Unmanned aerial vehicle pose control method based on active disturbance rejection and super-spiral sliding mode fusion control

Abstract

The invention relates to an unmanned aerial vehicle pose stability control method based on active disturbance rejection and super-spiral sliding mode fusion control, designs a technology combining active disturbance rejection control (ADRC) and super-spiral sliding mode control (ST-SMC), and realizes performance complementation. Under the condition of random airflow interference, by designing a fusion algorithm, the external disturbance can be stabilized by using the dynamic compensation characteristic of the ADRC, and the influence of the nonlinearity and uncertainty of the system on the control precision can be suppressed through the robustness of the ST-SMC. Compared with traditional PID and ADRC controllers, the method has remarkable advantages in disturbance response speed, stability and tracking precision.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) control technology, and specifically relates to a method for UAV attitude control based on the fusion control of active disturbance rejection and super-helical sliding mode. Background Technology

[0002] With the rapid development of drone technology, quadcopter drones have been widely used in military, agriculture, logistics and other fields. However, in actual flight, quadcopter drones often encounter various uncertain environmental factors, among which random airflow interference is a particularly common and challenging problem. Random airflow can affect the flight stability and control accuracy of drones.

[0003] To study the impact of wind disturbances on UAV control, current UAV control methods mainly include PID, sliding mode, and backstepping control. Some researchers have also proposed using neural networks to compensate for the uncertainties caused by disturbances. Although researchers have made continuous technological innovations in UAV control algorithms, when the random airflow is too large or too complex, UAV tracking always suffers from inaccurate tracking and poor attitude control stability. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings and deficiencies of the prior art and propose a UAV attitude control method based on the fusion control of active disturbance rejection and super-helical sliding mode, which solves the problems of low tracking accuracy and stable attitude control of UAV when encountering random airflow interference.

[0005] To achieve the above objectives, the technical solution provided by this invention is as follows:

[0006] A fusion control method based on active disturbance rejection and superhelical sliding mode includes the following steps:

[0007] Step 1: The motion of the drone's center of mass is controlled by mechanics, and can be described as follows:

[0008]

[0009] Where p = [x, y, z] T It is the position vector of the drone, v = [v x ,v y ,v z ] T This is the velocity vector, m is the mass of the UAV, and g = [0, 0, -g]. T It is the acceleration due to gravity, T = [0, 0, T] T It is the total thrust, along the UAV z b The direction, R, is the rotation matrix of the UAV from the body coordinate system to the inertial coordinate system.

[0010] Step 2: Design an active disturbance rejection controller to perform initial control of the UAV attitude. The active disturbance rejection control law is used as the initial control law and combined with the super-helical sliding mode control law in Step 3 in the fusion controller.

[0011] Step 3: Design a superhelical sliding mode controller. Similarly, the superhelical sliding mode control law is used as the initial control law and transmitted to the fusion controller for integration.

[0012] Step 4: Combine the active disturbance rejection control law in Step 3 and the super-helical sliding mode control law in Step 4 to form a new fused control law to control the attitude of the UAV.

[0013] Step 5: To ensure the designed controller has strong robustness, tracking accuracy, and dynamic response characteristics, the sliding surface dynamics are designed.

[0014] Step 6: Perform Lyapunov stability verification on the designed fusion controller;

[0015] Furthermore, the workflow for establishing the UAV dynamics model in step 1 includes the following steps:

[0016] Step 1.1: The attitude of the UAV is described by Euler angles, including roll, pitch, and yaw. Its dynamic model is as follows:

[0017]

[0018] Where ω=[p,q,r] T It is the angular velocity (mechanical system), I is the inertial matrix of the UAV, τ=[τ φ ,τ θ ,τ ψ ] T It refers to the control torque, η=[φ,θ,ψ] T It's Euler angles, T η It is a nonlinear mapping matrix of angle and angular velocity.

[0019] Step 1.2: The complete state-space model of the UAV is as follows:

[0020]

[0021] Where the state x = [p, v, η, ω] T Control input u = [T, τ φ ,τ θ ,τ ψ ] T .

[0022] Furthermore, the design of the UAV active disturbance rejection controller in step 2 includes the following steps:

[0023] Step 2.1: Introduce a perturbation d(t) into the model, and rewrite the dynamics as follows:

[0024]

[0025] Step 2.2: Estimate d(t) and system state using ESO:

[0026]

[0027] Where z1, z2, and z3 correspond to the values, first derivative, and second derivative of the observed variable, respectively, y is the measured system output, and β1, β2, and β3 are the gain coefficients of the ESO.

[0028] Step 2.3: To ensure the convergence and stability of the ESO, the gain coefficients β1, β2, and β3 can be determined using the pole placement method. Assuming the ESO's poles are located at -ω1, the gain coefficients are designed as follows:

[0029] β1=3ω0,

[0030] Step 2.4: ADRC control law based on ESO output:

[0031]

[0032] in, It is an estimate of the system dynamics. It is the estimated disturbance, k s It is the error feedback gain.

[0033] Step 2.5: To compensate for the unknown part of the system f(x)+d(t), define u ADRC for:

[0034]

[0035] in, and These are the ESO estimates of f(x) and d(t). Substituting them into the equation eliminates most of the unknown dynamics, and the remaining term represents the estimation error. and

[0036] Furthermore, the design of the UAV superhelical sliding mode controller in step 3 includes the following steps:

[0037] Step 3.1: The sliding surface is defined as:

[0038]

[0039] Where, e = xx refFor tracking error, λ is the sliding surface gain.

[0040] Step 3.2: The sliding surface gain determines the convergence rate of the sliding surface. Typically, it can be designed using the following formula:

[0041] k e =2ζω n

[0042] Where ζ is the damping ratio, ω n It is the natural frequency. By adjusting ζ and ω n It can control the convergence speed of the sliding surface.

[0043] Step 3.3: Incorporating higher-order terms into superspiral control to reduce chattering:

[0044]

[0045] Where k1 and k2 are sliding mode gains.

[0046] Step 3.4: To ensure the stability of the superspiral sliding mode control, the sliding mode gain needs to meet the following conditions:

[0047]

[0048] Here, ε is the upper bound of the system's uncertainty, and α is a positive number, usually taken as 1. These conditions ensure the robustness and stability of sliding mode control.

[0049] Step 3.5: To further suppress chattering in sliding mode control, a saturation function is added to the control law, resulting in the following control system:

[0050]

[0051] Where sat(S) is the saturation function, defined as:

[0052]

[0053] In the saturation function, δ is a positive number used to control the smoothness of the saturation function.

[0054] Furthermore, the design of the UAV fusion controller in step 4 includes the following steps in its workflow:

[0055] Step 4.1: Use the extended state observer in active disturbance rejection control to observe the system state and disturbances in real time. The state equation is:

[0056]

[0057]

[0058]

[0059] Where z1, z2, and z3 correspond to the values, first derivative, and second derivative of the observed variable, respectively, y is the measured system output, and β1, β2, and β3 are the gain coefficients of the ESO.

[0060] Step 4.1: Using the state observer in active disturbance rejection control, the system state and disturbance can be estimated as follows:

[0061]

[0062] Step 4.2: Disturbance compensation can eliminate the influence of system dynamics and external disturbances. The disturbance from the state observer is introduced into the control law for compensation. Therefore, the disturbance compensation term for active disturbance rejection control is:

[0063]

[0064] Step 4.3: By adding a saturation function to the original control law, chattering in sliding mode control is suppressed. The control law is:

[0065]

[0066] Step 4.4: The final control input u consists of an active disturbance rejection term and a super-helical sliding mode control:

[0067] u = u ADRC +u ST-SMC

[0068] Step 4.5: Put u ADRC and u ST-SMC Substituting into step 4.4, we get:

[0069]

[0070] Step 4.6: After processing, the final control input is obtained:

[0071]

[0072] Furthermore, the dynamic design of the sliding surface in step 5 includes the following steps:

[0073] Step 5.1: Based on the definition of the sliding surface S in Step 3.1, the sliding surface dynamics... It can be expressed as:

[0074]

[0075] Step 5.2: Substitute the final control input u to obtain:

[0076]

[0077] Step 5.3: Assuming the trajectory tracking error is zero, then:

[0078]

[0079] Step 5.4: Further simplified to:

[0080]

[0081] Furthermore, the design of Lyapunov stability in step 6 involves the following steps:

[0082] Step 6.1: Define the Lyapunov function as:

[0083]

[0084] Step 6.2: The derivative is:

[0085]

[0086] Step 6.3: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require the full context.] The terms are broken down into disturbance terms and sliding mode control terms, which will be discussed separately. First, the contribution of the disturbance term will be considered, since... and If it is bounded, then:

[0087]

[0088] Step 6.4: Next, consider the contribution of the superhelical sliding mode term. The two parts of the superhelical sliding mode control are -k1|S| 1 / 2 sat(S) and -k2S:

[0089] (1) For -k1|S| 1 / 2 Then sat(S) has:

[0090] S(-k1|S| 1 / 2 sat(S))=-k1|S| 1 / 2 S·sat(S)

[0091] Since sat(S) is sign(S) when |S|>δ and is [sign(S)] when |S|≤δ, [sign(S)] therefore:

[0092]

[0093]

[0094] (2) For -k2S, we have:

[0095] S(-k2S)=-k2S 2

[0096] Combining the disturbance term and the sliding mode control term, we get:

[0097]

[0098] Step 6.5: Based on the different cases of the sat(S) function, it can be divided into the following two main cases for detailed discussion.

[0099] Case 1: When |S|>δ, then sat(S)=sign(S), therefore:

[0100]

[0101] To ensure The following conditions must be met:

[0102]

[0103] Case 2: When |S|≤δ, then therefore:

[0104]

[0105] To ensure The following conditions must be met:

[0106]

[0107] Step 6.6: Finally, to ensure the asymptotic stability of the system in the Lyapunov sense, the sliding mode gain must satisfy the following conditions:

[0108] (1) When |S|>δ:

[0109]

[0110] (2) When |S|≤δ:

[0111]

[0112] When k1 and k2 satisfy the above conditions, the system is asymptotically stable in the Lyapunov sense.

[0113] The present invention proposes a fusion control method based on active disturbance rejection and superspiral sliding mode, which has the following advantages compared with the prior art:

[0114] 1. Improved system stability: The extended state observer in active disturbance rejection control can always estimate and compensate for disturbances in real time. Therefore, this algorithm can predict real-time disturbances of random disturbances under random airflow conditions, and can specifically address the system's disturbance rejection performance and response speed.

[0115] 2. Improve system smoothness: By introducing advanced sliding mode technology, the high-frequency jitter caused by traditional sliding mode control can be eliminated. It can not only track the system state more smoothly, but also avoid energy consumption and instability caused by switching of multiple sliding surfaces during rapid response.

[0116] 3. Optimized control accuracy: Through the active disturbance rejection optimization algorithm, disturbances are compensated in real time, ensuring high accuracy of UAV attitude control, especially when dealing with complex environments or disturbances, showing better control performance. Attached Figure Description

[0117] Figure 1 This is a block diagram illustrating the principle of an attitude control method for unmanned aerial vehicles (UAVs) based on active disturbance rejection and superspiral sliding mode proposed in this invention.

[0118] Figure 2 This is a flowchart of the UAV attitude control algorithm based on active disturbance rejection and superhelical sliding mode proposed in this invention.

[0119] Figure 3 The image shows the trajectory tracking curve along the X-axis in Experiment Example 1.

[0120] Figure 4 The Y-axis trajectory tracking curve is shown in Experiment Example 1.

[0121] Figure 5 The curve shown is the Z-axis trajectory tracking curve in Experiment Example 1.

[0122] Figure 6 The image shows the roll angle trajectory tracking curve in Experiment Example 1.

[0123] Figure 7 The pitch angle trajectory tracking curve is shown in Experiment Example 1.

[0124] Figure 8 The yaw angle trajectory tracking curve is shown in Experiment Example 1.

[0125] Figure 9 The X-axis error tracking curve is shown in Experiment Example 1.

[0126] Figure 10 This is the Y-axis error tracking curve in Experiment Example 1.

[0127] Figure 11 This is the Z-axis error tracking curve in Experiment Example 1.

[0128] Figure 12 This is the roll angle error tracking curve in Experiment Example 1.

[0129] Figure 13 This is the pitch angle error tracking curve in Experiment Example 1.

[0130] Figure 14 The yaw angle error tracking curve is shown in Experiment Example 1.

[0131] Figure 15 The roll angle and angular velocity curves are from Experiment Example 1.

[0132] Figure 16 The pitch angle and angular velocity curves are from Experiment Example 1.

[0133] Figure 17 The yaw angle and angular velocity curves are from Experiment Example 1. Detailed Implementation

[0134] The present application will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present application, but do not limit the present application in any way. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present application. These fall within the scope of protection of the present application.

[0135] Example 1

[0136] This application discloses a fusion control method based on active disturbance rejection and superhelical sliding mode, which includes the following steps:

[0137] Step 1: The motion of the drone's center of mass is controlled by mechanics, and can be described as follows:

[0138]

[0139] Where p = [x, y, z] T It is the position vector of the drone, v = [v x ,v y ,v z ] T This is the velocity vector, m is the mass of the UAV, and g = [0, 0, -g]. T It is the acceleration due to gravity, T = [0, 0, T] T It is the total thrust, along the UAV z b The direction, R, is the rotation matrix of the UAV from the body coordinate system to the inertial coordinate system.

[0140] Step 2: Design an active disturbance rejection controller to perform initial control of the UAV attitude. The active disturbance rejection control law is used as the initial control law and combined with the super-helical sliding mode control law in Step 3 in the fusion controller.

[0141] Step 3: Design a superhelical sliding mode controller. Similarly, the superhelical sliding mode control law is used as the initial control law and transmitted to the fusion controller for integration.

[0142] Step 4: Combine the active disturbance rejection control law in Step 3 and the super-helical sliding mode control law in Step 4 to form a new fused control law to control the attitude of the UAV.

[0143] Step 5: To ensure the designed controller has strong robustness, tracking accuracy, and dynamic response characteristics, the sliding surface dynamics are designed.

[0144] Step 6: Perform Lyapunov stability verification on the designed fusion controller;

[0145] Furthermore, the workflow for establishing the UAV dynamics model in step 1 includes the following steps:

[0146] Step 1.1: The attitude of the UAV is described by Euler angles, including roll, pitch, and yaw. Its dynamic model is as follows:

[0147]

[0148] Where ω=[p,q,r] T It is the angular velocity (mechanical system), I is the inertial matrix of the UAV, τ=[τ φ ,τ θ ,τ ψ ] T It refers to the control torque, η=[φ,θ,ψ] T It's Euler angles, T η It is a nonlinear mapping matrix of angle and angular velocity.

[0149] Step 1.2: The complete state-space model of the UAV is as follows:

[0150]

[0151] Where the state x = [p, v, η, ω] T Control input u = [T, τ φ ,τ θ ,τ ψ ] T .

[0152] Furthermore, the design of the UAV active disturbance rejection controller in step 2 includes the following steps:

[0153] Step 2.1: Introduce a perturbation d(t) into the model, and rewrite the dynamics as follows:

[0154]

[0155] Step 2.2: Estimate d(t) and system state using ESO:

[0156]

[0157] Where z1, z2, and z3 correspond to the values, first derivative, and second derivative of the observed variable, respectively, y is the measured system output, and β1, β2, and β3 are the gain coefficients of the ESO.

[0158] Step 2.3: To ensure the convergence and stability of the ESO, the gain coefficients β1, β2, and β3 can be determined using the pole placement method. Assuming the ESO's poles are located at -ω1, the gain coefficients are designed as follows:

[0159] β1=3ω0,

[0160] Step 2.4: ADRC control law based on ESO output:

[0161]

[0162] in, It is an estimate of the system dynamics. It is the estimated disturbance, k s It is the error feedback gain.

[0163] Step 2.5: To compensate for the unknown part of the system f(x)+d(t), define u ADRC for:

[0164]

[0165] in, and These are the ESO estimates of f(x) and d(t). Substituting them into the equation eliminates most of the unknown dynamics, and the remaining term represents the estimation error. and

[0166] Furthermore, the design of the UAV superhelical sliding mode controller in step 3 includes the following steps:

[0167] Step 3.1: The sliding surface is defined as:

[0168]

[0169] Where, e = xx ref For tracking error, λ is the sliding surface gain.

[0170] Step 3.2: The sliding surface gain determines the convergence rate of the sliding surface. Typically, it can be designed using the following formula:

[0171] k e =2ζω n

[0172] Where ζ is the damping ratio, ω n It is the natural frequency. By adjusting ζ and ω n It can control the convergence speed of the sliding surface.

[0173] Step 3.3: Incorporating higher-order terms into superspiral control to reduce chattering:

[0174]

[0175] Where k1 and k2 are sliding mode gains.

[0176] Step 3.4: To ensure the stability of the superspiral sliding mode control, the sliding mode gain needs to meet the following conditions:

[0177]

[0178] Here, ε is the upper bound of the system's uncertainty, and α is a positive number, usually taken as 1. These conditions ensure the robustness and stability of sliding mode control.

[0179] Step 3.5: To further suppress chattering in sliding mode control, a saturation function is added to the control law, resulting in the following control system:

[0180]

[0181] Where sat(S) is the saturation function, defined as:

[0182]

[0183] In the saturation function, δ is a positive number used to control the smoothness of the saturation function.

[0184] Furthermore, the design of the UAV fusion controller in step 4 includes the following steps in its workflow:

[0185] Step 4.1: Use the extended state observer in active disturbance rejection control to observe the system state and disturbances in real time. The state equation is:

[0186]

[0187]

[0188]

[0189] Where z1, z2, and z3 correspond to the values, first derivative, and second derivative of the observed variable, respectively, y is the measured system output, and β1, β2, and β3 are the gain coefficients of the ESO.

[0190] Step 4.1: Using the state observer in active disturbance rejection control, the system state and disturbance can be estimated as follows:

[0191]

[0192] Step 4.2: Disturbance compensation can eliminate the influence of system dynamics and external disturbances. The disturbance from the state observer is introduced into the control law for compensation. Therefore, the disturbance compensation term for active disturbance rejection control is:

[0193]

[0194] Step 4.3: By adding a saturation function to the original control law, chattering in sliding mode control is suppressed. The control law is:

[0195]

[0196] Step 4.4: The final control input u consists of an active disturbance rejection term and a super-helical sliding mode control:

[0197] u = u ADRC +u ST-SMC

[0198] Step 4.5: Put u ADRC and u ST-SMC Substituting into step 4.4, we get:

[0199]

[0200] Step 4.6: After processing, the final control input is obtained:

[0201]

[0202] Furthermore, the dynamic design of the sliding surface in step 5 includes the following steps:

[0203] Step 5.1: Based on the definition of the sliding surface S in Step 3.1, the sliding surface dynamics... It can be expressed as:

[0204]

[0205] Step 5.2: Substitute the final control input u to obtain:

[0206]

[0207] Step 5.3: Assuming the trajectory tracking error is zero, then:

[0208]

[0209] Step 5.4: Further simplified to:

[0210]

[0211] Furthermore, the design of Lyapunov stability in step 6 involves the following steps:

[0212] Step 6.1: Define the Lyapunov function as:

[0213]

[0214] Step 6.2: The derivative is:

[0215]

[0216] Step 6.3: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require the full context.] The terms are broken down into disturbance terms and sliding mode control terms, which will be discussed separately. First, the contribution of the disturbance term will be considered, since... and If it is bounded, then:

[0217]

[0218] Step 6.4: Next, consider the contribution of the superhelical sliding mode term. The two parts of the superhelical sliding mode control are -k1|S| 1 / 2 sat(S) and -k2S:

[0219] (1) For -k1|S| 1 / 2 Then sat(S) has:

[0220] S(-k1|S| 1 / 2 sat(S))=-k1|S| 1 / 2 S·sat(S)

[0221] Since sat(S) is sign(S) when |S|>δ and is [sign(S)] when |S|≤δ, [sign(S)] therefore:

[0222]

[0223]

[0224] (2) For -k2S, we have:

[0225] S(-k2S)=-k2S 2

[0226] Combining the disturbance term and the sliding mode control term, we get:

[0227]

[0228] Step 6.5: Based on the different cases of the sat(S) function, it can be divided into the following two main cases for detailed discussion.

[0229] Case 1: When |S|>δ, then sat(S)=sign(S), therefore:

[0230]

[0231] To ensure The following conditions must be met:

[0232]

[0233] Case 2: When |S|≤δ, then therefore:

[0234]

[0235] To ensure The following conditions must be met:

[0236]

[0237] Step 6.6: Finally, to ensure the asymptotic stability of the system in the Lyapunov sense, the sliding mode gain must satisfy the following conditions:

[0238] (1) When |S|>δ:

[0239]

[0240] (2) When |S|≤δ:

[0241]

[0242] When k1 and k2 satisfy the above conditions, the system is asymptotically stable in the Lyapunov sense.

[0243] The following simulation demonstrates the effectiveness and feasibility of the control strategy based on the fusion control method of active disturbance rejection and superhelical sliding mode disclosed in this application, as shown in the following details:

[0244] The parameters of the UAV in the simulation experiment are listed below: mass of the quadcopter UAV m = 0.698 kg, distance from the center of the rotor to the center of mass d = 0.171 m, and gravitational acceleration g = 9.8 m / s². -2 Lift coefficient C T=7.6184×10 (-8) ×(60 / (2×pi)) 2 N·s -2 X / Y moment of inertia J xx / J yy =0.0034 kg·m 2 Anti-torque coefficient C m =2.6839×10 (-9) ×(60 / (2×pi)) 2 N·m·s -2 Rotor moment of inertia J m =1.302×10 (-6) Kg·m 2 and Z-axis rotational inertia J zz =0.0060 kg·m 2 Gain coefficients β1 = 30, β2 = 300 and β3 = 1000, sliding mode gain k1 = 20.

[0245] The simulation results are as follows:

[0246] The figure shows the target tracking curve. The proportional-integral-derivative (PID) controller and active disturbance rejection controller (ADRC) are the control experimental curves. The attitude control algorithm designed in this invention is based on the fusion control of active disturbance rejection and super-spiral sliding mode (A-ST-SMC). Figures 3-8 Analysis of the curves reveals that initially, the system experiences significant fluctuations in the drone due to the influence of disturbed airflow. When random airflow disturbances are introduced between 0 and 5 seconds, observation of the PID and ADRC control curves reveals significant overshoot and oscillations in their output curves, indicating insufficient stability of the drone control under random airflow disturbances. However, observation of the A-ST-SMC shows that its overshoot is significantly lower than that of the traditional PID and ADRC algorithms, and its oscillation frequency and amplitude are superior, demonstrating stronger robustness against random airflow disturbances and a faster recovery to stability. In the final 5 seconds, all three control algorithms gradually stabilize, but comparison shows that the A-ST-SMC stabilizes in less time than PID and ADRC, indicating that PID and ADRC perform worse than A-ST-SMC when dealing with random airflow disturbances.

[0247] Observing the attitude angles reveals two phases: the initial stage from 0s to 5s and the steady-state stage after 5s. In the initial stage, where random airflow is introduced, the PID and ADRC control algorithm curves exhibit significant fluctuations and show substantial deviations after being disturbed by random airflow, requiring a considerable amount of time to recover stability. In contrast, the A-ST-SMC-based approach demonstrates better disturbance rejection performance. The curves show that with the introduction of random airflow, the A-ST-SMC curve exhibits almost no overshoot or prolonged jitter, and it converges rapidly to zero in the steady-state phase.

[0248] Figures 9-11 Error curves for X, Y, and Z were tracked. Random airflow disturbance was introduced from the initial 0s to 5s, and the output curves of the traditional control algorithm PID, ADRC, and ADRC-based super-spiral sliding mode control algorithm were observed to analyze the response performance of the three algorithms to random airflow disturbance. Observation of the curves revealed that during the disturbance phase, the error of the traditional control algorithm's output curve also exhibited more pronounced jitter. Although the amplitude of the jitter was relatively small, the time it took for the amplitude to decrease from large to small was relatively long.

[0249] Figures 12-14 The graph shows the attitude angle error. It can be observed that traditional control algorithms always exhibit significant overshoot during the disturbance phase, and the curves show a long jitter time. Although A-ST-SMC experiences an overshoot momentarily, the controller adjustment can always reduce the overshoot when the next jitter occurs, and the jitter amplitude is significantly lower than that of PID and ADRC control algorithms.

[0250] Figures 15-17 The graph shows the attitude angle and angular velocity curves. The results show that after encountering airflow, the UAV adjusts its attitude angular velocity to achieve balance. Comparing traditional PID control, active disturbance rejection (ADR) control, and a fusion controller based on ADR and super-helical sliding mode, it can be seen that the UAV can quickly achieve attitude stability under the fusion controller's control. However, PID and ADR are significantly faster and have much greater overshoot than the fusion controller, both in terms of response speed and overshoot.

[0251] The technical means disclosed in this invention are not limited to those disclosed in the above embodiments, but also include technical solutions that are any combination of the above technical features. It should be noted that for those skilled in the art, various improvements and modifications can be made without departing from the principle of this invention, and these modifications are also considered within the scope of protection of this invention.

Claims

1. A method for UAV attitude control based on the fusion of active disturbance rejection (ADRROC) and superspiral sliding mode control is provided for UAV flight attitude control. Since the extended state observer in ADRROC can always estimate and compensate for disturbances in real time, this algorithm can predict the real-time disturbances of random disturbances under random airflow conditions, and can specifically address the system's disturbance rejection performance and response speed. Similarly, superspiral sliding mode control, by introducing high-order sliding mode technology, can eliminate the high-frequency jitter phenomenon caused by traditional sliding mode control. The superspiral sliding mode algorithm can not only track the system state more smoothly, but also avoid energy consumption and instability caused by switching multiple sliding surfaces during rapid response, thus compensating for the shortcomings of ADRROC in robustness and smoothness. By combining the above two control algorithms, the UAV can achieve rapid response speed and disturbance rejection performance in suppressing external disturbances under random airflow conditions, as well as improve the robustness and accuracy of the control algorithm.

2. The UAV pose control method based on the fusion control of active disturbance rejection and superspiral sliding mode as described in claim 1, characterized in that, The method may include the following steps: Step 1: The motion of the drone's center of mass is controlled by mechanics, which can be described as: Where p = [x, y, z] T It is the position vector of the drone, v = [v x ,v y ,v z ] T This is the velocity vector, m is the mass of the UAV, and g = [0, 0, -g]. T It is the acceleration due to gravity, T = [0, 0, T] T It is the total thrust, along the UAV z b Direction, R is the rotation matrix of the UAV from the body coordinate system to the inertial coordinate system; Step 2: Design an active disturbance rejection controller to perform initial control of the UAV attitude. The active disturbance rejection control law is used as the initial control law and combined with the super-helical sliding mode control law in Step 3 in the fusion controller. Step 3: Design a superhelical sliding mode controller. Similarly, the superhelical sliding mode control law is used as the initial control law and transmitted to the fusion controller for integration. Step 4: Combine the active disturbance rejection control law in Step 3 and the super-helical sliding mode control law in Step 4 to form a new fused control law to control the attitude of the UAV. Step 5: To ensure the designed controller has strong robustness, tracking accuracy, and dynamic response characteristics, the sliding surface dynamics are designed. Step 6: Perform Lyapunov stability verification on the designed fusion controller.

3. The UAV pose control method based on the fusion control of active disturbance rejection and superspiral sliding mode as described in claim 2, characterized in that, The workflow for establishing the UAV dynamics model described in step 1 includes the following steps: Step 1.1: The attitude of the UAV is described by Euler angles, including roll, pitch, and yaw angles. Its dynamic model is as follows: Where ω=[p,q,r] T It is the angular velocity (mechanical system), I is the inertial matrix of the UAV, τ=[τ φ ,τ θ ,τ ψ ] T It refers to the control torque, η=[φ,θ,ψ] T It's Euler angles, T η It is a nonlinear mapping matrix between angle and angular velocity; Step 1.2: The complete state-space model of the UAV is as follows: Where the state x = [p, v, η, ω] T Control input u = [T, τ φ ,τ θ ,τ ψ ] T .

4. The UAV pose control method based on the fusion control of active disturbance rejection and superspiral sliding mode as described in claim 2, characterized in that, The design of the UAV active interference rejection controller described in step 2 includes the following steps: Step 2.1: Introduce a perturbation d(t) into the model, and rewrite the dynamics as follows: Step 2.2: Estimate d(t) and system state using ESO: Where z1, z2, and z3 correspond to the values, first derivative, and second derivative of the observed variable, respectively, y is the measured system output, and β1, β2, and β3 are the gain coefficients of the ESO. Step 2.3: To ensure the convergence and stability of the ESO, the gain coefficients β1, β2, and β3 can be determined using the pole placement method. Assuming the ESO's poles are located at -ω1, the gain coefficients are designed as follows: β1=3ω0, Step 2.4: ADRC control law based on ESO output: in, It is an estimate of the system dynamics. It is the estimated disturbance, k s It is the error feedback gain; Step 2.5: To compensate for the unknown part of the system f(x)+d(t), define u ADRC for: in, and These are the ESO estimates of f(x) and d(t). Substituting them into the equation eliminates most of the unknown dynamics, and the remaining term represents the estimation error. and 5. The UAV pose control method based on the fusion control of active disturbance rejection and superspiral sliding mode as described in claim 2, characterized in that, The design of the UAV superhelical sliding mode controller described in step 3 includes the following steps: Step 3.1: The sliding surface is defined as: Where, e = xx ref For tracking error, λ is the sliding surface gain; Step 3.2: The sliding surface gain determines the convergence rate of the sliding surface, which can typically be designed using the following formula: k e =2lives n Where ζ is the damping ratio, ω n It is the natural frequency, obtained by adjusting ζ and ω. n This allows control over the convergence speed of the sliding surface; Step 3.3: Incorporating higher-order terms into superspiral control to reduce chattering: Where k1 and k2 are sliding mode gains; Step 3.4: To ensure the stability of the superspiral sliding mode control, the sliding mode gain needs to meet the following conditions: Here, ε is the upper bound of the system's uncertainty, and α is a positive number, usually taken as 1; these conditions ensure the robustness and stability of sliding mode control. Step 3.5: To further suppress chattering in sliding mode control, a saturation function is added to the control law, resulting in the following control system: Where sat(S) is the saturation function, defined as: In the saturation function, δ is a positive number used to control the smoothness of the saturation function.

6. The UAV pose control method based on the fusion control of active disturbance rejection and superspiral sliding mode as described in claim 2, characterized in that, The design of the UAV fusion controller described in step 4 includes the following steps: Step 4.1: Use the extended state observer in active disturbance rejection control to observe the system state and disturbances in real time. The state equation is: Where z1, z2, and z3 correspond to the values, first derivative, and second derivative of the observed variable, respectively; y is the measured system output; and β1, β2, and β3 are the gain coefficients of the ESO. Step 4.1: Using the state observer in active disturbance rejection control, the system state and disturbance can be estimated as follows: Step 4.2: Disturbance compensation can eliminate the influence of system dynamics and external disturbances. The disturbance from the state observer is introduced into the control law for compensation. Therefore, the disturbance compensation term for active disturbance rejection control is: Step 4.3: By adding a saturation function to the original control law, chattering in sliding mode control is suppressed. The control law is: Step 4.4: The final control input u consists of an active disturbance rejection term and a super-helical sliding mode control: in=in ADRC +in ST-SMC Step 4.5: Put u ADRC and u ST-SMC Substituting into step 4.4, we get: Step 4.6: After processing, the final control input is obtained:

7. The UAV pose control method based on the fusion control of active disturbance rejection and superspiral sliding mode as described in claim 2, characterized in that, The dynamic design of the sliding surface described in step 5 includes the following steps: Step 5.1: Based on the definition of the sliding surface S in Step 3.1, the sliding surface dynamics... It can be expressed as: Step 5.2: Substitute the final control input u to obtain: Step 5.3: Assuming the trajectory tracking error is zero, then: Step 5.4: Further simplified to:

8. The UAV pose control method based on the fusion control of active disturbance rejection and superspiral sliding mode as described in claim 2, characterized in that, The design of Lyapunov stability described in step 6 includes the following steps: Step 6.1: Define the Lyapunov function as: Step 6.2: The derivative is: Step 6.3: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require the full context.] The terms are broken down into disturbance terms and sliding mode control terms, and will be introduced separately. First, the contribution of the disturbance term will be considered. and If it is bounded, then: Step 6.4: Next, consider the contribution of the superspiral sliding mode term; where the two parts of the superspiral sliding mode control are -k1|S| 1 / 2 sat(S) and -k2S: (1) For -k1|S| 1 / 2 Then sat(S) has: S(-k1|S| 1 / 2 sat(S))=-k1|S| 1 / 2 S·sat(S) Since sat(S) is sign(S) when |S|>δ and is [sign(S)] when |S|≤δ, [sign(S)] therefore: (2) For -k2S, we have: S(-k2S)=-k2S 2 Combining the disturbance term and the sliding mode control term, we get: Step 6.5: Based on the different cases of the sat(S) function, it can be divided into the following two main cases for detailed discussion; Case 1: When |S|>δ, then sat(S)=sign(S), therefore: To ensure The following conditions must be met: Case 2: When |S|≤δ, then therefore: To ensure The following conditions must be met: Step 6.6: Finally, to ensure the asymptotic stability of the system in the Lyapunov sense, the sliding mode gain must satisfy the following conditions: (1) When |S|>δ: (2) When |S|≤δ: When k1 and k2 satisfy the above conditions, the system is asymptotically stable in the Lyapunov sense.

Images