Unmanned aerial vehicle attitude control method and system based on single neuron and related device

Abstract

The application is suitable for the technical field of unmanned aerial vehicles, and particularly relates to an unmanned aerial vehicle attitude control method and system based on a single neuron and related equipment. The application establishes an unmanned aerial vehicle attitude dynamics model, calculates a basic output control amount of the unmanned aerial vehicle according to feedback errors of expected attitudes and actual attitudes of the unmanned aerial vehicle attitude dynamics model through a PID controller, constructs a fractional order extended state observer model, uses the fractional order extended state observer model to estimate a total disturbance value of the unmanned aerial vehicle, compensates the total disturbance value to the PID controller, and calculates a control input amount of the unmanned aerial vehicle, and establishes a mapping relationship between a single neuron adaptive model and preset parameters in the fractional order extended state observer model. Compared with the prior art, the application improves system state estimation and disturbance observation accuracy of the unmanned aerial vehicle, and significantly enhances robustness, anti-interference capability and noise insensitivity of the unmanned aerial vehicle control system.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) technology, specifically to a single-neuron-based UAV attitude control method, system, and related equipment. Background Technology

[0002] Extended State Observer (ESO) is the core technology of Active Disturbance Rejection Control (ADRC) framework. Its core function is to treat the dynamics of the UAV system and external disturbances as a unified "total disturbance" and estimate it in real time, so as to achieve disturbance suppression without the need for a precise mathematical model. It is widely used in state observation and anti-interference control in fields such as industrial control, aerospace, and robotics.

[0003] Although existing integer-order extended state observers have achieved good control results in many engineering applications, their performance still has significant limitations under complex conditions such as nonlinearity, strong coupling, and fast time-varying disturbances. On the one hand, extended state observers are built based on integer-order operators, which limits their dynamic response flexibility and makes them insufficiently adaptable to nonlinear characteristics and distributed memory effects, making it difficult to accurately characterize the disturbance behavior of complex systems. On the other hand, increasing the observation gain can improve the transient response, but it will significantly amplify high-frequency noise, causing output oscillations and observation errors, resulting in a contradiction between response speed and noise immunity. Furthermore, integer-order extended state observers have strong parameter coupling, and the tuning process is significantly dependent on engineering experience, further increasing the design difficulty in multivariable or highly dynamic systems.

[0004] To overcome the above contradictions, existing research has proposed a fractional extended state observer method, which utilizes the memory and nonlocality of fractional operators to enhance the modeling ability of systems and disturbances, and can achieve better observation accuracy compared with integer extended state observers.

[0005] However, existing fractional-order extended state observers still have the following shortcomings:

[0006] (1) The parameters are highly coupled and their tuning depends on experience; (2) While fractional-order structures improve observation accuracy, they remain sensitive to noise. (3) Lack of online learning and dynamic compensation mechanisms for disturbance characteristics. These limitations restrict the application and promotion of fractional extended state observers in a wider range of engineering scenarios.

[0007] Therefore, there is an urgent need for a new single-neuron-based method, system, and related equipment for UAV attitude control to solve the above-mentioned technical problems. Summary of the Invention

[0008] This invention provides a method, system, and related equipment for UAV attitude control based on a single neuron, aiming to solve the core problems of existing UAV control systems, such as low tracking accuracy, weak anti-disturbance capability, and complex parameter tuning in scenarios with strong disturbances and time-varying parameters.

[0009] In a first aspect, the present invention provides a method for attitude control of a UAV based on a single neuron, comprising the following steps: S1. Establish a UAV attitude dynamics model. Based on the feedback error between the desired attitude and the actual attitude of the UAV attitude dynamics model, calculate the basic output control quantity of the UAV through a PID controller. S2. Construct a fractional-order extended state observer model, which is used to estimate the total disturbance value of the UAV and compensate the total disturbance value to the PID controller to calculate the control input of the UAV. S3. Establish a single neuron adaptive model, establish a mapping relationship between the single neuron adaptive model and the preset parameters in the fractional-order extended state observer model, and adaptively update the preset parameters in the fractional-order extended state observer model based on preset rules to obtain an optimized fractional-order extended state observer model. S4. The disturbance value of the UAV is estimated by the optimized fractional extended state observer model to obtain the disturbance estimate. S5. Feedforward the disturbance estimate to the basic output control quantity of the PID controller to obtain the final control input quantity, thereby realizing the attitude control of the UAV.

[0010] Preferably, the attitude dynamics equations of the UAV attitude dynamics model satisfy the following conditions: ; in, as well as These represent the second derivatives of the roll angle, pitch angle, and yaw angle of the UAV, respectively. as well as These respectively indicate that the drone is in x axis, y shaft and z Moment of inertia under the axis. , as well as Both represent external unknown disturbance terms.

[0011] Preferably, the control torque of the UAV attitude dynamics model satisfies the following condition: ; in, These respectively indicate that the drone is in xaxis, y shaft and z Control torque under the shaft; as well as These represent the different motor speeds of the drone; This indicates the length of the drone's arm. Indicates the thrust coefficient. This represents the torque coefficient.

[0012] Preferably, the basic output control quantity of the PID controller satisfies the following conditions: ; in, This represents the basic output control quantity. This indicates the feedback error. This represents the first derivative of the feedback error. , as well as These represent the proportional parameter, integral parameter, and derivative coefficient parameter of the PID controller, respectively.

[0013] Preferably, the fractional-order extended state observer model estimates the perturbation estimate based on the following conditions: ; in, , , , as well as These represent the observer feedback gain coefficients, and Let represent the orders of the first and second fractional derivatives, respectively. This represents a first-order fractional differential operator; This represents a second-order fractional differential operator; express The estimated value, express The estimated value, This represents the estimated disturbance value. , as well as They represent , as well as The first derivative, This represents the control gain parameter. This represents the output quantity in the first state. This represents the output quantity of the second state.

[0014] Preferably, step S3 includes the following sub-steps: S31. Discretize the first-order fractional differential operator and the second-order fractional differential operator; S32. Map the preset parameters to the output of the single-neuron adaptive model; wherein the preset parameters include the first-order fractional derivative, the second-order fractional derivative, and the observer feedback gain coefficient.

[0015] Preferably, the preset rule is: to ensure that the preset parameters in the fractional-order extended state observer model satisfy the following conditions: ; in, and These represent the time intervals at which a single neuron is acquired. and Input weights, This represents the input weights after normalization. This represents the i-th input terminal of the single neuron; and These represent the sampling times. The preset parameters, Indicates the learning rate of the weights. This indicates the proportion of neuron update rates.

[0016] Secondly, the present invention also provides a single-neuron-based unmanned aerial vehicle (UAV) attitude control system, the UAV attitude control system comprising: The attitude establishment module is used to establish the attitude dynamics model of the UAV. Based on the feedback error between the expected attitude and the actual attitude of the UAV attitude dynamics model, the basic output control quantity of the UAV is calculated by the PID controller. The model building module is used to construct a fractional-order extended state observer model, which is used to estimate the total disturbance value experienced by the UAV and compensate the total disturbance value to the PID controller to calculate the control input and basic output control quantity of the UAV. The model optimization module is used to establish a single neuron adaptive model, establish a mapping relationship between the single neuron adaptive model and the preset parameters in the fractional-order extended state observer model, and update the preset parameters in the fractional-order extended state observer model based on preset rules to obtain an optimized fractional-order extended state observer model. The disturbance calculation module is used to estimate the disturbance value of the UAV through the optimized fractional extended state observer model to obtain the disturbance estimate value; The attitude control module is used to feed forward the disturbance estimate to the basic output control quantity of the PID controller to obtain the final control input quantity, thereby realizing the attitude control of the UAV.

[0017] Thirdly, the present invention also provides a computer device, including: a memory, a processor, and a single-neuron-based UAV attitude control program stored in the memory and executable on the processor, wherein when the processor executes the single-neuron-based UAV attitude control program, it implements the steps of the single-neuron-based UAV attitude control method as described in any of the above embodiments.

[0018] Fourthly, the present invention also provides a computer-readable storage medium storing a single-neuron-based UAV attitude control program, wherein the single-neuron-based UAV attitude control program, when executed by a processor, implements the steps of the single-neuron-based UAV attitude control method as described in any of the above embodiments.

[0019] Compared with existing technologies, this invention establishes a UAV attitude dynamics model. Based on the feedback error between the desired attitude and the actual attitude of the UAV, a PID controller calculates the basic output control quantity of the UAV. A fractional-order extended state observer model is constructed to estimate the total disturbance value experienced by the UAV and compensates this total disturbance value to the PID controller to calculate the control input quantity of the UAV. A single-neuron adaptive model is established, mapping the preset parameters in the single-neuron adaptive model to the preset parameters in the fractional-order extended state observer model. Based on preset rules, the preset parameters in the fractional-order extended state observer model are adaptively updated to obtain an optimized fractional-order extended state observer model. The disturbance value of the UAV is estimated using the optimized fractional-order extended state observer model to obtain the disturbance estimate. The disturbance estimate is then fed forward to compensate the basic output control quantity of the PID controller to obtain the final control input quantity, thus achieving attitude control of the UAV. This invention utilizes an error-driven single-neuron adaptive model learning mechanism to adjust fractional-order weights in real time. This enhances the observation and control capabilities of the fractional-order extended state observer model during periods of severe disturbance and suppresses them during periods of dominant noise, thus achieving a dynamic balance between observation accuracy and noise robustness. This improves the accuracy of UAV system state estimation and disturbance observation, enhances the anti-interference capability and noise insensitivity of UAV control, and ultimately improves the control performance of the UAV. Attached Figure Description

[0020] The present invention will now be described in detail with reference to the accompanying drawings. The above and other aspects of the present invention will become clearer and more readily understood through the detailed description following the accompanying drawings. In the drawings: Figure 1 This is a flowchart of the UAV attitude control method based on a single neuron provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a UAV attitude control system based on a single neuron provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0022] Example 1 Please refer to Figure 1 This invention provides a single-neuron-based UAV attitude control method, which includes the following steps: S1. Establish a UAV attitude dynamics model. Based on the feedback error between the desired attitude and the actual attitude of the UAV attitude dynamics model, calculate the basic output control quantity of the UAV through a PID controller.

[0023] In this embodiment of the invention, a UAV attitude dynamics model is constructed to clarify the physical relationship between the UAV's attitude angles, control torque, and motor speed. The attitude dynamics of the UAV attitude dynamics model satisfy the following conditions: ; in, as well as These represent the second derivatives of the roll angle, pitch angle, and yaw angle of the UAV, respectively. as well as These respectively indicate that the drone is in x axis, y shaft and z Moment of inertia under the axis. , as well as Both represent external unknown disturbance terms acting on the corresponding channel.

[0024] In this embodiment of the invention, the control torque of the UAV attitude dynamics model satisfies the following condition: ; in, These respectively indicate that the drone is in x axis, y shaft and z Control torque under the shaft; as well as These represent the different motor speeds of the drone; This indicates the length of the drone's arm. Indicates the thrust coefficient. This represents the torque coefficient.

[0025] In this embodiment of the invention, the basic output control quantity satisfies the following condition: ; in, This represents the basic output control quantity. This indicates the feedback error. This represents the first derivative of the feedback error. , as well as These represent the proportional parameter, integral parameter, and derivative coefficient parameter of the PID controller, respectively.

[0026] S2. Construct a fractional-order extended state observer model. The fractional-order extended state observer model is used to estimate the total disturbance value of the UAV and compensate the total disturbance value to the PID controller to calculate the control input of the UAV.

[0027] In this embodiment of the invention, the integer-order differential operators in the existing extended state observer are replaced with fractional-order differential operators to form a fractional-order extended state observer model. Utilizing the memory and nonlocality of fractional-order operators, the observer's ability to characterize complex disturbances is improved, enabling rapid and accurate estimation of the UAV's system state and external disturbances. The fractional-order extended state observer model estimates the total disturbance experienced during multi-rotor attitude control in real time. d ( t ), This leads to an attitude control framework based on a fractional-order extended state observer model, where the observer's inputs are the UAV's system measurable outputs (such as attitude angles) and PID control variables. u The output is a disturbance estimate, and the output is directly connected to the PID control loop to achieve disturbance compensation.

[0028] In this embodiment of the invention, the fractional-order extended state observer model estimates the perturbation estimate based on the following conditions: ; in, , , , as well as These represent the observer feedback gain coefficients, and Let represent the orders of the first and second fractional derivatives, respectively. , ; This represents a first-order fractional differential operator; This represents a second-order fractional differential operator; express The estimated value, express The estimated value, This represents the estimated disturbance value. , as well as They represent , as well as The first derivative, This represents the control gain parameter. This represents the output quantity in the first state. This represents the output quantity of the second state.

[0029] S3. Establish a single-neuron adaptive model, establish a mapping relationship between the single-neuron adaptive model and the preset parameters in the fractional-order extended state observer model, and adaptively update the preset parameters in the fractional-order extended state observer model based on preset rules to obtain an optimized fractional-order extended state observer model.

[0030] In this embodiment of the invention, step S3 includes the following sub-steps: S31. Discretize the first-order fractional differential operator and the second-order fractional differential operator to obtain the disturbance estimate in discrete time. The update equation is ; Right now: ; in, For a moment The estimated total system disturbance is essentially a weighted fusion of observation errors at different historical moments, determined by the fractional operator coefficients and the GL recursive weights.

[0031] S32. Map the preset parameters to the output of the single-neuron adaptive model; wherein the preset parameters include the first-order fractional derivative, the second-order fractional derivative, and the observer feedback gain coefficient.

[0032] When establishing a mapping relationship between the single-neuron adaptive model and the preset parameters in the fractional-order extended state observer model, and updating the preset parameters in the fractional-order extended state observer model based on preset rules, the following principles must be satisfied: By establishing a mapping relationship between the single-neuron adaptive model and the preset parameters in the fractional-order extended state observer model (i.e. , , , ), with respect to fractional derivative terms and Grünwald-Letnikov discretization is performed, and historical error signals are preserved through time-series input nodes, which fits the nonlocality of fractional calculus.

[0033] Each key parameter is mapped to a single neuron for independent output, avoiding interference between parameters with different physical meanings during the learning process and solving the parameter coupling problem; In this embodiment of the invention, the preset rule for the single-neuron adaptive model is: the constructed single-neuron structure uses fractional-order operator core parameters. To achieve the desired output adjustment, the fractional-order parameters are designed to dynamically adapt to changes in system disturbances. The update rules for the parameters and weights in this embodiment are as follows: ; in, and These represent the time intervals at which a single neuron is acquired. and Input weights, This represents the input weights after normalization. This represents the i-th input terminal of the single neuron; and These represent the sampling times. The preset parameters, The learning rate for the weights controls the step size for updating the weights, and its value range is [range missing]. ; This is a percentage of the neuron update rate, used to control the update speed of parameters; its value range is... .

[0034] Through the adaptive adjustment mechanism based on observation error information as described above, the parameters are automatically tuned online by updating the rules, without the need for manual intervention.

[0035] S4. The disturbance value of the UAV is estimated by the optimized fractional extended state observer model to obtain the disturbance estimate.

[0036] In this embodiment of the invention, the optimized fractional-order extended state observer model, improved by a fractional-order and single-neuron adaptive model, estimates the perturbation value of the UAV, which can effectively improve observation accuracy, enhance anti-interference capability, and suppress high-frequency noise. The perturbation estimate is the total perturbation value received by the optimized UAV.

[0037] S5. Feedforward the disturbance estimate to the basic output control quantity of the PID controller to obtain the final control input quantity, thereby realizing the attitude control of the UAV.

[0038] In this embodiment of the invention, a PID controller generates a basic control quantity based on the feedback error input value between the desired attitude and the actual attitude of the UAV attitude dynamics model. Based on the disturbance value of the UAV estimated by the optimized fractional extended state observer model, a disturbance estimate is obtained. The disturbance estimate is multiplied by the gain system control gain and subtracted from the basic control quantity to obtain the final control input quantity, which is input to the UAV to drive the motor adjustment, thereby realizing attitude control of the UAV.

[0039] Compared with existing technologies, this invention establishes a UAV attitude dynamics model. Based on the feedback error between the desired attitude and the actual attitude of the UAV, a PID controller calculates the basic output control quantity of the UAV. A fractional-order extended state observer model is constructed to estimate the total disturbance value experienced by the UAV and compensates this total disturbance value to the PID controller to calculate the control input quantity of the UAV. A single-neuron adaptive model is established, mapping the preset parameters in the single-neuron adaptive model to the preset parameters in the fractional-order extended state observer model. Based on preset rules, the preset parameters in the fractional-order extended state observer model are adaptively updated to obtain an optimized fractional-order extended state observer model. The disturbance value of the UAV is estimated using the optimized fractional-order extended state observer model to obtain the disturbance estimate. The disturbance estimate is then fed forward to compensate the basic output control quantity of the PID controller to obtain the final control input quantity, thus achieving attitude control of the UAV. This invention utilizes an error-driven single-neuron adaptive model learning mechanism to adjust fractional-order weights in real time. This enhances the observation and control capabilities of the fractional-order extended state observer model during periods of severe disturbance and suppresses them during periods of dominant noise, thus achieving a dynamic balance between observation accuracy and noise robustness. This improves the accuracy of UAV system state estimation and disturbance observation, enhances the anti-interference capability and noise insensitivity of UAV control, and ultimately improves the control performance of the UAV.

[0040] Example 2 This invention also provides a single-neuron-based UAV attitude control system, please refer to... Figure 2 , Figure 2 This is a schematic diagram of the structure of a single-neuron-based UAV attitude control system 200 provided in an embodiment of the present invention, which includes: 201. Attitude Establishment Module: This module is used to establish the attitude dynamics model of the UAV. Based on the feedback error between the desired attitude and the actual attitude of the UAV attitude dynamics model, the basic output control quantity of the UAV is calculated by a PID controller.

[0041] In this embodiment of the invention, a UAV attitude dynamics model is constructed to clarify the physical relationship between the UAV's attitude angles, control torque, and motor speed. The attitude dynamics of the UAV attitude dynamics model satisfy the following conditions: ; in, as well as These represent the second derivatives of the roll angle, pitch angle, and yaw angle of the UAV, respectively. as well as These respectively indicate that the drone is in x axis, y shaft and z Moment of inertia under the axis. , as well as Both represent external unknown disturbance terms acting on the corresponding channel.

[0042] In this embodiment of the invention, the control torque of the UAV attitude dynamics model satisfies the following condition: ; in, These respectively indicate that the drone is in x axis, y shaft and z Control torque under the shaft; as well as Both represent the motor speed of the drone; This indicates the length of the drone's arm. Indicates the thrust coefficient. This represents the torque coefficient.

[0043] In this embodiment of the invention, the basic output control quantity satisfies the following condition: ; in, This represents the basic output control quantity. This indicates the feedback error. This represents the first derivative of the feedback error. , as well as These represent the proportional parameter, integral parameter, and derivative coefficient parameter of the PID controller, respectively.

[0044] 202. Model building module, used to construct a fractional-order extended state observer model, which is used to estimate the total disturbance value of the UAV and compensate the total disturbance value to the PID controller to calculate the control input of the UAV.

[0045] In this embodiment of the invention, the integer-order differential operators in the existing extended state observer are replaced with fractional-order differential operators to form a fractional-order extended state observer model. Utilizing the memory and nonlocality of fractional-order operators, the observer's ability to characterize complex disturbances is improved, enabling rapid and accurate estimation of the UAV's system state and external disturbances. The fractional-order extended state observer model estimates the total disturbance experienced during multi-rotor attitude control in real time. d ( t ), This leads to an attitude control framework based on a fractional-order extended state observer model, where the observer's inputs are the UAV's system measurable outputs (such as attitude angles) and PID control variables. u The output is a disturbance estimate, and the output is directly connected to the PID control loop to achieve disturbance compensation.

[0046] In this embodiment of the invention, the fractional-order extended state observer model estimates the perturbation estimate based on the following conditions: ; in, , , , as well as These represent the observer feedback gain coefficients with different values. and Let represent the orders of the first and second fractional derivatives, respectively. , ; This represents a first-order fractional differential operator; This represents a second-order fractional differential operator; express The estimated value, express The estimated value, This represents the estimated disturbance value. , as well as They represent , as well as The first derivative, This represents the control gain parameter. This represents the output quantity in the first state. This represents the output quantity of the second state.

[0047] 203. Model optimization module, used to establish a single neuron adaptive model, establish a mapping relationship between the single neuron adaptive model and the preset parameters in the fractional-order extended state observer model, and adaptively update the preset parameters in the fractional-order extended state observer model based on preset rules to obtain an optimized fractional-order extended state observer model.

[0048] In this embodiment of the invention, the first-order fractional differential operator and the second-order fractional differential operator are discretized. The disturbance estimate in discrete time can be obtained. The update equation is ; Right now: ; in, For a moment The estimated total system disturbance is essentially a weighted fusion of observation errors at different historical moments, determined by the fractional operator coefficients and the GL recursive weights.

[0049] The preset parameters are mapped to the weights or nodes of the single-neuron adaptive model; wherein the preset parameters include the first-order fractional derivative, the second-order fractional derivative, and the observer feedback gain coefficient.

[0050] When establishing a mapping relationship between the single-neuron adaptive model and the preset parameters in the fractional-order extended state observer model, and updating the preset parameters in the fractional-order extended state observer model based on preset rules, the following principles must be satisfied: By establishing a mapping relationship between the single-neuron adaptive model and the preset parameters in the fractional-order extended state observer model (i.e. , , , ), with respect to fractional derivative terms and Grünwald-Letnikov discretization is performed, and historical error signals are preserved through time-series input nodes, which fits the nonlocality of fractional calculus.

[0051] Each key parameter is mapped to a single neuron for independent output, avoiding interference between parameters with different physical meanings during the learning process and solving the parameter coupling problem; In this embodiment of the invention, the preset rule is: the constructed single-neuron structure uses fractional-order operator core parameters. To achieve the desired output adjustment, the fractional-order parameters are designed to dynamically adapt to changes in system disturbances. The parameter and weight update rules in this embodiment are as follows: The preset rule is: the constructed single-neuron structure uses the core parameters of the fractional-order operator... To achieve the desired output adjustment, the fractional-order parameters are designed to dynamically adapt to changes in system disturbances. The update rules for the parameters and weights in this embodiment are as follows: ; in, and These represent the time intervals at which a single neuron is acquired. and Input weights, This represents the input weights after normalization. This represents the i-th input terminal of the single neuron; and These represent the sampling times. The preset parameters, The learning rate for the weights controls the step size for updating the weights, and its value range is [range missing]. ; This is a percentage of the neuron update rate, used to control the update speed of parameters; its value range is... .

[0052] Through the adaptive adjustment mechanism based on observation error information as described above, the parameters are automatically tuned online by updating the rules, without the need for manual intervention.

[0053] 204. Disturbance calculation module, used to estimate the disturbance value of the UAV through the optimized fractional extended state observer model, and obtain the disturbance estimate value.

[0054] In this embodiment of the invention, the optimized fractional-order extended state observer model, improved by fractional-order and single-neuron adaptive models, estimates the perturbation value of the UAV, which can effectively improve observation accuracy, enhance anti-interference capability, and suppress high-frequency noise.

[0055] 205. Attitude control module, used to feed forward the disturbance estimate to the basic output control quantity of the PID controller to obtain the final control input quantity, thereby realizing attitude control of the UAV.

[0056] The single-neuron-based UAV attitude control system 200 can implement the steps in the single-neuron-based UAV attitude control method in the above embodiments and achieve the same technical effect. Refer to the description in the above embodiments, which will not be repeated here.

[0057] Example 3 This invention also provides a computer device, please refer to... Figure 3 , Figure 3 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. The computer device 300 includes: a memory 302, a processor 301, and a single-neuron-based UAV attitude control program stored in the memory 302 and capable of running on the processor 301.

[0058] The processor 301 calls the single-neuron-based UAV attitude control program stored in the memory 302 and executes the steps in the single-neuron-based UAV attitude control method provided in this embodiment of the invention. Please refer to... Figure 1 Specifically, it includes the following steps: S1. Establish a UAV attitude dynamics model. Based on the feedback error between the desired attitude and the actual attitude of the UAV attitude dynamics model, calculate the basic output control quantity of the UAV through a PID controller. S2. Construct a fractional-order extended state observer model, which is used to estimate the total disturbance value of the UAV and compensate the total disturbance value to the PID controller to calculate the control input of the UAV. S3. Establish a single neuron adaptive model, establish a mapping relationship between the single neuron adaptive model and the preset parameters in the fractional-order extended state observer model, and adaptively update the preset parameters in the fractional-order extended state observer model based on preset rules to obtain an optimized fractional-order extended state observer model. S4. The disturbance value of the UAV is estimated by the optimized fractional extended state observer model to obtain the disturbance estimate. S5. Feedforward the disturbance estimate to the basic output control quantity of the PID controller to obtain the final control input quantity, thereby realizing the attitude control of the UAV.

[0059] The computer device 300 provided in this embodiment of the invention can implement the steps in the single-neuron-based UAV attitude control method as described in the above embodiments, and can achieve the same technical effect. Refer to the description in the above embodiments, which will not be repeated here.

[0060] Example 4 This invention also provides a computer-readable storage medium storing a single-neuron-based UAV attitude control program. When executed by a processor, the single-neuron-based UAV attitude control program implements the various processes and steps of the single-neuron-based UAV attitude control method provided in this invention and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0061] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by hardware related to computer programs or instructions. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0062] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0063] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0064] The embodiments of the present invention have been described above with reference to the accompanying drawings. The disclosed embodiments are merely preferred embodiments of the present invention. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many equivalent changes in form under the guidance of the present invention without departing from the spirit and scope of the claims. All such changes are within the protection scope of the present invention.

Claims

1. A method for attitude control of a UAV based on a single neuron, characterized in that, Includes the following steps: S1. Establish a UAV attitude dynamics model. Based on the feedback error between the desired attitude and the actual attitude of the UAV attitude dynamics model, calculate the basic output control quantity of the UAV through a PID controller. S2. Construct a fractional-order extended state observer model, which is used to estimate the total disturbance value of the UAV and compensate the total disturbance value to the PID controller to calculate the control input of the UAV. S3. Establish a single neuron adaptive model, establish a mapping relationship between the single neuron adaptive model and the preset parameters in the fractional-order extended state observer model, and adaptively update the preset parameters in the fractional-order extended state observer model based on preset rules to obtain an optimized fractional-order extended state observer model. S4. The disturbance value of the UAV is estimated by the optimized fractional extended state observer model to obtain the disturbance estimate. S5. Feedforward the disturbance estimate to the basic output control quantity of the PID controller to obtain the final control input quantity, thereby realizing the attitude control of the UAV.

2. The UAV attitude control method based on a single neuron as described in claim 1, characterized in that, The attitude dynamics equations of the UAV attitude dynamics model satisfy the following conditions: ； in, as well as These represent the second derivatives of the roll angle, pitch angle, and yaw angle of the UAV, respectively. as well as These respectively indicate that the drone is in x axis, y shaft and z Moment of inertia under the axis. , as well as Both represent external unknown disturbance terms.

3. The UAV attitude control method based on a single neuron as described in claim 2, characterized in that, The control torque of the UAV attitude dynamics model satisfies the following condition: ； in, These respectively indicate that the drone is in x axis, y shaft and z Control torque under the shaft; as well as These represent the different motor speeds of the drone; This indicates the length of the drone's arm. Indicates the thrust coefficient. This represents the torque coefficient.

4. The UAV attitude control method based on a single neuron as described in claim 3, characterized in that, The basic output control quantity of the PID controller satisfies the following condition: ； in, This represents the basic output control quantity. This indicates the feedback error. This represents the first derivative of the feedback error. , as well as These represent the proportional parameter, integral parameter, and derivative coefficient parameter of the PID controller, respectively.

5. The UAV attitude control method based on a single neuron as described in claim 4, characterized in that, The fractional-order extended state observer model estimates the perturbation estimate based on the following conditions: ； in, , , , as well as These represent the observer feedback gain coefficients, and Let represent the orders of the first and second fractional derivatives, respectively. This represents a first-order fractional differential operator; This represents a second-order fractional differential operator; express The estimated value, express The estimated value, This represents the estimated disturbance value. , as well as They represent , as well as The first derivative, This represents the control gain parameter. This represents the output quantity in the first state. This represents the output quantity of the second state.

6. The UAV attitude control method based on a single neuron as described in claim 5, characterized in that, Step S3 includes the following sub-steps: S31. Discretize the first-order fractional differential operator and the second-order fractional differential operator; S32. Map the preset parameters to the output of the single-neuron adaptive model; wherein the preset parameters include the first-order fractional derivative, the second-order fractional derivative, and the observer feedback gain coefficient.

7. The UAV attitude control method based on a single neuron as described in claim 6, characterized in that, The preset rule is to ensure that the preset parameters in the fractional-order extended state observer model satisfy the following conditions: ； in, and These represent the time intervals at which a single neuron is acquired. and Input weights, This represents the input weights after normalization. This represents the i-th input terminal of the single neuron; and These represent the sampling times. The preset parameters, Indicates the learning rate of the weights. This indicates the proportion of neuron update rates.

8. A single-neuron-based unmanned aerial vehicle (UAV) attitude control system, characterized in that, The UAV attitude control system includes: The attitude establishment module is used to establish the attitude dynamics model of the UAV. Based on the feedback error between the expected attitude and the actual attitude of the UAV attitude dynamics model, the basic output control quantity of the UAV is calculated by the PID controller. The model building module is used to construct a fractional-order extended state observer model, which is used to estimate the total disturbance value of the UAV and compensate the total disturbance value to the PID controller to calculate the control input of the UAV. The model optimization module is used to establish a single neuron adaptive model, establish a mapping relationship between the single neuron adaptive model and the preset parameters in the fractional-order extended state observer model, and update the preset parameters in the fractional-order extended state observer model based on preset rules to obtain an optimized fractional-order extended state observer model. The disturbance calculation module is used to estimate the disturbance value of the UAV through the optimized fractional extended state observer model to obtain the disturbance estimate value; The attitude control module is used to feed forward the disturbance estimate to the basic output control quantity of the PID controller to obtain the final control input quantity, thereby realizing the attitude control of the UAV.

9. A computer device, characterized in that, include: The system includes a memory, a processor, and a single-neuron-based UAV attitude control program stored in the memory and executable on the processor. When the processor executes the single-neuron-based UAV attitude control program, it implements the steps of the single-neuron-based UAV attitude control method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a single-neuron-based UAV attitude control program, which, when executed by a processor, implements the steps of the single-neuron-based UAV attitude control method as described in any one of claims 1-7.

Images