077] Initialization self-test: Control the tiltrotor aircraft to the corresponding flight mode according to the selected flight mode, assign values ​​to the design variables of the associated flight control law, give control instructions for basic operations, such as altitude, speed and yaw control, etc., check the various loops of the tiltrotor aircraft flight control system, check the wind tunnel equipment, and measure the actual flight state parameters of the initial state.

[0078] The computer device sets the flight mode of the tilt-rotor aircraft: for example, in the helicopter mode, the computer device sends a control value electrical signal to drive the nine-channel servo steering engine in the wind tunnel simulation device to adjust the tilt-rotor aircraft to the helicopter mode;

[0079] The autopilot calls the flight control law: such as the PID control law, and the computer device sends an electrical signal to drive the autopilot to switch to the PID control law in helicopter mode;

[0080] Computer equipment sets flight operations: Computer equipment sets ADS-33EPRF operations (such as hovering, hovering turn, centripetal turn, vertical maneuver, acceleration and deceleration, transient turn, etc.), and the data processing system obtains the expected flight state according to the flight operation; among them, the expected flight state can be determined according to the flight operation, and the flight state parameters of the tilt-rotor aircraft are position, attitude, linear velocity and angular velocity. The expected flight state includes the expected position, expected attitude, expected linear velocity and expected angular velocity.

[0081] After the desired flight state is determined, the tiltrotor is controlled to change from the initial flight state to the desired flight state. When the desired flight state is reached, it means that the tiltrotor has completed the set flight operation. For example, if the flight operation is hovering, the corresponding desired flight state parameters are the state parameters in the hovering state. When the tiltrotor is controlled to change from the initial flight state to the desired state, it means that the tiltrotor has completed the hovering action. Similarly, if the flight operation is a hovering turn, the corresponding desired flight state parameters are the various state parameters in the hovering turn state. When the tiltrotor is controlled to change from the initial flight state to the desired state, it means that the tiltrotor has completed the hovering turn action.

[0082] The data processing system obtains the initial flight state parameters: the inertial measurement unit in the wind tunnel simulation device measures the flight state parameters of the tiltrotor aircraft at the current moment and feeds them back to the data transmission system.

[0083] Step 2: Start the semi-physical closed-loop simulation control formed by the wind tunnel simulation device and computer equipment:

[0084] The data processing system transmits the error value between the expected instruction and the initial flight state parameter to the autopilot. The flight control law of the autopilot calculates the control value. The computer device sends a control value electrical signal to drive the nine-channel servo steering gear in the wind tunnel simulation device to change the corresponding control amount of the tiltrotor aircraft. The tiltrotor aircraft flight dynamics simulation system calculates the flight state parameters at the next moment. The data processing system sends a flight state electrical signal to the six-degree-of-freedom motion mechanism and the wind tunnel laboratory. The six-degree-of-freedom motion mechanism drives the motion mechanism and the wind tunnel to cooperate in blowing air to make the tiltrotor aircraft reach the flight state. The inertial measurement unit in the wind tunnel simulation device measures the current tiltrotor aircraft. The flight state parameters of the rotorcraft are fed back to the data transmission system (updating the initial flight state parameters). The data processing system transmits the error value between the expected command and the current actual flight state parameters to the autopilot. The above process is repeated until the error value is within the allowable range. At this point, a cycle T is iterated. After the iteration, the data processing system outputs the response curve of each loop. The computer calls the algorithm to obtain the dynamic index, steady-state index value and ADS-33EPRF evaluation level of the response curve and gives the control effect score. The data processing system stores a data chain in the form of expected command-flight mode-flight control law-flight control law parameters-control effect score. Among them, the control effect score is obtained according to the requirements of the control system on the automatic control principle. The driving quality of the maneuvering flight can also be evaluated according to the bandwidth and phase delay time of the tilt-rotor aircraft. The evaluation level of the typical maneuvering action given by ADS-33EPRF can be obtained by computer programming. ADS-33EPRF: The bandwidth and phase delay time of a tiltrotor aircraft can be used to evaluate the quality of the maneuvering flight. ADS-33EPRF provides evaluation levels for typical maneuvers, such as hovering, hovering turns, centripetal turns, vertical maneuvers, acceleration and deceleration, transient turns, etc. A sample library of maneuvering operations is established for comparison and error correction.

[0085] The responses of the stability augmentation loop, attitude angle loop, altitude control loop and speed control loop of the tiltrotor flight control system are combined with the performance index requirements / ADS-33EPRF requirements given when designing the tiltrotor. The passing score is determined by the ability to complete the corresponding expected actions, and the full score is determined by the ability to maximize the tiltrotor design and complete the expected actions. The evaluation score is dynamically adjusted according to the control effect to supervise the wind tunnel assisted training process.

[0086] Step 3: The computer randomly changes the flight control law parameters and repeats step 2 N times. The data processing system stores N data links in the expert knowledge base.

[0087] The computer calls the intelligent algorithm to learn the process, extracts the characteristic parameters of the flight control law to predict the response, and trains the flight control law parameters according to the performance index requirements / ADS-33EPRF level requirements until the optimization termination conditions are met, eliminates the data links with substandard control effect scores in the expert knowledge base, and updates and develops data links with higher control effect scores. Characteristic parameters: Automatic adjustment of characteristic parameters usually requires repeated execution of the entire training process. The adjustment basis of the characteristic parameters is the prediction of the response by the intelligent algorithm, and the update and development basis of the characteristic parameters is the score of the control effect by the intelligent algorithm.

[0088] Step 4: The computer repeats steps 1, 2, and 3 until all flight control laws adapted to the three modes are trained under the expected instructions of all ADS-33EPRF operations.

[0089] Step 5: The computer sets new desired instructions and trains all flight control laws adapted to the three modes as above.

[0090] (II) Application of flight control laws obtained after training:

[0091] Step 1: The computer selects the flight mission, and the data processing system discretizes the flight mission into i segments, marking i+1 track point information, such as Figure 5 As shown, the computer inputs the offline environment of each track point, flight performance constraints and other constraint information;

[0092] Step 2: The computer gives an expected instruction according to the track point information labeled 1, and the expert knowledge base randomly selects a data link according to the expected instruction. The computer controls the tilt-rotor aircraft to fly from the current position to the track point position in combination with the offline constraint information and the track point information. The computer gives an expected instruction according to the track point information labeled 2, and repeats the above process until the tilt-rotor aircraft flies to the track point position labeled i+1. The computer calls an intelligent algorithm to give an evaluation score for the above flight mission in combination with the time cost, safety cost, fuel consumption cost, etc., and stores the above flight strategy in the form of flight mission-data link 1-…-data link i-evaluation score;

[0093] Step 3, repeat step 2 M times, the computer calls the intelligent algorithm (neural network, deep learning, deep reinforcement learning) to learn the process, intelligently change the flight control law characteristic parameters of the flight strategy and develop a new flight strategy, and train the expert knowledge base flight strategy autonomous optimization function according to performance index requirements / maneuvering flight criteria and real-time constraint information;

[0094] Step 4: The computer sets a new flight mission and trains the expert knowledge base as above.

[0095] In summary, the error value between the actual flight state parameter measured on the tiltrotor built in the wind tunnel simulation device and the expected flight state parameter given by the computer device is used as the input of the flight control law of the tiltrotor built in the autopilot, the flight control law solves the control amount and serves as the input of the tiltrotor built in the computer device's flight dynamics simulation system, the dynamics system calculates the theoretical flight state parameter and serves as the input of the wind tunnel simulation device, the wind tunnel simulation device simulates the flight state, the intelligent algorithm built in the computer device learns the wind tunnel test dynamic data, extracts characteristic parameters and trains the flight control law built in the autopilot, develops the trained data link and expands it to an expert knowledge base that can be used for autonomous optimization of flight strategies. Characteristic parameters: The intelligent algorithm automatically adjusts the characteristic parameters and usually needs to repeatedly perform the entire training process. The adjustment basis of the characteristic parameters is the prediction of the response by the intelligent algorithm, and the update and development basis of the characteristic parameters is the score of the control effect by the intelligent algorithm.

[0096] (III) The following describes the specific training scheme and specific application of the flight control law of this embodiment.

[0097] S1: Initialize the wind tunnel simulation device for self-checking, set the flight operation of the tiltrotor aircraft and the corresponding expected flight state parameters, flight mode, and flight control law under the current flight mode.

[0098] The flight modes include helicopter mode, tilt transition mode and airplane mode.

[0099] The flight control law includes PID control law, fuzzy control, neural network control, optimal control of transition section, adaptive control based on model inversion, H ∞ Loop shaping control and model following control.

[0100] The flight operations include hovering, hovering turns, centripetal turns, vertical maneuvers, acceleration and deceleration, and transient turns.

[0101] The computer equipment sequentially calls the expected instructions from the ADS-33EPRF operation library to train different flight control laws and different flight control law parameters under different flight modes. After training, the data link is used as the flight control law library, which is convenient for direct retrieval during subsequent flight missions.

[0102] Specifically, step S1 specifically includes:

[0103] Driving the nine-channel servo steering engine in the wind tunnel simulation device to adjust the tilt-rotor aircraft to the current flight mode;

[0104] Driving the autopilot in the tiltrotor aircraft to switch to the flight control law in the current flight mode;

[0105] The current flight operation of the tiltrotor aircraft is set, and the expected flight state parameters are obtained according to the current flight operation.

[0106] S2: using the inertial measurement unit in the wind tunnel simulation device to obtain initial flight state parameters of the tiltrotor aircraft; setting the initial flight state parameters as current actual flight state parameters.

[0107] S3: Calculate the error value between the expected flight state parameter and the current actual flight state parameter, and determine whether the current error value is less than a preset value.

[0108] S4: If not, drive the autopilot to calculate the control value update amount according to the current error value and the current flight control law.

[0109] S5: adjusting the flight state parameters of the tiltrotor aircraft to target flight state parameters corresponding to the current manipulation value update amount according to the current manipulation value update amount.

[0110] Specifically, step S5 specifically includes:

[0111] According to the current control value update amount, a nine-channel servo steering engine in the wind tunnel simulation device is driven to change the current control amount of the tilt-rotor aircraft;

[0112] The corresponding target flight state parameters are calculated according to the manipulation value update amount, and the six-degree-of-freedom motion mechanism in the wind tunnel simulation device is driven according to the calculation result, and the flight state parameters of the tilt-rotor aircraft are adjusted to the current target flight state parameters in cooperation with the wind tunnel laboratory blowing.

[0113] S6: Utilize the inertial measurement unit in the wind tunnel simulation device to obtain the current actual flight state parameters of the tiltrotor aircraft, and return to the step of "calculating the error value between the expected flight state parameters and the current actual flight state parameters, and determining whether the current error value is less than a preset value" until the current error value is less than the preset value.

[0114] S7: If so, determine the control effect score of the tiltrotor aircraft reaching the current actual flight state parameters, and construct a first data link including expected flight state parameters, flight mode, flight control law, flight control law parameters and control effect score; each data link corresponds to a process from an initial flight state (initial position) to an expected flight state (expected position).

[0115] The data link is generated after each iteration, in the form of expected command-flight mode-flight control law-flight control law parameters-control effect score. The iteration ends when the computer device gives the expected command signal, and the wind tunnel assists in training the tilt-rotor aircraft flight control law, iterating the actual flight state parameter value of the tilt-rotor aircraft from the initial value to the expected value, thus completing the response.

[0116] After each iteration, the response curves of the actual flight state parameters of the tilt-rotor aircraft and the animation simulation of the flight process of the tilt-rotor aircraft are obtained.

[0117] Specifically, determining the control effect score of the tiltrotor aircraft reaching the current actual flight state parameter specifically includes:

[0118] Outputting loop response curves of the stabilization loop, attitude angle loop, altitude control loop, and speed control loop when the tiltrotor aircraft reaches the current actual flight state parameters;

[0119] A control effect score for the tiltrotor aircraft to achieve the current actual flight state parameters is given based on each of the loop response curves in combination with the performance index requirements when designing the tiltrotor aircraft or the ADS-33EPRF requirements.

[0120] S8: Randomly change the parameter value of the flight control law, and make the flight control law after the parameter change the current flight control law, and return to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter" until the number of parameter value changes reaches the preset number.

[0121] When changing the flight control law parameters, considering that some parameters may have little effect on the control effect, in order to select parameters that are more relevant to the control effect, the characteristic parameters of the flight control law can be extracted, that is, the parameters that will affect the control effect. Subsequently, only the selected characteristic parameters can be randomly changed, thereby improving the training efficiency.

[0122] When the parameter value is changed a preset number of times, the following step is also included:

[0123] Determine whether all the control effect scores are equal to the target scores; if so, execute the step of “changing the type of the current flight control law”;

[0124] If not, after deleting all the first data links corresponding to the control effect scores that do not meet the standards, execute the step of "changing the type of the current flight control law".

[0125] S9: Change the type of the current flight control law, and make the flight control law after the type change the flight control law under the current target flight mode, return to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter, and determining whether the current error value is less than the preset value", until all types of the flight control laws under the current target flight mode are traversed.

[0126] S10: Change the type of the current flight mode, make the changed flight mode the current target flight mode, and return to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter, and determining whether the current error value is less than a preset value", until all the flight control laws under all types of the target flight modes are traversed; until all flight mode types are traversed.

[0127] S11: Change the current flight operation and the current expected flight state parameters, make the changed flight operation the current flight operation, and the changed expected flight state parameters the current expected flight state parameters, and return to the step of "calculating the error value between the expected flight state parameters and the current actual flight state parameters, and determining whether the current error value is less than the preset value" until all flight operation types are traversed.

[0128] S12: storing all the flight operations, all the flight modes, all the flight control laws and all the first data links under the parameters of the flight control laws in a first database.

[0129] S13: The first data links stored in the first database determine the flight strategy of the tiltrotor aircraft when performing a flight mission. This step can also be simulated using a wind tunnel simulation experiment device to obtain a database storing flight strategies under various circumstances, which is convenient for subsequent direct retrieval and application.

[0130] The expert knowledge base (second database) stores a large number of data links that meet the control requirements, and can intelligently match the optimal data link according to the requirements of each flight track of the flight mission, that is, it has the function of autonomous optimization of the flight strategy of the tilt-rotor aircraft autopilot. For example, it can autonomously optimize and use genetic algorithms to select the best individual (data link) from a large number of data links that meet the control requirements.

[0131] The specific step S13 specifically includes:

[0132] The path between the starting point and the end point of the flight mission is divided into a plurality of sub-paths; each sub-path has two track points at both ends; all the track points are numbered according to the direction of the starting point and the end point of the flight mission;

[0133] According to the second expected flight state parameter corresponding to the i-th track point; i=1, 2, ..., n; n is the total number of track points;

[0134] randomly selecting the first data link from the first database according to the second expected flight state parameter, and controlling the tiltrotor aircraft to fly from a current position to the next track point in combination with the flight performance constraint information and the track point information;

[0135] Let i=i+1, and return to step "according to the second expected flight state parameter corresponding to the i-th track point" until the last track point is reached;

[0136] Scoring the completion of the flight mission according to the time cost, safety cost and fuel consumption cost of the flight mission, and constructing a second data link including the flight mission, data link 1, data link 2, ..., data link i, ..., and evaluation scores and storing the second data link in a second database;

[0137] Determine whether the number of the second data links is equal to a preset value, if not, return to the step of "according to the second expected flight state parameter corresponding to the i-th track point" until the number of the second data links is equal to the preset value;

[0138] If so, autonomous optimization is performed on all the second data links in the second database to obtain the optimal flight strategy for completing the flight mission.

[0139] In this embodiment, after the computer device gives the expected command signal, the tiltrotor aircraft completes the control from the initial flight state to the expected flight state under the action of a certain flight control law, randomly changes the flight control law parameters and repeats the control process N times, stores N data links, and the intelligent algorithm learns the response curves of the N data links, mines the laws and obtains experience or knowledge, that is, extracts the flight control law characteristic parameters. The intelligent algorithm corrects and optimizes the flight control law characteristic parameters according to the ADS-33EPRF level requirements, and trains the flight control law characteristic parameters on this basis until the optimization termination conditions are met, and updates the data links to the expert knowledge base according to the reward and punishment evaluation development. In this way, the flight control law built into the tiltrotor autopilot is trained, and the optimal solution of the characteristic parameters of the flight control law adapted in different flight modes under different expected instructions; after the computer equipment gives the flight mission, combined with the track information of each segment, environmental constraints and flight performance constraints, the expert knowledge base randomly matches the data link for each segment of the track to control the tiltrotor to complete the entire flight mission, and repeats this process M times. The intelligent algorithm learns the flight strategy of the tiltrotor and trains the flight strategy optimization algorithm with the maximum performance advantage of the designed tiltrotor as the reward and punishment judgment, so as to train the autonomous optimization function of the flight strategy of the tiltrotor autopilot.

[0140] The actual flight state parameters measured by the wind tunnel simulation device are used instead of the theoretical flight state parameters to reduce the error caused by the feedback parameter part. Computer equipment is connected to the wind tunnel simulation device to learn wind tunnel test data and train the tilt-rotor aircraft flight control law. The three independent tasks are integrated into a wind tunnel laboratory for iteration and integration to complete the flight control law training of different modes, shortening the cycle while reducing losses. Intelligent algorithms are used to train the autopilot flight strategy to autonomously optimize the function.

[0141] The solution implemented in this paper has the following technical effects:

[0142] First, the error accumulation in the design of the tiltrotor flight control law is reduced, which is mainly reflected in: the wind tunnel and the tiltrotor semi-physical simulation device form a wind tunnel simulation device, which can simulate the flight state of the tiltrotor in a real environment, and measure the actual flight state parameters from the tiltrotor, instead of the theoretical flight state parameters output by the MATLAB / Simulink flight dynamics model based on a large number of assumptions in the prior art, which are used as feedback parameters and the expected flight instructions of the computer device are subtracted as the input of the flight control law, which effectively reduces the error accumulation between theory and actual flight and improves the confidence of the tiltrotor flight control law in the actual flight system.

[0143] Second, it shortens the iteration cycle of the flight control law of the tiltrotor aircraft from theory to practice, effectively reducing the consumption of manpower, financial resources and time, which is mainly reflected in: integrating the flight control law design process of manual control law parameter adjustment, MATLAB / Simulink simulation verification and field test flight verification into a wind tunnel assisted training tiltrotor aircraft flight control law system and method formed by "wind tunnel simulation device simulating the actual flight state of the tiltrotor aircraft, artificial intelligence technology / data-driven technology learning wind tunnel dynamic test data, computer equipment and wind tunnel simulation device closed-loop iterative training", shortening the iteration cycle, and using computer equipment instead of manual control law parameter adjustment and parameter optimization, and completing simulation verification and test flight verification in parallel, developing the data chain of expected instructions-flight mode-flight control law-flight control law parameters-control effect score and expanding it to the expert knowledge base, effectively reducing losses and improving the flight quality of the tiltrotor aircraft, not only meeting the ADS-33EPRF level requirements, but also giving full play to its maximum performance advantages.

[0144] Thirdly, the confidence and mission effectiveness of the tiltrotor aircraft's flight control law are improved, which is mainly reflected in: the wind tunnel simulation device and the computer equipment form a closed loop, integrating and completing the flight control law training of different modes, using intelligent algorithms to autonomously learn and train the flight control laws under different flight missions, and at the same time training the autopilot's flight strategy autonomous optimization function. It can perform global flight strategy optimization for the entire trajectory of an offline flight mission, and can also perform local flight strategy optimization for a real-time flight mission, thereby improving the mission effectiveness of the flight control system, and using intelligent algorithms to optimize the flight control law parameters to maximize the performance advantages of the designed tiltrotor aircraft. The expert knowledge base can spontaneously activate the optimal flight strategy for different flight missions. After continuous pure intelligent training, when the flight control law library is updated, there is no need to repeat the above-mentioned parameter adjustment training and the simulation results can still be accurate.

[0145] Compared with the flight control law design method of MATLAB / Simulink software numerical simulation and field flight verification, the above-mentioned semi-physical simulation method, which relies on artificial intelligence technology / data-driven technology and wind tunnel experiment technology, has high confidence, short cycle, and low loss. It has simulation verification of flight missions that real flight control systems will face, and can switch to the relatively optimal controller according to the adaptability of a certain maneuver operation, effectively improving the flight performance of the tilt-rotor aircraft.

[0146] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for wind tunnel assisted training of tiltrotor aircraft flight control law, characterized in that: include: Initialize and self-check the wind tunnel simulation device, set the flight operation of the tilt-rotor aircraft and the corresponding expected flight state parameters, flight mode, and flight control law under the current flight mode; Using the inertial measurement unit in the wind tunnel simulation device to obtain the initial flight state parameters of the tiltrotor aircraft; setting the initial flight state parameters as the current actual flight state parameters; Calculating an error value between the expected flight state parameter and the current actual flight state parameter, and determining whether the current error value is less than a preset value; If not, the autopilot is driven to calculate an update amount of the control value according to the current error value and the current flight control law; Adjusting the flight state parameter of the tiltrotor aircraft to a target flight state parameter corresponding to the current manipulation value update amount according to the current manipulation value update amount; Using the inertial measurement unit in the wind tunnel simulation device to obtain the current actual flight state parameter of the tiltrotor aircraft, and returning to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter, and determining whether the current error value is less than a preset value", until the current error value is less than the preset value; If yes, determining the control effect score of the tiltrotor aircraft to achieve the current actual flight state parameter, and constructing a first data link including the expected flight state parameter, the flight mode, the flight control law, the flight control law parameter and the control effect score; Randomly changing the parameter value of the flight control law, and making the flight control law after the parameter change the current flight control law, and returning to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter" until the number of parameter value changes reaches a preset number; Changing the type of the current flight control law, and setting the flight control law after the type change as the flight control law under the current target flight mode, returning to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter, and determining whether the current error value is less than a preset value", until all types of the flight control laws under the current target flight mode are traversed; Changing the type of the current flight mode, setting the changed flight mode as the current target flight mode, and returning to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter, and determining whether the current error value is less than a preset value", until all the flight control laws under all types of the target flight modes are traversed; Until all flight mode types are traversed; Change the current flight operation and the current expected flight state parameter, set the changed flight operation as the current flight operation, and the changed expected flight state parameter as the current expected flight state parameter, and return to the step of "calculating the error value between the expected flight state parameter and the current actual flight state parameter, and determining whether the current error value is less than a preset value", until all flight operation types are traversed; storing each of the first data links under all the flight operations, all the flight modes, all the flight control laws and all the parameters of the flight control laws in a first database; Determining a flight strategy for the tiltrotor aircraft when performing a flight mission according to each of the first data links stored in the first database; Wherein, adjusting the flight state parameter of the tiltrotor aircraft to a target flight state parameter corresponding to the current manipulation value update amount according to the current manipulation value update amount specifically includes: According to the current control value update amount, a nine-channel servo steering engine in the wind tunnel simulation device is driven to change the current control amount of the tilt-rotor aircraft; The corresponding target flight state parameters are calculated according to the manipulation value update amount, and the six-degree-of-freedom motion mechanism in the wind tunnel simulation device is driven according to the calculation result, and the flight state parameters of the tilt-rotor aircraft are adjusted to the current target flight state parameters in cooperation with the wind tunnel laboratory blowing.

2. The method according to claim 1, characterized in that The initialization self-check of the wind tunnel simulation device, setting the flight operation of the tiltrotor aircraft and the corresponding expected flight state parameters, flight mode, and flight control law under the current flight mode specifically includes: Driving the nine-channel servo steering engine in the wind tunnel simulation device to adjust the tilt-rotor aircraft to the current flight mode; Driving the autopilot in the tiltrotor aircraft to switch to the flight control law in the current flight mode; The current flight operation of the tiltrotor aircraft is set, and the expected flight state parameters are obtained according to the current flight operation.

3. The method according to claim 1, characterized in that Determining the control effect score of the tiltrotor aircraft reaching the current actual flight state parameter specifically includes: Outputting loop response curves of the stabilization loop, attitude angle loop, altitude control loop, and speed control loop when the tiltrotor aircraft reaches the current actual flight state parameters; A control effect score for the tiltrotor aircraft to achieve the current actual flight state parameters is given based on each of the loop response curves in combination with the performance index requirements when designing the tiltrotor aircraft or the ADS-33EPRF requirements.

4. The method according to claim 1, characterized in that After the parameter value is changed a preset number of times, the following further includes: Determine whether all the control effect scores are equal to the target scores; if so, execute the step of "changing the type of the current flight control law"; If not, after deleting all the first data links corresponding to the control effect scores that do not meet the standards, execute the step of "changing the type of the current flight control law".

5. The method according to claim 1, characterized in that Determining the flight strategy of the tiltrotor aircraft when performing a flight mission according to each of the first data links stored in the first database specifically includes: The path between the starting point and the end point of the flight mission is divided into a plurality of sub-paths; each sub-path has two track points at both ends; all the track points are numbered according to the direction of the starting point and the end point of the flight mission; According to the second expected flight state parameter corresponding to the i-th track point; i=1, 2, ..., n; n is the total number of track points; randomly selecting the first data link from the first database according to the second expected flight state parameter, and controlling the tiltrotor aircraft to fly from a current position to the next track point in combination with the flight performance constraint information and the track point information; Let i=i+1, and return to step "according to the second expected flight state parameter corresponding to the i-th track point" until the last track point is reached; Scoring the completion of the flight mission according to the time cost, safety cost and fuel consumption cost of the flight mission, and constructing a second data link including the flight mission, data link 1, data link 2, ..., data link i, ..., and evaluation scores and storing the second data link in a second database; Determine whether the number of the second data links is equal to a preset value, and if not, return to step "according to the second expected flight state parameter corresponding to the i-th track point" until the number of the second data links is equal to the preset value; If so, autonomous optimization is performed on all the second data links in the second database to obtain the optimal flight strategy for completing the flight mission.

6. The method according to claim 1, characterized in that The flight modes include helicopter mode, tilt transition mode and airplane mode.

7. The method according to claim 1, characterized in that The flight control law includes PID control law, fuzzy control, neural network control, optimal control of transition section, adaptive control based on model inversion, H ∞ Loop shaping control and model following control.

8. The method according to claim 1, characterized in that The flight operations include hovering, hovering turns, centripetal turns, vertical maneuvers, acceleration and deceleration, and transient turns.

Images