A method and control system for reducing sway in suspended flight of unmanned aerial vehicle

Abstract

The present invention discloses a method and control system for reducing the sway of a suspended UAV in flight. The method uses an observer to observe the suspended UAV flight system using a control signal U and an observer, and adds an estimated value obtained by the observation when generating a new control signal to offset the adverse effects of disturbances on the flight control of the system. Furthermore, the control signal U output by the control method of the present invention enables the UAV to achieve high-precision position tracking and suppress the sway of the suspended load, ensuring that the UAV position tracking error and the load sway angle vary within a constrained range. The UAV suspended flight reduction control system of the present invention is used to implement the above-mentioned control method. By observing the UAV through an observer to obtain an estimated value, and outputting a control signal to the UAV flight system through a backstepping controller, the UAV can achieve precise control of the UAV position and sway reduction control of the suspended load.

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicle (UAV) control, and in particular to a sway reduction control method and control system for a UAV suspended flight. Background Art

[0002] There are two existing methods for transporting items using drones: one is to secure the payload to the drone, such as in drone food delivery; the other is to suspend the payload from the drone via a cable. The first method avoids the problem of sway suppression and treats the drone and payload as a single entity for position and attitude control. PID control is the primary method employed. However, securing the payload to the drone affects attitude control, reducing flight maneuverability. Furthermore, the shape and clamping mechanism of the suspended payload must be considered, increasing the complexity and weight of the transport system and reducing the drone's flight time. The second method, since the payload is suspended from the drone via a cable, utilizes a simpler, more practical method. While the system gains two degrees of freedom, the control input remains the same, exacerbating the underactuation problem and increasing the difficulty of implementing position and sway control. Currently employed methods include PID control, linear quadratic regulator (LQR) control, active disturbance rejection control, backstepping control, sliding mode control, and feedback linearization.

[0003] In the prior art:

[0004] Linear control mainly includes PID and LQR control, but the UAV's suspended flight system is nonlinear, underdriven, and strongly coupled. If the system is limited to hovering or small-range waypoint flight, linear control can be used. However, most situations in UAV suspended transportation do not meet linearization conditions. Moreover, when using linear control, the swing angle of the UAV's suspended load is very large, which in turn affects the UAV's flight stability.

[0005] The main nonlinear control methods include active disturbance rejection control, backstepping control, sliding mode control, feedback linearization and other methods. Among them, feedback linearization depends on an accurate mathematical model, and not all systems can be feedback linearized, and the system stability changes with the different controlled outputs; the disadvantage of the sliding mode control method is that when the system state trajectory reaches the sliding mode surface, it is difficult to slide strictly along the sliding mode surface to the equilibrium point, but it approaches the equilibrium point by crossing back and forth on both sides of the sliding mode surface, which is prone to high-frequency vibration; backstepping control also relies on an accurate model of the system, so it is not robust to noise and interference, and the calculation and stability analysis of the backstepping controller are relatively cumbersome.

[0006] In summary, there is an urgent need for a method and control system for reducing the sway of a UAV during suspended flight to solve the problems of sway suppression and wind disturbance resistance in the prior art of UAVs during flight with suspended loads. Summary of the Invention

[0007] The present invention aims to provide a method and control system for reducing sway during the flight of a suspended UAV, so as to solve the problems of sway suppression and wind disturbance resistance in the prior art of UAVs with suspended loads. The specific technical solutions are as follows:

[0008] A method for reducing sway during suspended flight of an unmanned aerial vehicle comprises the following steps:

[0009] Step S1: Establish a dynamic model of the UAV hanging flight system;

[0010] Step S2: Based on the dynamic model, the total disturbance of the UAV is observed by an observer and combined with the control signal to obtain an estimated value of the total disturbance;

[0011] Step S3: Based on the desired position, the output of the UAV's suspension flight system, and the estimated value in step S2, a control signal is generated. The control signal is transmitted to the dynamic model in step S1 to update the state of the UAV, and the control signal is returned to step S2 to calculate a new estimated value.

[0012] Step S4: Return the updated output of the UAV suspension flight system to steps S2 and S3 respectively to calculate a new estimated value and a new control signal.

[0013] The above technical solution is preferably, in step S1, the dynamic model of the UAV hanging flight system is as shown in formula 1):

[0014]

[0015] Among them, q represents the state vector of the UAV suspension flight system; F u represents the total disturbance to the UAV; M c0 Indicates M c (q) -1 Nominal value, M c (q) represents the inertia matrix of the UAV suspension flight system; U represents the control signal.

[0016] The above technical solution is preferred, in formula 1), F u As shown in formula 1.1):

[0017]

[0018] in, represents the centripetal force matrix of the UAV hanging flight system; G(q) represents the gravity vector of the UAV hanging flight system; F d represents the air resistance vector;

[0019] M c (q), G(q) and F d As shown below:

[0020]

[0021]

[0022] G(q)=[0 (m q +m l )gm l glsinγ] T 1.13);

[0023]

[0024] Among them, m q Indicates the mass of the drone; m l represents the mass of the load; l represents the length of the string used to suspend the load; γ represents the swing angle between the load and the vertical direction;

[0025] d x d z and d γ Both represent the air damping coefficient; and They represent the speed of the UAV in the x and z directions in the inertial coordinate system respectively.

[0026] The above technical solution is preferred. In step S2, an observer is used to observe the total disturbance of the UAV, as shown in formula 2):

[0027]

[0028] in, represents the estimated value of q; Express The estimated value of ; L is the gain matrix; is the state error vector; is the compensation vector; λ is the filter parameter matrix; is the total disturbance F u estimated value.

[0029] The above technical solution is preferred, in formula 2), L and λ are as follows:

[0030]

[0031]

[0032] Among them, l1 and l2 are both third-order diagonal matrices.

[0033] The above technical solution is preferred, 1i 、l2i and λ i (Satisfies the following characteristic equation:

[0034] s 3 +l 1i s 2 +(λ i l 1i +l 2i )s+λ i l 2i =0;

[0035] S represents a complex variable; l 1i =9w i ; w i represents the bandwidth of observer channel i; subscript i = x, z, γ.

[0036] In the above technical solution, preferably, in step S3, obtaining the control signal includes:

[0037] Step S3.1: Based on the barrier Lyapunov function, combined with the expected position q d , the state vector q and the estimated value of q Get the virtual control quantity ε of the inner loop;

[0038] Step S3.2: Based on the barrier Lyapunov function, the virtual control variable ε and the state error vector e and the estimated value of the total disturbance obtained in step S2 are combined The control signal U is obtained by design.

[0039] The above technical solution is preferably, wherein step S3.1 includes:

[0040] Step S3.11, select the first obstacle Lyapunov function V1, and take the derivative of V1 to obtain the first-order derivative of the first obstacle Lyapunov function As shown in the following formula 3):

[0041]

[0042] Step S3.12, based on The virtual control quantity ε is designed as shown in Equation 4):

[0043]

[0044] Step S3.2 includes:

[0045] Step S3.21: Select the second obstacle Lyapunov function V2 and take the derivative of V2 to obtain the first-order derivative of the second obstacle Lyapunov function As shown in formula 5):

[0046]

[0047] Step S3.22, based on The preselected control signal U' is designed and the estimated value of the total disturbance obtained in step S2 is used. Replace the total perturbation F in U' u , and the final control signal U is obtained, as shown in the following equations 6) and 7):

[0048]

[0049]

[0050] Among them, e1 is the outer loop error vector; e2 is the inner loop error vector; κ q represents the constraint of e1; τ1 and τ2 both represent coefficients.

[0051] The above technical solution is preferred, κ q =[κ x ,κ z ,κ γ ] T ; Among them, κ x , κ z and κ γ They represent the fixed limit value of position x tracking error, the fixed limit value of position z tracking error and the fixed limit value of swing angle error, respectively, satisfying |e x |<κ x ,|e z |<κ z ,|e γ |<κ γ ;e x 、e z and e γ They represent the x tracking error of the UAV coordinate position, the z tracking error of the coordinate position, and the swing angle error of the hanging load, respectively.

[0052] A UAV suspended flight sway reduction control system, used to implement the UAV suspended flight sway reduction control method, comprising a tracker, an observer and a backstepping controller;

[0053] The tracker and observer are both connected to the backstepping controller. The tracker is used to obtain the transition signal of the input desired position; the observer is connected to the UAV's flight system to estimate the UAV's total disturbance; and the backstepping controller is connected to the UAV's flight system to output a control signal to control the UAV's stable operation.

[0054] The application of the technical solution of the present invention has the following beneficial effects:

[0055] (1) The method for controlling the suspended flight of a UAV of the present invention uses a control signal U and an observer to observe the suspended flight system of the UAV, and adds an estimated value obtained by the observation when generating a new control signal to offset the adverse effects of the disturbance on the flight control of the system. The control signal U output by the control method of the present invention enables the UAV to achieve high-precision position tracking and suppress the swing of the suspended load, thereby ensuring that the UAV position tracking error and the load swing angle change within a constraint range.

[0056] (2) In the present invention, the established dynamic model integrates the position and load swing angle information of the UAV flight system into a column vector, and the dynamic model of the system is designed with the column vector as the state variable. The system position and load swing angle information are regarded as an overall variable, and a controller is designed for the overall variable (i.e., a control input signal is designed) to achieve the same frequency control of the UAV position and the hanging load swing, which simplifies the design of the controller and optimizes the control effect.

[0057] (4) In the present invention, a disturbance observer (i.e., compensation observer) is designed to address the total disturbance of the UAV suspension flight system (such as unmodeled dynamics, air turbulence, and parameter perturbations). The observer adopts a pure integral structure and makes full use of the system state variables and their differential information. By subtracting the system state variables and differential information estimated by the observer, the state variable error and the error of the state variable differential are obtained. These errors are then integrated to compensate for the total disturbance, thereby eliminating the adverse effects of the total disturbance on the system's swing reduction control. Zero error convergence can be achieved for total disturbances in the form of constants, velocity functions, and parabolic functions. The turbulence encountered during the flight of the UAV can be estimated with high precision, and the estimation accuracy of the total disturbance and the convergence of the tracking error are better than other observers or observation methods.

[0058] (5) In the present invention, in view of the requirement that the position tracking error and the swing angle in the UAV hanging flight mission need to change within a fixed constraint range, a backstepping controller based on BLF (Barrier Lyapunov Function) is designed to perform fixed and restricted control on the UAV position error and the load swing angle, and a stability proof is given to ensure that the UAV position tracking error and the load swing angle change within a fixed constraint range, meeting the actual mission requirements; in view of the disadvantage that the backstepping control relies on an accurate mathematical model, the high-precision estimate of the total disturbance by the compensation observer is fed back to the backstepping controller to offset its adverse effects on the hanging flight swing reduction control, so as to achieve the purpose of coping with the internal nonlinear uncertainty and external disturbance of the system, thereby realizing the precise control of the UAV position and the swing reduction control of the hanging load.

[0059] (6) The UAV suspension flight anti-sway control system of the present invention is used to implement the above-mentioned control method. By outputting control signals to the UAV flight system through the backstepping controller, it can achieve precise control of the UAV position and anti-sway control of the suspension load.

[0060] In addition to the above-described objects, features and advantages, the present invention has other objects, features and advantages. The present invention will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The drawings constituting a part of this application are used to provide a further understanding of the present invention. The illustrative embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0062] In the attached figure:

[0063] Figure 1 is a schematic diagram of the UAV in this embodiment (illustrating the hanging load);

[0064] Figure 2 This is a simplified structural diagram of the UAV hanging flight system in this embodiment;

[0065] Figure 3 This is a schematic diagram of the principle of the UAV suspension flight anti-sway control system in this embodiment;

[0066] Figure 4 It is a schematic diagram of wind disturbance;

[0067] Figure 5 This is a comparison graph of the UAV position x changing with time under the three control methods;

[0068] Figure 6 This is a comparison graph of the change of the UAV position z over time under the three control methods;

[0069] Figure 7 This is a comparison curve of the load swing angle γ changing with time under the three control methods;

[0070] Among them, 1. Drone; 2. Payload; 3. String. DETAILED DESCRIPTION

[0071] The embodiments of the present invention are described in detail below with reference to the accompanying drawings. However, the present invention can be implemented in many different ways as defined and covered by the claims.

[0072] Example:

[0073] This embodiment discloses a method and control system for controlling the sway reduction during suspended flight of a UAV. The control system in this embodiment will be described first.

[0074] like Figure 3 As shown, the UAV suspension flight anti-sway control system in this embodiment includes a tracker (specifically a third-order differential tracker), an observer (specifically a compensation function observer), and a backstepping controller (specifically a backstepping controller based on the obstacle Lyapunov function);

[0075] The tracker is connected to the backstepping controller, and the tracker is used to arrange the transition process and obtain the tracking signal v1, the first-order differential signal v2 and the second-order differential signal v3 of a given signal (i.e., the desired position, i.e., the target position that the drone wants to go to);

[0076] The observer is connected to the backstepping controller and the flight system of the UAV respectively. The observer can use the control signal U output by the backstepping controller and the output of the UAV suspension flight system (The system output is measured by sensors in the UAV flight system) to estimate the model bias, internal disturbance and external disturbance (here the model bias, internal disturbance and external disturbance are collectively referred to as total disturbance), and the estimated value of the total disturbance is Feedback is given to the backstepping controller to offset the adverse effects of disturbances on the flight control of the system.

[0077] The backstepping controller is connected to the flight system of the UAV. The backstepping controller is used to generate the lift vector required for the UAV to fly to the target position, that is, the control signal U. The control signal is input to the UAV flight system to enable the UAV to fly to the target position and ensure that the hanging load swing angle is within the set range. Specifically, the backstepping controller is based on the estimated value of the observer. The output of the UAV's hanging flight system and the tracking signal, first-order differential signal and second-order differential signal of the given signal generate the control signal U, which enables the UAV to achieve high-precision position tracking and suppress the swing of the hanging load, ensuring that the UAV position tracking error and the load swing angle change within the constraint range.

[0078] Among them, the drone and the hanging load in this embodiment are as follows Figure 1 As shown;

[0079] Considering that when the UAV is flying in a suspended transport mode, in most cases the suspended load is at the o b x b z b The UAV suspension flight system is simplified into a two-dimensional plane model, such as Figure 2 shown.

[0080] Figure 2 In, o i x i z i is the inertial coordinate system, o b x b z bis the coordinate system of the UAV, f is the total lift generated by the four rotors of the UAV, and the masses of the UAV and the payload are m respectively. q and m l , the length of the string (used to hang the load) is l, the swing angle between the load and the vertical direction is γ, and the pitch angle of the drone is θ.

[0081] The method for controlling the sway reduction of a UAV in suspended flight disclosed in this embodiment includes steps S1 to S4, specifically as follows:

[0082] Step S1: Establish the dynamic model of the UAV hanging flight system:

[0083] Among them, since the UAV hanging flight system adds two degrees of freedom brought by the hanging load, but the control input does not increase, the traditional vector mechanics method needs to consider the constraint force of the hanging tether (i.e., the thin rope) when modeling, and the modeling process of the UAV hanging system is cumbersome. In this embodiment, the Lagrangian mechanics method is used to establish the dynamic model of the UAV hanging flight system. The number of equations is the same as the degree of freedom of the system. The form is simple and the structure is compact. Only the known active force needs to be analyzed, and the unknown constraint force on the tether does not need to be considered, which can achieve the effect of simplifying the analysis, as shown below:

[0084] The dynamic model of the UAV suspension flight system can be established based on the Lagrange equation, as shown in Equation 1.a):

[0085]

[0086] Where q = [x q ,z q ,γ] T , represents the state vector of the UAV suspension flight system; It should be noted that: in this embodiment, the superscript dot represents the derivative, for example, and They represent the first-order derivative and the second-order derivative of q respectively. This representation method is well known in mathematics and will not be described in detail below.

[0087] M c (q) represents the inertia matrix of the UAV suspension flight system, M c (q) is shown in Equation 1.11):

[0088]

[0089] Represents the centripetal force matrix of the UAV suspension flight system, As shown in formula 1.12:

[0090]

[0091] G(q) represents the gravity vector of the UAV suspension flight system, and G(q) is shown in Equation 1.13):

[0092] G(q)=[0 (m q +m l )gm l glsinγ] T 1.13);

[0093] Q represents the generalized force vector of the UAV suspension flight system:

[0094]

[0095] Among them, U=[u x u z 0] T is the control signal output by the backstepping controller, which is also the control input of the UAV suspension flight system. x and u z They represent the components of the total lift f generated by the UAV along the x-direction and the z-direction in the inertial coordinate system respectively;

[0096] F d Represents the air resistance vector of the UAV suspension flight system, F d As shown in the following formula 1.14):

[0097]

[0098] Among them, d x , d z , d γ are air damping coefficients; and They represent the speed of the UAV in the x and z directions in the inertial coordinate system respectively;

[0099] From the above, Equation 1.a) can be expressed as Equation 1), which is the dynamic model of the UAV hanging flight system:

[0100]

[0101] Among them, M c0 M c (q) -1 Nominal value of F u is the total disturbance received by the UAV, F u As shown in the following formula 1.1):

[0102]

[0103] Step S2: Design an observer based on the dynamic model of step S1 to observe the state of the UAV suspension flight system, and combine it with the control signal output by the backstepping controller to obtain the estimated value of the total disturbance as follows:

[0104] First, the principle is explained: Due to the total disturbance F in the above formula 1) u Contains the state vector q, the first-order derivative and the functions they constitute, and in addition include the air resistance vector F d As well as the dynamics that are not modeled, these are unknown quantities that make the design of the backstepping controller difficult. Therefore, it is necessary to use an observer to calculate F u Estimate it.

[0105] Secondly, based on the established dynamic model of the UAV hanging flight system (i.e., Formula 1), the state equation of the UAV hanging flight system can be obtained, which is as follows:

[0106]

[0107] Where x1 = q, Here, the vector y composed of x1 and x2 is the output vector of the drone;

[0108] Furthermore, in this embodiment, a compensation function observer is used to observe the total disturbance. The expression of the compensation function observer is shown in Formula 2):

[0109]

[0110] in, represents the estimated value of q; Express The estimated value of; L = [l1, l2] is the gain matrix; is the state error vector; is the compensation vector; λ is the filter parameter matrix (third-order diagonal matrix); is the total disturbance F u estimated value.

[0111] In Equation 2), L = [l1, l2] and λ is as follows, where l1, l2, and λ are all third-order diagonal matrices:

[0112]

[0113] Among them, l 1i 、l 2i and λ i Satisfies the following characteristic equation:

[0114] s 3 +l1i s 2 +(λ i l 1i +l 2i )s+λ i l 2i =0

[0115] S represents a complex variable. To make the observer asymptotically stable, the pole of the observer is configured to be S1 = -w i ,S2=S3=-4w i , we can get l 1i =9w i ; w i represents the bandwidth of observer channel i; subscript i = x, z, γ.

[0116] The observer designed in this way can timely estimate the dynamics of the UAV suspension flight system that have not yet been modeled, internal disturbances of the system (such as changes in the mass of the UAV and the payload), and external disturbances (such as turbulent wind interference encountered during the flight of the UAV).

[0117] It should be noted that: since there is no control signal at the initial moment, the observer calculates the estimated value at the initial moment When the initial control signal U is 0, the initial estimated value is obtained by calculating the initial control signal After the control signal U at the next moment is calculated using the initial estimated value, the newly calculated control signal U needs to be used when calculating the estimated value at the next moment (ie, the control signal U is not 0 at this time).

[0118] Step S3: Based on the desired position, the output of the UAV's suspension flight system, and the estimated value of the total disturbance, a control signal is designed and transmitted to the dynamic model of step S1 to update the state of the UAV (updating the UAV state with a control signal is common knowledge in the art and will not be described in detail in this embodiment) to maintain the flight stability of the UAV, as follows:

[0119] First, the principle is explained: in the UAV hanging flight mission, the UAV's position tracking error and load swing angle generally need to be constrained. For example, in the UAV mine detection mission, the UAV's position and the swing angle of the hanging detector need to change within a fixed range. Therefore, considering the needs of high-precision position tracking and hanging load swing suppression during the actual flight of the UAV hanging flight system, and to ensure that the UAV position and load swing angle change within the constraint range, it is necessary to design a controller (that is, design a control signal) so that the UAV position tracking error and swing angle error always change within a fixed range.

[0120] This step S3 includes the following steps:

[0121] Step S3.1: Based on the barrier Lyapunov function (BLF), combined with the expected position q d The difference between the state vector q and the estimated value of q is used to obtain the virtual control quantity ε of the inner loop, as follows:

[0122] Step S3.11:

[0123] Let e1 = x1 - q d =[e x ,e z ,e γ ] T is the outer loop error vector; e2 = x2-ε, e2 is the inner loop error vector, ε is the virtual control quantity of the inner loop (ε is generated by the outer loop); where q d =[x qd ,z qd ,0] T , x qd and z qd They represent the expected positions of the UAV along the x and z directions in the inertial coordinate system, e x 、e z and e γ They represent the x tracking error of the UAV coordinate position, the z tracking error of the coordinate position, and the γ error of the hanging load swing angle respectively;

[0124] For the outer loop of the system, the first obstacle Lyapunov function V1 is selected, as shown in Equation 3.1):

[0125]

[0126] Among them, κ q =[κ x ,κ z ,κ γ ] T is the constraint of e1, κ x , κ z and κ γ They represent the fixed limit value of position x tracking error, the fixed limit value of position z tracking error and the fixed limit value of swing angle error, respectively, satisfying |e x |<κ x ,|e z |<κ z ,|e γ |<κ γ ;

[0127] Derivative V1 to obtain the first-order derivative of the first obstacle Lyapunov function As shown in the following formula 3):

[0128]

[0129] Step S3.12, based on the first-order derivative of the first obstacle Lyapunov function, the virtual control quantity ε is designed. Specifically, in order to ensure the existence of the negative square term in Equation 3 and eliminate Then the virtual control quantity ε is designed as shown in formula 4):

[0130]

[0131] The virtual control variable ε is obtained through the above steps.

[0132] Before proceeding to step S3.2, it should be noted that τ1 and τ2 (τ2 is described below) are both coefficients and are positive numbers. τ1 is much larger than τ2. For example, τ1 = 500 and τ2 = 0.01. The specific magnitude relationship between τ1 and τ2 is selected based on the actual situation. Substituting Equation 4) into Equation 3) yields Equation 3.2).

[0133]

[0134] The second term in Equation 3.2 (i.e. ) affected The negative determination can be offset when designing the inner loop control law (i.e., eliminating the item through step S3.2), so that the outer loop is stabilized under the action of the virtual control variable ε, thereby achieving the reachability of the outer loop.

[0135] Step S3.2: Based on the barrier Lyapunov function, the virtual control variable ε is combined with the state error vector e obtained in step S2 and the estimated value of the total disturbance The control signal U is designed as follows:

[0136] Step S3.21: To ensure that the drone's position error and the hanging load's swing angle converge to 0 during the flight to the target position in step S3.1, the second obstacle Lyapunov function V2 is selected for the inner loop. V2 is shown in Equation 5.1 below:

[0137]

[0138] Derivative the second obstacle Lyapunov function to obtain the first-order derivative of the second obstacle Lyapunov function As shown in formula 5.2):

[0139]

[0140] Furthermore, according to formula 5.2) and the above formula 3.2), formula 5.2) can be expressed as formula 5):

[0141]

[0142] Step S3.22: First-order derivative of the Lyapunov function based on the second obstacle The design obtains the preselected control signal U', specifically: to ensure that the error vectors e1 and e2 converge asymptotically to zero vectors and eliminate Therefore, the preselected control signal U' is designed as shown in formula 6):

[0143]

[0144] Among them, τ2 is a coefficient (τ2 has been explained above), so The system can achieve asymptotic stability, so the inner loop control law of the system can stabilize the UAV hanging flight system, and the outer loop control law can enable the system to track the desired position and swing angle;

[0145] Then use the estimated value of the total disturbance obtained in step S2 Replace the total perturbation F in U' u , the final control signal U is obtained, that is, the expression of the backstepping controller finally designed is shown in Equation 7):

[0146]

[0147] Finally, the backstepping controller outputs the generated control signal U to the dynamic model of step S1 to update the state of the drone, and at the same time outputs the control signal U to the observer to calculate the new estimated value

[0148] In this embodiment, the BLF (Barrier Lyapunov Function) and the backstepping controller are combined to design the control signal for the UAV's suspended flight swing reduction. On the one hand, it can make the system state variable tracking error vary within a fixed range, and on the other hand, it can make the UAV achieve internal stability and external performance during the suspended flight.

[0149] Step S4: After the drone is controlled by the control signal U, the output of the drone hanging flight system after the state update is returned to step S3 to participate in the calculation of the new control signal. At the same time, the output of the drone hanging flight system is returned to step S2 to participate in the new estimated value. Calculation.

[0150] The following is an explanation of the principle of how the control method in this embodiment suppresses the swing of the suspended load:

[0151] From Equation 3), we can see that when BLF is bounded under the action of the designed controller, the function variable will always remain within the constraint boundary, that is, |e x |<κ x ,|e z |<κ z ,|eγ |<κ γ , And from x1=q d +e1, we can see that x q (t) = x qd (t)+e x (t), z q (t) = z qd (t)+e z (t), γ(t)=e γ (t), t represents time, indicating the state quantity x q (t), z q (t), γ(t) are both limited. In addition, when τ1>0 and τ2>0, V2>0, This shows that the system is asymptotically stable, so V2→0. From Equation 5), we know that V1→0, e2→0. From Equation 3), we know that e1→0, so x q (t)→x qd (t), z q (t)→z qd (t), γ(t)→0, thus achieving the purpose of suppressing the swing during the UAV hanging transportation process.

[0152] The above is a description of the control system and control method of this embodiment.

[0153] The following is a test conducted in this embodiment to verify that the control method in this embodiment is superior to other existing control methods:

[0154] First: Set the parameters of the quadcopter UAV hanging system to: m q =1.0082kg,m l =0.076kg, l=1.085m, g=9.81m / s 2 . Set the initial position of the hanging flight system to: x q0 =0m、z q0 =1.5m, target position is: x qd =-1.5m, z qd =3m.

[0155] In [10, 30]s, Figure 4 The x-channel and z-channel of the wind disturbance injection system are shown in Figure 2, and compared with the linear active disturbance rejection control (LADRC) and linear quadratic regulator (LQR) control methods, as shown in Figure 2. Figures 5 to 7 shown.

[0156] The relevant parameters of the observer (CFO) in this embodiment are: x =6,w z =6,w γ=6, the parameters of the extended state observer (ESO) in the LADRC used for comparison are: ox =80,ω oz =80,ω oγ =500,ω ox 、ω oz and ω oγ are the bandwidths of the ESO at position x, position z, and swing angle γ channels, respectively.

[0157] The parameters of the BLF-based backstepping controller in this embodiment are: τ1 = 100, τ2 = 0.01, κ q =[0.2,0.2,0.1] T , the relevant parameters of the ADRC controller used for comparison are: ω cx =1.5,ω cz =1.5,ω cγ =0.2,ω cx 、ω cz and ω cγ are the bandwidths of the existing linear active disturbance rejection controller in the position x, position z and swing angle γ channels respectively.

[0158] The comparison is as follows:

[0159] like Figures 5 to 7 As shown, Figures 5 to 7 The changes of the UAV position and the load swing angle over time under the three control methods of CFO-BLF (i.e., the control method of this embodiment), LADRC, and LQR are described respectively.

[0160] The tracking error of the LQR method fluctuates greatly, especially after being disturbed by wind, and is always in an oscillating state. During the disturbance, the standard deviations of the position x, z and the swing angle γ are 0.5922, 0.6025, and 5.286, respectively.

[0161] The LADRC method will produce small fluctuations after being disturbed by wind, and the standard deviations of the position x, z and swing angle γ are 0.01347, 0.01254, and 0.352, respectively.

[0162] Compared with the previous two methods, the CFO-BLF method has almost no fluctuation during the disturbance and is more robust. The standard deviations of the positions x, z and the swing angle γ are 0.0003283, 0.0004008, and 0.01729, respectively.

[0163] Both the CFO-BLF and LADRC methods use observers to estimate the unmodeled dynamics of the system, approximating the system to an integral series type and ensuring its asymptotic stability. When the aircraft is subject to wind disturbances, the LQR's anti-interference performance is poor, with significant oscillations occurring in both the position and angle channels. This is because the LQR fails to compensate for the disturbances. After the disturbance disappears, the LQR requires 4.5 seconds to settle to eliminate the steady-state error.

[0164] The CFO-BLF backstepping controller in this embodiment has almost no fluctuation after being subjected to wind disturbance, and its robustness is significantly better than that of LADRC. This is because the CFO-BLF backstepping controller uses CFO to estimate the disturbance. CFO is a Type III system with high disturbance estimation accuracy. As a result, the BLF controller can accurately compensate for the adverse effects of wind disturbances. In addition, because the CFO-BLF method limits the error, the error always remains within the constraint range. From the previous stability analysis, it can be seen that BLF has a strong error constraint effect, so that the steady-state error eventually converges to near 0, ensuring high steady-state accuracy.

[0165] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A method for reducing sway during suspended flight of an unmanned aerial vehicle, characterized in that: The steps include: Step S1: Establish a dynamic model of the UAV hanging flight system; In step S1, the dynamic model of the UAV suspension flight system is shown in formula 1): Among them, q represents the state vector of the UAV suspension flight system; F u represents the total disturbance to the UAV; M c0 Indicates M c (q) -1 Nominal value, M c (q) represents the inertia matrix of the UAV suspension flight system; U represents the control signal; Step S2: Based on the dynamic model, the total disturbance of the UAV is observed by an observer and combined with the control signal to obtain an estimated value of the total disturbance; In step S2, the total disturbance of the UAV is observed by an observer, as shown in Equation 2): in, represents the estimated value of q; Express The estimated value of ; L is the gain matrix; is the state error vector; is the compensation vector; λ is the filter parameter matrix; is the total disturbance F u estimated value of; In formula 2), L and λ are as follows: L＝[l1,l2]； Among them, l1 and l2 are both third-order diagonal matrices, l 1i =9w i ; w i represents the bandwidth of observer i channel; subscript i = x, z, γ; Step S3: Based on the desired position, the output of the UAV's suspension flight system, and the estimated value in step S2, a control signal is generated. The control signal is transmitted to the dynamic model in step S1 to update the state of the UAV, and the control signal is returned to step S2 to calculate a new estimated value. In step S3, obtaining the control signal includes: Step S3.1: Based on the barrier Lyapunov function, combined with the expected position q d , the state vector q and the estimated value of q Get the virtual control quantity ε of the inner loop; Step S3.2: Based on the barrier Lyapunov function, the virtual control variable ε and the state error vector e and the estimated value of the total disturbance obtained in step S2 are combined The design obtains the control signal U; Step S4: Return the updated output of the UAV suspension flight system to steps S2 and S3 respectively to calculate a new estimated value and a new control signal.

2. The method for reducing sway during suspended flight of a UAV according to claim 1, characterized in that: In formula 1), F u As shown in formula 1.1): in, represents the centripetal force matrix of the UAV hanging flight system; G(q) represents the gravity vector of the UAV hanging flight system; F d represents the air resistance vector; M c (q), G(q) and F d As shown below: G(q)=[0(m q +m l )g m l glsinγ] T 1.13); Among them, m q Indicates the mass of the drone; m l represents the mass of the load; l represents the length of the string used to suspend the load; γ represents the swing angle between the load and the vertical direction; d x d z and d γ Both represent the air damping coefficient; and They represent the speed of the UAV in the x and z directions in the inertial coordinate system respectively.

3. The method for reducing sway during suspended flight of a UAV according to claim 2, characterized in that: l 1i 、l 2i and λ i Satisfies the following characteristic equation: s 3 +l 1i s 2 +(λ i l 1i +l 2i )s+λ i l 2i =0; s represents a complex variable.

4. The method for controlling the sway reduction during suspended flight of a UAV according to claim 3, wherein: The step S3.1 includes: Step S3.11, select the first obstacle Lyapunov function V1, and take the derivative of V1 to obtain the first-order derivative of the first obstacle Lyapunov function As shown in the following formula 3): Step S3.12, based on The virtual control quantity ε is designed as shown in Equation 4): Step S3.2 includes: Step S3.21: Select the second obstacle Lyapunov function V2 and take the derivative of V2 to obtain the first-order derivative of the second obstacle Lyapunov function As shown in formula 5): Step S3.22, based on The preselected control signal U' is designed and the estimated value of the total disturbance obtained in step S2 is used. Replace the total perturbation F in U' u , and the final control signal U is obtained, as shown in the following equations 6) and 7): Among them, e1 is the outer loop error vector; e2 is the inner loop error vector; κ q represents the constraint of e1; τ1 and τ2 both represent coefficients.

5. The method for controlling the sway reduction during suspended flight of a UAV according to claim 4, characterized in that: κ q =[κ x ,κ z ,κ γ ] T ; Among them, κ x , κ z and κ γ They represent the fixed limit value of position x tracking error, the fixed limit value of position z tracking error and the fixed limit value of swing angle error, respectively, satisfying |e x |<κ x ,|e z |<κ z ,|e γ |<κ γ ;e x 、e z and e γ They represent the x tracking error of the UAV coordinate position, the z tracking error of the coordinate position, and the swing angle error of the hanging load, respectively.

6. A UAV suspended flight anti-sway control system, used to implement the UAV suspended flight anti-sway control method according to any one of claims 1 to 5, characterized in that: Includes tracker, observer and backstepping controller; The tracker and observer are both connected to the backstepping controller, the tracker is used to obtain the input desired position; the observer is connected to the flight system of the UAV to estimate the total disturbance of the UAV; The backstepping controller is connected to the flight system of the UAV and is used to output control signals to control the UAV to operate stably.

Images