Autonomous flight attitude control method and system for UAV based on inertial navigation

Abstract

This invention discloses a method and system for autonomous flight attitude control of unmanned aerial vehicles (UAVs) based on inertial navigation, belonging to the field of UAV technology. The method includes the following steps: acquiring the UAV's current remaining battery power, external environment complexity data, and associated deviation data; generating a battery power contraction intensity signal, an environment contraction intensity signal, and a memory contraction intensity signal; fusing them into a final interval contraction command; acquiring and adjusting the dynamic flight attitude error confidence interval in real time; acquiring the real-time flight attitude error; and performing attitude correction if the real-time flight attitude error exceeds the boundary of the confidence interval. This invention achieves controllable long-endurance pure inertial navigation error by real-time contraction of the confidence interval through a triple linkage of battery power, environment, and memory, and by triggering attitude correction in advance. It solves the problem in existing technologies where the inability to dynamically assess error threats and passively correct them leads to mission failure.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) technology, and in particular to a method and system for controlling the autonomous flight attitude of UAVs based on inertial navigation. Background Technology

[0002] With the widespread application of unmanned aerial vehicles (UAVs) in civilian inspection, emergency rescue, and military reconnaissance missions, higher demands are being placed on autonomous flight capabilities that are characterized by "long endurance, independence from external signals, and controllable errors." Inertial navigation systems (INS), due to their independence from external signals and high short-term accuracy, have become the core navigation method for autonomous UAV flight.

[0003] Existing technologies typically employ a loosely or tightly coupled approach of "inertial navigation + satellite / vision / magnetism": GNSS periodically corrects accumulated INS errors, while vision or magnetism provides auxiliary attitude constraints. However, in complex scenarios such as GNSS signal obstruction or interference, loss of visual features, or abrupt changes in the magnetic environment, external correction sources fail, and the errors of pure inertial navigation accumulate exponentially over time, leading to a chain reaction of risks: attitude drift - track deviation - mission failure.

[0004] Existing technologies lack a dynamic judgment mechanism to determine when an error has threatened the current mission. They cannot adjust the error tolerance in real time based on remaining battery power, environmental threats, and historical experience. In other words, the system can sense the accumulation of errors, but lacks a dynamic decision-making mechanism to determine whether the error has threatened the current flight mission. Summary of the Invention

[0005] This application provides a method and system for autonomous flight attitude control of unmanned aerial vehicles based on inertial navigation. It solves the problem in the prior art that it is impossible to dynamically assess the threat level of pure inertial navigation errors to the mission. It realizes the adaptive contraction of the reliable interval and the early attitude correction under the triple linkage of power, environment and memory, and ensures the safe completion of long-endurance missions.

[0006] This application provides a method for autonomous flight attitude control of a UAV based on inertial navigation, including: acquiring the current remaining battery power of the UAV, external environment complexity data, and related deviation data extracted from the historical flight memory database;

[0007] Based on the remaining energy data of the organism, the complexity data of the external environment, and the associated deviation data, the energy contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal are generated.

[0008] The steps for generating the electrical shrinkage intensity signal, the environmental shrinkage intensity signal, and the memory shrinkage intensity signal include:

[0009] The remaining battery power data of the body is converted into a battery contraction intensity signal using an S-shaped response function;

[0010] The S-shaped response function is:

[0011] ;

[0012] In the formula, For electrical contraction intensity, This represents the current remaining battery power of the device. For the collective critical energy threshold, The parameter represents the position of the battery inflection point. This is the power sensitivity coefficient. It is a natural constant;

[0013] The external environment complexity data is constructed into a multi-dimensional environment state vector, and the degree of deviation between the multi-dimensional environment state vector and the preset safe environment vector is obtained by the multi-dimensional distance calculation formula. The degree of deviation is used as the environmental contraction intensity signal.

[0014] The steps for calculating the deviation between the multidimensional environment state vector and the preset safe environment vector using the multidimensional distance calculation formula include:

[0015] Wind shear intensity, geomagnetic disturbance, and satellite loss-lock rate values ​​were extracted from external environmental complexity data.

[0016] The wind shear intensity value, geomagnetic disturbance value, and satellite loss-lock rate value are combined to form a measured vector;

[0017] A reference vector is formed by constructing a reference vector from the corresponding reference values ​​in the preset safe environment vector. The degree of deviation is obtained by the multidimensional distance calculation formula, and the degree of deviation is used as the environmental shrinkage intensity signal.

[0018] The multidimensional distance calculation formula is as follows:

[0019] ;

[0020] In the formula, This is an environmental shrinkage intensity signal. The measured vector is the first Each component value The first reference vector Each component value For the first The normalized coefficients of each parameter, For the first Weighting factors of each parameter

[0021] The current remaining battery power data and external environment complexity data are input into the deviation prediction model based on the historical flight memory bank. The deviation prediction model outputs the predicted deviation value and uses the predicted deviation value as the memory contraction intensity signal.

[0022] The electrical contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal are fused together to form the final interval contraction command;

[0023] Based on the final interval contraction command, the reliable interval of dynamic flight attitude error is acquired and adjusted in real time.

[0024] The real-time flight attitude error calculated by the UAV's inertial navigation function is acquired in real time and compared with the confidence interval.

[0025] If the real-time flight attitude error exceeds the boundary of the confidence interval, an attitude correction operation is performed.

[0026] Furthermore, the steps of inputting the current remaining battery power data and external environmental complexity data into the deviation prediction model established based on the historical flight memory bank, and outputting the predicted deviation value through the deviation prediction model include:

[0027] The remaining battery power, external environment complexity, and corresponding attitude error peak values ​​are extracted from the historical flight memory database.

[0028] The training dataset is composed of the stored remaining battery power data, external environment complexity data, and corresponding attitude error peak values.

[0029] A Gaussian process regression model is established for the training dataset, where the input is the remaining battery power of the machine and the complexity of the external environment, and the output is the peak value of the attitude error;

[0030] The obtained data on the current remaining battery power of the organism and the complexity of the external environment are used as query points to input into the Gaussian process regression model, and the predicted deviation value is output as the memory contraction intensity signal.

[0031] Furthermore, the step of fusing the electrical contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal into the final interval contraction command includes:

[0032] The values ​​of the electrical shrinkage intensity signal, the environmental shrinkage intensity signal, and the memory shrinkage intensity signal are sorted from largest to smallest;

[0033] Based on the sorted values ​​of the electrical contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal, the final interval contraction command is obtained by performing calculations using a preset non-additive aggregation function.

[0034] The non-additive aggregation function is:

[0035] ;

[0036] In the formula, This is the final interval contraction instruction. , , The ordered contraction intensity signal satisfies... , As the dominant factor, and ,in As the dominant sensitivity coefficient, This is the preset dominant threshold.

[0037] Furthermore, the steps for acquiring and adjusting the confidence interval of dynamic flight attitude error in real time include:

[0038] The width of the confidence interval of the dynamic flight attitude error is obtained by using a preset exponential decay function;

[0039] The exponential decay function is:

[0040] ;

[0041] In the formula, The width of the confidence interval for dynamic flight attitude error. The minimum confidence interval width is preset. The preset initial confidence interval width, The attenuation intensity coefficient, This is the final interval contraction instruction. It is a natural constant.

[0042] Furthermore, the steps for obtaining the real-time flight attitude error calculated by the UAV's inertial navigation function include:

[0043] Real-time acquisition of the drone's angular velocity and specific force measurements;

[0044] The real-time angular velocity and specific force measurements are input into the strapdown inertial navigation algorithm to obtain the current flight attitude of the UAV in real time.

[0045] The current flight attitude is compared with the initial flight attitude of the UAV, which is the flight attitude recorded when the UAV enters the inertial navigation mode;

[0046] The difference between the current flight attitude and the initial flight attitude of the UAV is taken as the real-time flight attitude error.

[0047] Furthermore, if the real-time flight attitude error exceeds the boundary of the confidence interval, the steps for performing attitude correction operations include:

[0048] Obtain the difference between the real-time flight attitude error and the boundary value of the confidence interval that it exceeds, and determine the difference as the over-limit error;

[0049] The excess error is calculated with a preset correction gain coefficient to generate a correction control command;

[0050] The correction control command is input to the attitude control actuator of the UAV to adjust the flight attitude of the UAV until the real-time flight attitude error returns to the confidence range of the dynamic flight attitude error.

[0051] This application provides an inertial navigation-based autonomous flight attitude control system for unmanned aerial vehicles (UAVs), which is used to implement an inertial navigation-based autonomous flight attitude control method for UAVs, including: a data acquisition module, a signal generation module, a command generation module, a confidence interval acquisition module, an attitude error comparison module, and an attitude correction module.

[0052] The data acquisition module is used to acquire the current remaining battery power of the UAV, the complexity of the external environment, and the associated deviation data extracted from the historical flight memory database.

[0053] The signal generation module is used to generate a power contraction intensity signal, an environmental contraction intensity signal, and a memory contraction intensity signal based on the remaining power data of the body, the complexity data of the external environment, and the associated deviation data.

[0054] The instruction generation module is used to fuse the electrical contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal into a final interval contraction instruction.

[0055] The confidence interval acquisition module is used to acquire and adjust the confidence interval of dynamic flight attitude error in real time based on the final interval contraction command.

[0056] The attitude error comparison module is used to acquire the real-time flight attitude error calculated by the UAV's inertial navigation function in real time and compare it with the confidence interval.

[0057] The attitude correction module is used to perform attitude correction operations if the real-time flight attitude error exceeds the boundary of the confidence interval.

[0058] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0059] 1. By generating a dynamic reliable interval through the coupling of three signals: power, environment, and memory, the attitude error boundary is adaptively tightened based on real-time risks, thereby achieving controllable error accumulation during long-term inertial navigation flights. This effectively solves the problem in existing technologies that cannot dynamically determine whether errors threaten the current mission.

[0060] 2. By using a non-additive aggregation decision model, a strong dominant response is automatically triggered when any risk factor increases sharply, while conservative decisions from multiple parties are retained when the risks are balanced. This solves the problem that traditional linear fusion cannot adapt to emergency situations and eliminates the risk of flight loss of control caused by misjudgment of a single risk.

[0061] 3. By mapping the interval width to the contraction command one by one through the exponential decay function, the elastic boundary between the minimum and maximum is output in real time, thereby realizing precise control of the attitude correction triggering time. Attached Figure Description

[0062] Figure 1 A flowchart of an inertial navigation-based unmanned aerial vehicle (UAV) attitude control method provided in an embodiment of this application;

[0063] Figure 2 A schematic diagram of the structure of an inertial navigation-based unmanned aerial vehicle (UAV) autonomous flight attitude control system provided in an embodiment of this application. Detailed Implementation

[0064] This application provides a method and system for controlling the attitude of an unmanned aerial vehicle (UAV) in autonomous flight based on inertial navigation. This solves the problem that pure inertial navigation errors accumulate over time and lack a dynamic risk assessment mechanism in the prior art. By fusing the three signals of power, environment, and memory, the reliable interval is shrunk in real time and attitude correction is triggered in advance, thus achieving controllable error and safe mission completion under long-endurance conditions without external assistance.

[0065] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0066] like Figure 1 The diagram shown is a flowchart of an autonomous flight attitude control method for unmanned aerial vehicles (UAVs) based on inertial navigation, provided in an embodiment of this application. This method is applied to an autonomous flight attitude control system for UAVs based on inertial navigation and includes the following steps: acquiring the remaining battery power of the UAV, the complexity of the external environment, and the associated deviation data extracted from the historical flight memory bank.

[0067] Based on the remaining energy data of the organism, the complexity data of the external environment, and the associated deviation data, quantification processing is performed to generate energy contraction intensity signal, environmental contraction intensity signal, and memory contraction intensity signal.

[0068] The steps for generating the electrical shrinkage intensity signal, the environmental shrinkage intensity signal, and the memory shrinkage intensity signal include:

[0069] The remaining battery power data of the body is converted into a battery power contraction intensity signal by an S-shaped response function. This function makes the value of the battery power contraction intensity signal increase non-linearly as the battery power data decreases.

[0070] The S-shaped response function is:

[0071] ;

[0072] In the formula, The intensity of electrical contraction (0 = no effect, 1 = maximum effect). The current remaining electrical charge of the machine (unit: joules). The collective critical electrical threshold (unit: joules). The parameter is the position of the inflection point of the charge (dimensionless). The power sensitivity coefficient (unit: J⁻¹) Let be a natural constant, when The signal surges, enabling a low-battery priority response;

[0073] The external environment complexity data is constructed into a multi-dimensional environment state vector, and the degree of deviation between the multi-dimensional environment state vector and the preset safe environment vector is obtained by the multi-dimensional distance calculation formula. The degree of deviation is used as the environmental contraction intensity signal.

[0074] The steps for calculating the deviation between the multidimensional environment state vector and the preset safe environment vector using the multidimensional distance calculation formula include:

[0075] Wind shear intensity, geomagnetic disturbance, and satellite loss-lock rate values ​​were extracted from external environmental complexity data.

[0076] The wind shear intensity value, geomagnetic disturbance value, and satellite loss-lock rate value are combined to form a measured vector;

[0077] A reference vector is formed by constructing a reference vector from the corresponding reference values ​​in the preset safe environment vector. The degree of deviation is obtained by the multidimensional distance calculation formula, and the degree of deviation is used as the environmental shrinkage intensity signal.

[0078] The multidimensional distance calculation formula is as follows:

[0079] ;

[0080] In the formula, This is the environmental shrinkage intensity signal (dimensionless). The measured vector is the first Each component value specifically includes: The measured value of wind shear intensity (unit: m / s²) The values ​​represent measured geomagnetic disturbances (unit: μT). The measured value of satellite lock-off rate (unit: %). The first reference vector Each component value specifically includes: The preset wind shear intensity safety threshold (unit: m / s²) is used. Preset geomagnetic disturbance safety threshold (unit: μT). Preset satellite lock-off rate safety threshold (unit: %). For the first The normalized coefficients of each parameter include: The range of wind shear intensity (unit: m / s²). The range of geomagnetic disturbance is given in μT. Satellite lock-off rate range (unit: %). For the first Weighting factors for each parameter;

[0081] The current remaining battery power data and external environment complexity data are input into the deviation prediction model based on the historical flight memory bank. The deviation prediction model outputs the predicted deviation value and uses the predicted deviation value as the memory contraction intensity signal.

[0082] The electrical contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal are fused together to form the final interval contraction command;

[0083] Based on the final interval contraction command, the reliable interval of dynamic flight attitude error is acquired and adjusted in real time.

[0084] The real-time flight attitude error calculated by the UAV's inertial navigation function is acquired in real time and compared with the confidence interval.

[0085] If the real-time flight attitude error exceeds the boundary of the confidence interval, then an attitude correction operation is performed.

[0086] After performing attitude correction operations or at critical points in the flight mission, the current battery level, environmental status, and corresponding attitude error are updated to the historical flight memory database.

[0087] Furthermore, the steps of inputting the current remaining battery power data and external environmental complexity data into the deviation prediction model established based on the historical flight memory bank, and outputting the predicted deviation value through the deviation prediction model include:

[0088] The remaining battery power, external environment complexity, and corresponding attitude error peak values ​​are extracted from the historical flight memory database.

[0089] The training dataset is composed of the stored remaining battery power data, external environment complexity data, and corresponding attitude error peak values.

[0090] A Gaussian process regression model is established for the training dataset, where the input is the remaining battery power of the machine and the complexity of the external environment, and the output is the peak value of the attitude error;

[0091] The obtained data on the current remaining battery power of the organism and the complexity of the external environment are used as query points to input into the Gaussian process regression model, and the predicted deviation value is output as the memory contraction intensity signal.

[0092] Furthermore, the step of fusing the electrical contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal into the final interval contraction command includes:

[0093] The values ​​of the electrical shrinkage intensity signal, the environmental shrinkage intensity signal, and the memory shrinkage intensity signal are sorted from largest to smallest;

[0094] Based on the sorted values ​​of the electrical contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal, the final interval contraction command is obtained by performing calculations using a preset non-additive aggregation function.

[0095] The non-additive aggregation function is:

[0096] ;

[0097] In the formula, This is the final interval contraction instruction (dimensionless). , , The ordered contraction intensity signal satisfies... , As the dominant factor, and ,in As the dominant sensitivity coefficient, The preset dominant threshold;

[0098] The larger the value of any of the signals, the greater its contribution to the final interval contraction command value.

[0099] Furthermore, the steps for acquiring and adjusting the confidence interval of dynamic flight attitude error in real time include:

[0100] The width of the confidence interval of the dynamic flight attitude error is obtained by using a preset exponential decay function;

[0101] The exponential decay function is:

[0102] ;

[0103] In the formula, The width of the confidence interval for dynamic flight attitude error (unit: degrees). The preset minimum confidence interval width (unit: degrees). The preset initial confidence interval width (unit: degrees). This is the attenuation intensity coefficient (dimensionless, preset value > 0). This is the final interval contraction instruction. It is a natural constant;

[0104] The input to the exponential decay function is the final interval contraction command. The output of the exponential decay function is inversely proportional to the input, that is, the larger the value of the final interval contraction command, the smaller the calculated reliable interval width.

[0105] Furthermore, the steps for obtaining the real-time flight attitude error calculated by the UAV's inertial navigation function include:

[0106] Real-time acquisition of the drone's angular velocity and specific force measurements;

[0107] The real-time angular velocity and specific force measurements are input into the strapdown inertial navigation algorithm to obtain the current flight attitude of the UAV in real time.

[0108] The specific steps are as follows:

[0109] Substitute the real-time angular velocity into the quaternion update formula to calculate the quaternion at the current moment;

[0110] Convert the current quaternion into a direction cosine matrix;

[0111] The velocity increment is obtained by integrating the direction cosine matrix and real-time specific force measurement in the Earth coordinate system.

[0112] The current speed is obtained by adding the speed increment to the speed at the previous moment;

[0113] Integrate the current speed to obtain the current position;

[0114] The current roll angle, pitch angle, and yaw angle are calculated based on the current quaternion to form the current flight attitude.

[0115] The current flight attitude is compared with the initial flight attitude of the UAV, which is the flight attitude recorded when the UAV enters the inertial navigation mode;

[0116] The difference between the current flight attitude and the initial flight attitude of the UAV is taken as the real-time flight attitude error.

[0117] Furthermore, if the real-time flight attitude error exceeds the boundary of the confidence interval, the steps for performing attitude correction operations include:

[0118] Obtain the difference between the real-time flight attitude error and the boundary value of the confidence interval that it exceeds, and determine the difference as the over-limit error;

[0119] The excess error is calculated with a preset correction gain coefficient to generate a correction control command;

[0120] The correction control command is input to the attitude control actuator of the UAV to adjust the flight attitude of the UAV until the real-time flight attitude error returns to the confidence range of the dynamic flight attitude error.

[0121] like Figure 2 The diagram shown is a structural schematic of an inertial navigation-based UAV autonomous flight attitude control system provided in an embodiment of this application. It is used to implement the inertial navigation-based UAV autonomous flight attitude control method. The inertial navigation-based UAV autonomous flight attitude control system provided in this embodiment includes: a data acquisition module, a signal generation module, a command generation module, a confidence interval acquisition module, an attitude error comparison module, and an attitude correction module.

[0122] The data acquisition module is used to acquire the current remaining battery power of the UAV, the complexity of the external environment, and the associated deviation data extracted from the historical flight memory database.

[0123] The signal generation module is used to generate a power contraction intensity signal, an environmental contraction intensity signal, and a memory contraction intensity signal based on the remaining power data of the body, the complexity data of the external environment, and the associated deviation data.

[0124] The instruction generation module is used to fuse the electrical contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal into a final interval contraction instruction.

[0125] The confidence interval acquisition module is used to acquire and adjust the confidence interval of dynamic flight attitude error in real time based on the final interval contraction command.

[0126] The attitude error comparison module is used to acquire the real-time flight attitude error calculated by the UAV's inertial navigation function in real time and compare it with the confidence interval.

[0127] The attitude correction module is used to perform attitude correction operations if the real-time flight attitude error exceeds the boundary of the confidence interval.

[0128] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0129] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0130] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0131] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0132] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0133] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0134] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for autonomous flight attitude control of unmanned aerial vehicles based on inertial navigation, characterized in that, Includes the following steps: Acquire data on the current remaining battery power of the drone, the complexity of the external environment, and the associated deviation data extracted from the historical flight memory database; Based on the remaining energy data of the organism, the complexity data of the external environment, and the associated deviation data, the energy contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal are generated. The steps for generating the electrical shrinkage intensity signal, the environmental shrinkage intensity signal, and the memory shrinkage intensity signal include: The remaining battery power data of the body is converted into a battery contraction intensity signal using an S-shaped response function; The S-shaped response function is: ； In the formula, For electrical contraction intensity, This represents the current remaining battery power of the device. For the collective critical energy threshold, The parameter represents the position of the battery inflection point. This is the power sensitivity coefficient. It is a natural constant; The external environment complexity data is constructed into a multi-dimensional environment state vector, and the degree of deviation between the multi-dimensional environment state vector and the preset safe environment vector is obtained by the multi-dimensional distance calculation formula. The degree of deviation is used as the environmental contraction intensity signal. The steps for calculating the deviation between the multidimensional environment state vector and the preset safe environment vector using the multidimensional distance calculation formula include: Wind shear intensity, geomagnetic disturbance, and satellite loss-lock rate values ​​were extracted from external environmental complexity data. The wind shear intensity value, geomagnetic disturbance value, and satellite loss-lock rate value are combined to form a measured vector; A reference vector is formed by constructing a reference vector from the corresponding reference values ​​in the preset safe environment vector. The degree of deviation is obtained by the multidimensional distance calculation formula, and the degree of deviation is used as the environmental shrinkage intensity signal. The multidimensional distance calculation formula is as follows: ; In the formula, This is an environmental shrinkage intensity signal. The measured vector is the first Each component value The first reference vector Each component value For the first The normalized coefficients of each parameter, For the first Weighting factors for each parameter; The current remaining battery power data and external environment complexity data are input into the deviation prediction model based on the historical flight memory bank. The deviation prediction model outputs the predicted deviation value and uses the predicted deviation value as the memory contraction intensity signal. The electrical contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal are fused together to form the final interval contraction command; Based on the final interval contraction command, the reliable interval of dynamic flight attitude error is acquired and adjusted in real time. The real-time flight attitude error calculated by the UAV's inertial navigation function is acquired in real time and compared with the confidence interval. If the real-time flight attitude error exceeds the boundary of the confidence interval, an attitude correction operation is performed.

2. The method for autonomous flight attitude control of an unmanned aerial vehicle based on inertial navigation as described in claim 1, characterized in that, The steps of inputting the current remaining battery power data and external environmental complexity data into the deviation prediction model based on the historical flight memory bank, and outputting the predicted deviation value through the deviation prediction model, include: The remaining battery power, external environment complexity, and corresponding attitude error peak values ​​are extracted from the historical flight memory database. The training dataset is composed of the stored remaining battery power data, external environment complexity data, and corresponding attitude error peak values. A Gaussian process regression model is established for the training dataset, where the input is the remaining battery power of the machine and the complexity of the external environment, and the output is the peak value of the attitude error; The obtained data on the current remaining battery power of the organism and the complexity of the external environment are used as query points to input into the Gaussian process regression model, and the predicted deviation value is output as the memory contraction intensity signal.

3. The method for autonomous flight attitude control of an unmanned aerial vehicle based on inertial navigation as described in claim 1, characterized in that, The step of fusing the electrical contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal into the final interval contraction command includes: The values ​​of the electrical shrinkage intensity signal, the environmental shrinkage intensity signal, and the memory shrinkage intensity signal are sorted from largest to smallest; Based on the sorted values ​​of the electrical contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal, the final interval contraction command is obtained by performing calculations using a preset non-additive aggregation function. The non-additive aggregation function is: ； In the formula, This is the final interval contraction instruction. , , The ordered contraction intensity signal satisfies... , As the dominant factor, and ,in As the dominant sensitivity coefficient, This is the preset dominant threshold.

4. The method for autonomous flight attitude control of an unmanned aerial vehicle based on inertial navigation as described in claim 1, characterized in that, The steps for real-time acquisition and adjustment of the confidence interval of dynamic flight attitude error include: The width of the confidence interval of the dynamic flight attitude error is obtained by using a preset exponential decay function; The exponential decay function is: ； In the formula, The width of the confidence interval for dynamic flight attitude error. The minimum confidence interval width is preset. The preset initial confidence interval width, The attenuation intensity coefficient, This is the final interval contraction instruction. It is a natural constant.

5. The method for autonomous flight attitude control of an unmanned aerial vehicle based on inertial navigation as described in claim 1, characterized in that, The steps for obtaining the real-time flight attitude error calculated by the UAV's inertial navigation function include: Real-time acquisition of the drone's angular velocity and specific force measurements; The real-time angular velocity and specific force measurements are input into the strapdown inertial navigation algorithm to obtain the current flight attitude of the UAV in real time. The current flight attitude is compared with the initial flight attitude of the UAV, which is the flight attitude recorded when the UAV enters the inertial navigation mode; The difference between the current flight attitude and the initial flight attitude of the UAV is taken as the real-time flight attitude error.

6. The method for autonomous flight attitude control of an unmanned aerial vehicle based on inertial navigation as described in claim 5, characterized in that, If the real-time flight attitude error exceeds the boundary of the confidence interval, the steps for performing attitude correction operations include: Obtain the difference between the real-time flight attitude error and the boundary value of the confidence interval that it exceeds, and determine the difference as the over-limit error; The excess error is calculated with a preset correction gain coefficient to generate a correction control command; The correction control command is input to the attitude control actuator of the UAV to adjust the flight attitude of the UAV until the real-time flight attitude error returns to the confidence range of the dynamic flight attitude error.

7. An inertial navigation-based autonomous flight attitude control system for unmanned aerial vehicles (UAVs), used to implement the inertial navigation-based autonomous flight attitude control method for UAVs as described in any one of claims 1-6, characterized in that, include: Data acquisition module, signal generation module, instruction generation module, confidence interval acquisition module, attitude error comparison module, attitude correction module; The data acquisition module is used to acquire the current remaining battery power of the UAV, the complexity of the external environment, and the associated deviation data extracted from the historical flight memory database. The signal generation module is used to generate a power contraction intensity signal, an environmental contraction intensity signal, and a memory contraction intensity signal based on the remaining power data of the body, the complexity data of the external environment, and the associated deviation data. The instruction generation module is used to fuse the electrical contraction intensity signal, the environmental contraction intensity signal, and the memory contraction intensity signal into a final interval contraction instruction. The confidence interval acquisition module is used to acquire and adjust the confidence interval of dynamic flight attitude error in real time based on the final interval contraction command. The attitude error comparison module is used to acquire the real-time flight attitude error calculated by the UAV's inertial navigation function in real time and compare it with the confidence interval. The attitude correction module is used to perform attitude correction operations if the real-time flight attitude error exceeds the boundary of the confidence interval.

Images