A robust control method for unmanned aerial vehicles based on extended state observer

Abstract

The invention discloses a method for robust control of an unmanned aerial vehicle based on an extended state observer. A tracking differentiator is used to track a trajectory signal of the unmanned aerial vehicle, a lumped disturbance is introduced to reconstruct a state error system, an extended state observer is designed in combination with adaptive theory and a galn function, a state error change rate estimate and a lumped disturbance estimate of each channel are estimated, a non-singular terminal sliding mode self-disturbance rejection control strategy is determined, and a control quantity of each channel of the unmanned aerial vehicle is obtained; the state error change rate estimate is used as the state error change rate of the current unmanned aerial vehicle, and the lumped disturbance estimate is used as the lumped disturbance of the current unmanned aerial vehicle to realize speed sensorless control, and the control quantity output by the speed sensorless controller is fed back to the extended state observer; according to the control quantity allocation strategy, the control quantity is converted into a thrust instruction of a motor of a quad-rotor unmanned aerial vehicle, and while controlling the movement of the unmanned aerial vehicle, the real-time position and attitude information of the unmanned aerial vehicle is fed back to complete a control closed loop, thereby realizing robust control of the unmanned aerial vehicle.

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicle control, and in particular to a method for robust control of unmanned aerial vehicles based on an extended state observer. Background Art

[0002] Quadrotors play a key role in agricultural monitoring, disaster assessment, traffic monitoring and other scenarios due to their compact size, low cost and excellent maneuverability, and are widely used in data collection, image analysis and other monitoring tasks. In order to ensure that quadrotors perform flight missions according to the predetermined trajectory, high-precision trajectory tracking control is critical. However, quadrotors are nonlinear, strongly coupled and under-actuated complex systems, which are extremely susceptible to multi-source unknown disturbances such as internal parameter perturbations, external airflow changes and inaccurate modeling, which increases the complexity of designing efficient and reliable drone trajectory tracking control strategies.

[0003] The current mainstream solution is to accurately estimate multi-source unknown disturbances by designing an extended state observer, and use these estimated disturbance information to drive the controller to attenuate or offset the impact of the disturbance. Although the existing solution can alleviate the impact of internal uncertainty and external disturbances from the UAV to a certain extent, it is difficult to balance the contradiction between rapidity and overshoot, has high requirements for sensors, and is difficult to process expected signals that do not have closed-form solutions, resulting in complex UAV control systems, high costs, and poor control accuracy when encountering unknown time-varying disturbances. Summary of the invention

[0004] Purpose of the invention: The present invention aims to provide a robust control method for an unmanned aerial vehicle (UAV) using a non-singular terminal sliding mode auto-disturbance rejection control strategy based on an extended state observer optimized by a tracking differentiator, adaptive theory and galn function.

[0005] Technical solution: The UAV robust control method based on extended state observer of the present invention comprises the following steps:

[0006] (1) Establish a dynamic model of a quadrotor drone based on multi-source disturbances;

[0007] (2) Introducing virtual control variables into the quadrotor UAV dynamics model and determining the control variable allocation strategy;

[0008] (3) Design a tracking differentiator using the fastest control synthesis function to track the trajectory signal of the UAV;

[0009] (4) Based on the UAV position and attitude information and the tracking signal obtained by the tracking differentiator, the state error information is obtained, the state error system is determined, and the lumped disturbance is introduced to reconstruct the state error system;

[0010] (5) Based on the reconstructed state error system, combined with adaptive theory and galn function, an extended state observer is designed. Combined with the current control quantity of each channel, the state error change rate estimation and lumped disturbance estimation of each channel are estimated;

[0011] (6) Construct a non-singular terminal integral sliding mode controller. According to the reconstructed state error system, the state error change rate and the lumped disturbance, determine the non-singular terminal sliding mode active disturbance rejection control strategy and obtain the control quantity of each channel of the UAV.

[0012] (7) The estimated state error change rate obtained in step (5) is used as the state error change rate of the current UAV.

[0013] The lumped disturbance estimation is used as the lumped disturbance of the current UAV to realize speed sensorless control, and the control quantity output by the speed sensorless controller is fed back to the extended state observer in step (5);

[0014] (8) According to the control quantity allocation strategy of step (2), the control quantity obtained in step (7) is converted into the thrust command of the motor of the quadrotor UAV, and the real-time position and attitude information of the UAV is fed back to step (4) while controlling the movement of the UAV, thus completing the control closed loop and realizing robust control of the UAV.

[0015] Furthermore, the control quantity allocation strategy in step (2) is

[0016] ;

[0017] .

[0018] Furthermore, in step (3), the tracking differentiator is designed using the fastest control synthesis function as

[0019] .

[0020] Furthermore, in step (4), the lumped perturbation is

[0021] ;

[0022] .

[0023] Furthermore, in step (4), after introducing the lumped disturbance, the reconstructed state error system is

[0024] .

[0025] Furthermore, in step (5), the design of the extended state observer includes an extended state observer for the position loop state error system and an extended state observer for the attitude loop state error system; the extended state observer for the position loop state error system is

[0026] ;

[0027] The extended state observer of the attitude loop state error system is:

[0028] .

[0029] Furthermore, step (6) is specifically as follows: according to the reconstructed state error system, the state error change rate and the lumped disturbance, the non-singular terminal integral sliding surface function is determined; based on the adaptive theory, the sliding mode convergence law adopts the hyperbolic tangent function to obtain the adaptive sliding mode convergence law; according to the first-order derivative of the non-singular terminal integral sliding surface function and the adaptive sliding mode convergence law, the control amount of each channel of the UAV is obtained; the non-singular terminal integral sliding surface function is

[0030] ;

[0031] .

[0032] Furthermore, in step (6), the adaptive sliding mode reaching law is

[0033] ;

[0034] ;

[0035] ;

[0036] .

[0037] Furthermore, in step (6), the control quantity of each channel of the drone is

[0038] .

[0039] Furthermore, in step (7), the speed sensorless controller is specifically as follows:

[0040] ;

[0041] ;

[0042] .

[0043] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: 1. The present invention adopts the fastest control comprehensive function to design the tracking differentiator to track the desired trajectory signal of the UAV, which significantly improves the tracking accuracy of the UAV for complex trajectories, and while ensuring rapid response, effectively reduces the overshoot phenomenon of the system, optimizes the dynamic performance of the system, improves the control efficiency, reduces energy consumption, and provides strong support for the stable flight of the UAV under multi-source interference conditions; 2. The present invention combines adaptive theory and galn function to optimize the traditional extended state observer, improves the observation accuracy of the observer, enhances its adaptability to dynamic changes of the system, simplifies the parameter adjustment process, improves the performance in uncertain environments, and ensures the stability and reliability of the observer in the face of model errors and external disturbances; 3. The present invention reconstructs the state error system, incorporates the dynamic characteristics of the desired signal and its second-order derivative information, which are usually difficult to express with a closed-form solution, into the category of lumped disturbances, and uses the optimized extended state observer to accurately observe this lumped disturbance, effectively bypassing the closed-form solution of the desired signal. The problem of loss not only ensures the rigor and practicality of the control strategy, but also improves the control performance of the system when processing complex desired trajectories; 4. The present invention adopts an optimized extended state observer to accurately observe the rate of change of the state error, and uses this observation value to replace the traditional real feedback signal, and then designs a non-singular terminal sliding mode controller, which reduces the dependence of the drone on the speed sensor, thereby reducing the complexity and cost of the entire drone system, and realizing the speed sensorless control of the drone, which not only improves the system's integration and economy, but also opens up a new path for drone control technology, and realizes a more efficient and economical flight control solution; 5. The present invention determines the non-singular terminal integral sliding mode surface function based on the reconstructed state error system, the state error change rate and the lumped disturbance, which can effectively avoid the singular value problem, thereby improving the convergence speed of the system; 6. The adaptive sliding mode convergence law adopted by the present invention increases the convergence speed, reduces the system chattering, further improves the overall performance of the system, and the control quantity output becomes smoother, thereby improving the control accuracy and robustness of the drone in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 is a flow chart of the present invention;

[0045] Figure 2 It is the control principle diagram of the present invention;

[0046] Figure 3 This is a control schematic diagram of step (4) to step (7) of the present invention;

[0047] Figure 4 In the case of multiple source disturbances Channel observer effect diagram;

[0048] Figure 5 In the case of multiple source disturbances Channel observer effect diagram;

[0049] Figure 6 This is the simulation effect diagram of position trajectory tracking under multi-source disturbance;

[0050] Figure 7 This is the simulation effect diagram of attitude trajectory tracking under multi-source disturbance. DETAILED DESCRIPTION

[0051] The following is combined with Figure 1 To Attachment Figure 7 The present invention is further described.

[0052] The UAV robust control method based on the extended state observer of the present invention comprises the following steps:

[0053] (1) Establish a quadrotor UAV dynamic model based on multi-source disturbances.

[0054] Assuming that the quadrotor drone has uniform and symmetrical rigid body characteristics, its geometric center coincides with the center of gravity. Under the small angle approximation, the attitude angle change rate of the quadrotor drone is directly equivalent to its angular velocity. Based on the Newton-Euler dynamics equation and taking into account the influence of air resistance that the drone is subjected to during flight, the dynamic model of the quadrotor drone is established as follows:

[0055] ;

[0056] in, Indicates the position of the drone; Indicates the roll angle, pitch angle and yaw angle of the drone; represents the total lift provided by the drone propeller, represents the torque acting on the drone; Respectively represent the body coordinate system ( , , ) The moment of inertia of the shaft; Indicates the quality of the drone; is the acceleration due to gravity; represents the moment of inertia of a single propeller; , , represents the translational air resistance coefficient; , , represents the damping moment coefficient; Indicates external interference; represents the propeller speed and, ;in represents the rotation speed of each propeller; the relationship between the rotation speed of the propeller and the total lift and torque it generates can be described as

[0057] ;

[0058] in, represents the total lift provided by the drone propeller, represents the torque acting on the drone, represents the propeller lift coefficient, represents the propeller anti-torque coefficient, Indicates the radius of the fuselage.

[0059] (2) Introduce virtual control variables into the quadrotor UAV dynamics model and determine the control variable allocation strategy.

[0060] In order to decouple attitude control from position control, a virtual control variable is introduced to transform the position control problem into an attitude control problem. By adjusting the thrust of the four motors, the six degrees of freedom of the quadcopter can be controlled indirectly, effectively overcoming the under-actuated characteristics of the system. , , , as follows:

[0061] ;

[0062] By virtual control of the position loop , , By performing an inverse solution of the attitude angle, we can obtain the specific expression of the total lift and the expected input of the attitude, and determine the control amount allocation strategy as follows:

[0063] ;

[0064] in, , , represent the expected inputs of the roll channel, pitch channel, and yaw channel respectively.

[0065] (3) The tracking differentiator is designed using the fastest control synthesis function to track the trajectory signal of the UAV.

[0066] According to the principle of the fastest control synthesis, the tracking differentiator is designed to achieve accurate tracking of the desired trajectory signal. The tracking differentiator is designed using the fastest control synthesis function as follows:

[0067] ;

[0068] in, is the fastest control synthesis function, for The expected input of the channel, is the tracking signal of the expected signal, is the differential signal of the desired signal, is the system speed factor, is the filtering factor, , , Indicates the location coordinates of the drone, , increasing the filter silver factor can effectively enhance the filtering effect.

[0069] To simplify the formula, the following intermediate variables are defined:

[0070] ;

[0071] ;

[0072] in, , , , , , , is the intermediate variable, is the system speed factor, is the filter factor, is a symbolic function.

[0073] (4) Based on the UAV position and attitude information and the tracking signal obtained by the tracking differentiator, the state error information is obtained, the state error system is determined, and the lumped disturbance is introduced to reconstruct the state error system.

[0074] By comparing the UAV system status with the tracking signal obtained in step 3, the position and attitude angle tracking errors of the quadrotor UAV are defined. The state error information is as follows:

[0075] ;

[0076] ;

[0077] ;

[0078] ;

[0079] ;

[0080] ;

[0081] in, Indicates the state error of each channel, Indicates the expected tracking signal of each channel. Indicates the position of the drone; represents the roll angle, pitch angle and yaw angle of the drone, then the state error system is

[0082] ;

[0083] ;

[0084] in, Indicates the state error of each channel, Indicates the expected tracking signal of each channel. They represent the moment of inertia around the (x, y, z) axis of the body coordinate system.

[0085] The system internal modeling uncertainty, external disturbance and the second-order derivative of the desired signal are uniformly defined as lumped disturbances, as follows:

[0086] ;

[0087] ;

[0088] in, represents the total disturbance of each channel, Indicates the expected tracking signal for each channel The second order inverse of , , express , , The position derivative of Indicates the quality of the drone, is the acceleration due to gravity, , , is the translational air resistance coefficient, , , Respectively represent the body coordinate system ( , , ) axis moment of inertia, , , represents the damping torque coefficient, Indicates external interference, represents the propeller speed and, represents the moment of inertia of a single propeller; , , express , , The attitude derivative of .

[0089] , reconstruct the state error system according to the lumped disturbance and unify the state error system into the following form:

[0090] ;

[0091] in, Indicates the state error of each channel The second-order derivative of represents the total disturbance of each channel, represents the control gain, , , , , Indicates the control amount of each channel. , , , , , , , , Represents the torque acting on the drone.

[0092] (5) Based on the reconstructed state error system, combined with adaptive theory and galn function, an extended state observer is designed. Combined with the current control variables of each channel, the state error change rate estimate and the lumped disturbance estimate of each channel are estimated.

[0093] For the reconstructed state error system, the extended state observer is designed by integrating adaptive theory and galn function. By integrating the control input and error signal of each channel, the observer can accurately estimate the state error change rate and lumped disturbance of each channel, and then output accurate estimation information, which provides a solid guarantee for the high-precision robust control of the quadrotor UAV. The designed extended state observer includes the extended state observer of the position loop state error system and the extended state observer of the attitude loop state error system; the extended state observer of the position loop state error system is

[0094] ;

[0095] ;

[0096] ;

[0097] in, , , Represents an intermediate variable , , The first derivative of Indicates the state error of the position channel, is an adaptive law based on a power function, is the position loop adaptive law gain, is the position loop adaptive law parameter, is the position loop adaptive law index, ; is the galn function, , is the adjustable parameter of the galn function, , , Respectively represent the change rate of position loop state error , , The estimated value of , , They represent the lumped disturbance of the position loop , , The estimated value of .

[0098] Similarly, the extended state observer of the attitude loop state error system is

[0099] ;

[0100] ;

[0101] ;

[0102] in, , , Respectively represent the attitude loop state error change rate , , The estimated value of , , They represent the attitude loop lumped disturbance , , The estimated value of .

[0103] (6) Construct a non-singular terminal integral sliding mode controller. According to the reconstructed state error system, the state error change rate and the lumped disturbance, determine the non-singular terminal sliding mode anti-disturbance control strategy and obtain the control quantity of each channel of the UAV.

[0104] According to the reconstructed state error system, state error change rate and lumped disturbance, the non-singular terminal integral sliding surface function is determined; based on the adaptive theory, the sliding mode convergence law adopts the hyperbolic tangent function to obtain the adaptive sliding mode convergence law; according to the first-order derivative of the non-singular terminal integral sliding surface function and the adaptive sliding mode convergence law, the control amount of each channel of the UAV is obtained; the non-singular terminal integral sliding surface function is

[0105] ;

[0106] ;

[0107] in, , , , , , They are , , , , , The sliding surface function of the channel, , , , is the non-singular terminal integral sliding mode adjustable parameter, ; , is a positive odd number and satisfies , t is the current time, , , , , , is the state error of each channel.

[0108] Derivative the sliding surface function to obtain its first-order derivative:

[0109] ;

[0110] ;

[0111] Based on the adaptive theory, an adaptive sliding mode reaching law is designed. The hyperbolic tangent function is used instead of the sign function. The adaptive sliding mode reaching law is:

[0112] ;

[0113] ;

[0114] ;

[0115] ;

[0116] in, , , , , , They represent the first-order derivative of the channel sliding surface function, is the sliding mode reaching law, is the sliding surface function, is the adaptive approach function, , , is the adjustable parameter of the adaptive approach function, , , is the hyperbolic tangent function, is the adaptive parameter, is the adaptive gain coefficient, is the hyperbolic tangent parameter, , is the power exponent of the adaptive parameter, .

[0117] According to the first-order derivative of the sliding surface function and the sliding reaching law, the control quantity of each channel of the UAV is obtained as follows:

[0118] .

[0119] (7) The estimated state error change rate obtained in step (5) is used as the state error change rate of the current UAV.

[0120] The lumped disturbance estimation is used as the lumped disturbance of the current UAV to realize speed sensorless control. The control quantity output by the speed sensorless controller is fed back to the extended state observer in step (5).

[0121] Since the lumped disturbance is composed of unknown internal modeling uncertainty, external disturbance and the second-order derivative of the desired signal, its actual value cannot be directly obtained. Therefore, the lumped disturbance estimate obtained in step (5) is used as a substitute. At the same time, the state error change rate is usually obtained by a speed sensor, but in order to reduce the complexity of control and reduce costs, the present invention does not use a speed sensor. Accordingly, the state error change rate estimate obtained by step (5) is used to meet the control requirements. In this way, we not only ensure the performance of the controller, but also optimize the economy and implementation difficulty of the system. The speed sensorless controller is as follows:

[0122] ;

[0123] in, is the estimated value of the sliding surface function, as follows:

[0124] ;

[0125] .

[0126] (8) According to the control quantity allocation strategy of step (2), the control quantity obtained in step (7) is converted into the thrust command of the motor of the quadrotor UAV, and the real-time position and attitude information of the UAV is fed back to step (4) while controlling the movement of the UAV, thus completing the control closed loop and realizing robust control of the UAV.

[0127] ;

[0128] Through this conversion, efficient control of the quadrotor UAV is achieved, and the closed-loop operation of the entire control system is completed, thus ensuring the stability and reliability of the UAV under multi-source interference conditions.

[0129] In order to verify the effectiveness and rationality of the method proposed in the present invention, the following examples are given:

[0130] In the MATLAB2021 environment, a simulation verification test was carried out on a UAV robust control scheme based on an optimized extended state observer designed in the present invention: First, a simulation module was built in MATLAB / SIMULINK to carry out a simulation experiment of the entire quadrotor aircraft. The model parameters used in the quadrotor aircraft are shown in Table 1.

[0131] Table 1

[0132] ;

[0133] The given quadrotor drone reference trajectory and heading angle are set as

[0134] ;

[0135] in, , , , are the expected horizontal position, expected height and expected yaw angle of the drone, t is the current time, and the initial position and attitude are set to

[0136] ;

[0137] In the present invention, the parameters of the optimized extended state observer and the improved fast non-singular terminal integral sliding mode controller have a direct impact on the performance of the flight control system; the larger the parameter value, the stronger the control ability, but it will also cause a larger control error. If the parameter is too small, the system is prone to oscillation; combined with the control ability and control performance requirements of the quadrotor actuator, the appropriate control parameters are selected as shown in Table 2-4.

[0138] Table 2

[0139] ;

[0140] Table 3

[0141] ;

[0142] Table 4

[0143] ;

[0144] Depend on Figure 4 and Figure 5 It can be seen that the extended state observer of the present invention can observe the speed and unknown disturbance of the quadrotor system in a very short time, which fully demonstrates the excellent performance of the observer. Figure 6 and Figure 7 It can be seen that even in the presence of disturbances, the actual trajectory can quickly catch up with the desired trajectory, and the control effect is very significant, which further confirms the superiority of the controller designed by the present invention in terms of control effect.

Claims

1. A robust control method for unmanned aerial vehicles based on an extended state observer, characterized in that: The following steps are involved: (1) Establish a dynamic model of a quadrotor drone based on multi-source disturbances; (2) Introducing virtual control variables into the quadrotor UAV dynamics model and determining the control variable allocation strategy; (3) Design a tracking differentiator using the fastest control synthesis function to track the trajectory signal of the UAV; (4) Based on the UAV position and attitude information and the tracking signal obtained by the tracking differentiator, the state error information is obtained, the state error system is determined, and the lumped disturbance is introduced to reconstruct the state error system; (5) Based on the reconstructed state error system, combined with adaptive theory and galn function, an extended state observer is designed. Combined with the current control quantity of each channel, the state error change rate estimation and lumped disturbance estimation of each channel are estimated; (6) Construct a non-singular terminal integral sliding mode controller. According to the reconstructed state error system, the state error change rate and the lumped disturbance, determine the non-singular terminal sliding mode active disturbance rejection control strategy and obtain the control quantity of each channel of the UAV. (7) The estimated state error change rate obtained in step (5) is used as the state error change rate of the current UAV. The lumped disturbance estimation is used as the lumped disturbance of the current UAV to realize speed sensorless control, and the control quantity output by the speed sensorless controller is fed back to the extended state observer in step (5); (8) According to the control quantity allocation strategy of step (2), the control quantity obtained in step (7) is converted into the thrust command of the motor of the quadrotor UAV, and the real-time position and attitude information of the UAV is fed back to step (4) while controlling the movement of the UAV, thus completing the control closed loop and realizing robust control of the UAV.

2. The UAV robust control method based on extended state observer according to claim 1 is characterized in that: The control quantity allocation strategy in step (2) is: ； ； in, represents the total lift provided by the drone propeller, , , is the virtual control quantity of the position loop, , , denote the expected inputs of the roll channel, pitch channel, and yaw channel respectively, , , They represent the roll angle, pitch angle and yaw angle of the drone respectively.

3. The UAV robust control method based on extended state observer according to claim 2 is characterized in that: In step (3), the tracking differentiator is designed using the fastest control synthesis function as ； in, is the fastest control synthesis function, for The expected input of the channel, is the tracking signal of the expected signal, is the differential signal of the desired signal, is the system speed factor, is the filtering factor, , , Indicates the position coordinates of the drone, .

4. The UAV robust control method based on extended state observer according to claim 3 is characterized in that: In step (4), the lumped perturbation is ； ； in, represents the total disturbance of each channel, Indicates the expected tracking signal for each channel The second order inverse of , , express , , The position derivative of Indicates the quality of the drone, is the acceleration due to gravity, , , is the translational air resistance coefficient, , , Respectively represent the body coordinate system ( , , ) axis moment of inertia, , , represents the damping torque coefficient, Indicates external interference, represents the propeller speed and, represents the moment of inertia of a single propeller; , , express , , The attitude derivative of .

5. The UAV robust control method based on extended state observer according to claim 4 is characterized in that: In step (4), after introducing the lumped disturbance, the reconstructed state error system is ； in, Indicates the state error of each channel The second-order derivative of represents the total disturbance of each channel, represents the control gain, , , , , Indicates the control amount of each channel. , , , , , , , , Represents the torque acting on the drone.

6. The UAV robust control method based on extended state observer according to claim 5 is characterized in that: In step (5), the design of the extended state observer includes an extended state observer of the position loop state error system and an extended state observer of the attitude loop state error system; the extended state observer of the position loop state error system is ； ； ； in, , , Represents an intermediate variable , , The first derivative of Indicates the state error of the position channel, is an adaptive law based on a power function, is the position loop adaptive law gain, is the position loop adaptive law parameter, is the index of the position loop adaptive law, ; is the galn function, , is the adjustable parameter of the galn function, , , Respectively represent the change rate of position loop state error , , The estimated value of , , They represent the lumped disturbance of the position loop , , An estimated value of The extended state observer of the attitude loop state error system is: ； ； ； in, , , Respectively represent the attitude loop state error change rate , , The estimated value of , , They represent the attitude loop lumped disturbance , , The estimated value of .

7. The UAV robust control method based on extended state observer according to claim 6 is characterized in that: Step (6) is specifically as follows: determining a non-singular terminal integral sliding surface function according to the reconstructed state error system, the state error change rate and the lumped disturbance; Based on the adaptive theory, the sliding mode reaching law adopts the hyperbolic tangent function to obtain the adaptive sliding mode reaching law; According to the first-order derivative of the non-singular terminal integral sliding mode surface function and the adaptive sliding mode reaching law, the control quantity of each channel of the UAV is obtained; The non-singular terminal integral sliding surface function is: ； ； in, , , , , , They are , , , , , The sliding surface function of the channel, , , , is the non-singular terminal integral sliding mode adjustable parameter, ; , is a positive odd number and satisfies , t is the current time, , , , , , is the state error of each channel.

8. The UAV robust control method based on extended state observer according to claim 7 is characterized in that: In step (6), the adaptive sliding mode reaching law is ； ； ； ； in, , , , , , They represent the first-order derivative of the channel sliding surface function, is the sliding mode reaching law, is the sliding surface function, is the adaptive approach function, , , is the adjustable parameter of the adaptive approach function, , , is the hyperbolic tangent function, is the adaptive parameter, is the adaptive gain coefficient, is the hyperbolic tangent parameter, , is the power exponent of the adaptive parameter, .

9. The UAV robust control method based on extended state observer according to claim 8 is characterized in that: In step (6), the control quantity of each channel of the drone is 。 10. The UAV robust control method based on extended state observer according to claim 9 is characterized in that: In step (7), the speed sensorless controller is as follows: ； in, is the estimated value of the sliding surface function, as follows: ； 。

Images