A Quadrotor Suspended Transport Control Method Based on Dual Time-Varying Disturbance Estimator

Abstract

The application discloses a kind of four-rotor hanging transport control methods based on double time-varying interference estimator, comprising the following steps: S1, the dynamics model of four-rotor unmanned aerial vehicle and hanging load is established;S2, the model is simplified using feedback linearization technique, and the nonlinear coupling system is decoupled into two subsystems of anti-swing and trajectory tracking;S3, according to the swing model of the goods being hung, time-varying uncertainty and interference estimator are constructed in the anti-swing controller, and time-varying interference is dynamically aggregated to eliminate the swing of heavy object;S4, according to the kinematics model of four-rotor, time-varying uncertainty and interference estimator are constructed in the trajectory tracking controller, to realize the accurate trajectory tracking of four-rotor unmanned aerial vehicle;S5, the stability of estimation error system is proved using linear time-varying system theory, and the interference estimation performance is optimized by adjusting auxiliary parameter;S6, the stability of anti-swing subsystem and trajectory tracking subsystem is proved, and the interference estimation performance is optimized by adjusting auxiliary parameter.

Description

Technical Field

[0001] This invention belongs to the field of aircraft automatic control technology, specifically relating to a quadcopter sling transport control method based on a dual time-varying disturbance estimator. Background Technology

[0002] In recent years, quadcopter drones have been widely used in various fields due to their low cost, high maneuverability, and excellent hovering capabilities, successfully completing aerial tasks such as aerial photography, formation performances, and remote sensing. As an emerging mode of transportation, quadcopter drones are also being used in the field of air transport. For example, in the event of natural disasters such as earthquakes and mudslides, quadcopter drones, because they can take off and land vertically regardless of terrain and hover, are being tasked with delivering emergency relief supplies such as water, food, and medicine to trapped individuals, enabling them to save themselves, especially in areas without suitable landing areas, such as mountains, jungles, and islands.

[0003] Quadrotor drones face complex interference from gusts, air resistance, and waypoint changes. These time-varying disturbances often exacerbate load sway, which, if not eliminated, can severely affect drone stability or even lead to loss of control. Similarly, these time-varying disturbances and the resulting load sway pose a significant challenge to achieving precise trajectory tracking, especially when the external environment imposes strict constraints on the drone's flight path. Without time-varying disturbance-resistant control for trajectory tracking, collisions and crashes may occur, causing significant danger and losses. Therefore, simultaneous anti-sway and interference-resistant control during transport and trajectory tracking is of paramount importance. Summary of the Invention

[0004] The purpose of this invention is to solve the above-mentioned problems and provide a control framework based on dual time-varying uncertainty and disturbance estimator (TVUDE) to solve the problem of time-varying multiple disturbances encountered by quadrotor UAVs carrying loads in different flight stages, and to achieve efficient load transportation; and to coordinate anti-sway control and trajectory tracking control, and simultaneously realize the quadrotor carrying transport control method of both.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is: a quadcopter sling transport control method based on a dual time-varying disturbance estimator, comprising the following steps:

[0006] S1. Establish a dynamic model of the quadcopter UAV and its suspended load, including the transformation relationships between the world coordinate system, the body coordinate system, the suspension coordinate system, and the load coordinate system;

[0007] S2. The model is simplified using feedback linearization techniques, and the nonlinearly coupled system is decoupled into two subsystems: anti-sway and trajectory tracking.

[0008] S3. Based on the swaying model of the hoisted goods, a time-varying uncertainty and disturbance estimator (TVUDE) is constructed in the sway elimination controller to dynamically aggregate time-varying disturbances to eliminate the swaying of the heavy objects;

[0009] S4. Based on the quadrotor kinematic model, a time-varying uncertainty and disturbance estimator (TVUDE) is constructed in the trajectory tracking controller to achieve accurate trajectory tracking of the quadrotor UAV.

[0010] S5. Prove the stability of the estimation error system using the theory of linear time-varying systems, and optimize the performance of disturbance estimation by adjusting auxiliary parameters;

[0011] S6. Prove the stability of the anti-swing subsystem and the trajectory tracking subsystem, and optimize the disturbance estimation performance by adjusting the auxiliary parameters.

[0012] Furthermore, the rotation matrix of the quadcopter in S1 is given by the following formula:

[0013]

[0014] Where C(·) and S(·) represent the trigonometric functions sin(·) and cos(·) respectively; θ, φ, ψ represent the Euler angles of the quadcopter in the inertial frame; B refers to the machine system, W refers to the world system, P refers to the load system, and S refers to the suspension system. It is the rotation matrix from the slave system to the world system; other rotation matrices are defined in the same way.

[0015] Furthermore, the dynamic model of the suspended load in S1 includes a model of cargo swaying, which is expressed as follows:

[0016]

[0017] Where λ x ,λ y The two swing angles of the cargo are represented by L; L is the length of the rope used to hoist the cargo; m p It refers to the quality of the goods; This represents the acceleration vectors along the three axes of a quadcopter drone; the symbol "T" in the upper right corner represents the transpose of the vector or matrix; F pd =[F pdx ,F pdy ,F pdz ] T is the air resistance vector acting on the cargo; g is the acceleration due to gravity.

[0018] Furthermore, based on the aforementioned cargo swaying model, a comprehensive quadrotor load model is established, and the quadrotor linear motion dynamics formula in the world coordinate system is given by the following equation:

[0019]

[0020] Where m q It is the mass of the quadcopter; F qt =[0,0,F qt ] T It is the quadrotor thrust command vector in the body coordinate system; F is the cable tension vector on the quadcopter in the UAV's world coordinate system; qd This is the disturbance signal vector of the quadcopter position loop. The disturbance signals include model uncertainty, wind interference, and load oscillation interference. The derivatives of the disturbances are all bounded, and g = [0, 0, g]. T It is gravitational acceleration.

[0021] Furthermore, in step S2, the model is simplified using feedback linearization technology, and the simplified swing angle model is as follows:

[0022]

[0023] Where λ=[λ x ,λ y ] T It's the tilt angle of the quadcopter, u λ =[u λx ,u λy ] T It is a virtual control input, and has:

[0024]

[0025] f λ =[f λx ,f λy ] T It is a lumped interference, and has:

[0026]

[0027] To prevent the quadcopter from producing aggressive vertical maneuvers that could lead to collisions, the system input is... To sum up to the interference term f λ middle.

[0028] Furthermore, the quadrotor load synthesis model (3), after using feedback linearization technology, yields a simplified model as follows:

[0029]

[0030] Where u p It is virtual input:

[0031]

[0032] f p It is lumped interference:

[0033]

[0034] This completes the simplification of the model, facilitating subsequent control design.

[0035] Furthermore, in S3, the swing angle converges rapidly during the swing and actively resists interference, which is the uncertainty of the model and external interference. An outer loop and an inner loop are set up. The outer loop is the control loop and the inner loop is the attitude loop. The first component of the outer loop is designed as the swing cancellation controller, which rewrites the simplified swing angle model (4) into a state space expression:

[0036]

[0037] Where x1 and x2 represent the yaw angle and yaw angular velocity of the quadcopter, respectively, assuming that the required state is generated by the following reference model:

[0038]

[0039] The robust control algorithm based on uncertainty and disturbance estimator designed for the model of formula (10) is expressed as follows:

[0040]

[0041] u λ0 It is a nominal controller, designed as follows:

[0042]

[0043] Where u λd It is a feedforward term, which can be considered zero in this robust stabilization problem, kp = diag{kpx,kpy} and k d =diag{k dx ,k dy} is the feedback gain. It is the quadcopter tilt angle tracking error. It's an angular velocity tracking error, and in this problem, the swing angle x... 1d and x 2d The expected value can also be considered as zero; substituting formula (13) into formula (12), we get:

[0044]

[0045] The interference estimation signal generated by TVUDE is designed in the time domain as follows:

[0046]

[0047] Where T λ(t)=diag{T λx (t),T λy (t)} is a time-varying parameter matrix designed by humans. Combining formula (4) and formula (12), we get

[0048] Furthermore, the trajectory tracking controller of TVUDE in S4 is the second component of the outer loop. The simplified trajectory tracking model is represented by formula (7), and its state-space expression is:

[0049]

[0050] in and These represent the position and speed of the quadcopter, respectively.

[0051] Furthermore, in S5, the stability of the estimation error system is proven using linear time-varying system theory, and when optimizing the interference estimation performance by adjusting auxiliary parameters, the anti-sway interference estimator is taken into account. And trajectory tracking interference estimator The system maintains formal consistency, and in the proof, it is only necessary to prove that one of the estimation error systems is stable. Therefore, all subscripts in the system will be removed in the subsequent proof. The estimation error is defined as:

[0052]

[0053] Integrating both sides of equation (25) with respect to time t, and then substituting it into equation (15), we obtain the estimation error subsystem:

[0054]

[0055] According to the theory of linear systems, the state transition matrix of the linear time-varying system (26) is:

[0056]

[0057] The solution to system (26) is:

[0058]

[0059] Time-varying parameter T = diag{T x (t),T y (t),T z (t)} is a bounded positive number that is artificially defined. For the x-axis, we define 0 < T. xmin ≤T x (t)≤T xmax And there are From formula (28), we know that the solution of the system consists of two components: the zero-input response and the zero-state response. As t→∞, the zero-input response... Then, scaling the zero-state response yields:

[0060]

[0061] in As t→∞, The zero-state response can be simplified to:

[0062]

[0063] It can be proven using the same method. and It is also convergent; as t→∞, we have:

[0064]

[0065] in T m =diag{T xmax ,T ymax ,T zmax};make The advantage of doing this is that the time-varying parameter matrix T can be changed by adjusting the magnitude of the scalar ζ. m The magnitudes of each component, thus proving that when the external disturbance f is bounded, the output... It is also bounded, which proves that system (26) is stable.

[0066] Furthermore, in S6, when proving the stability of the anti-swing subsystem and the trajectory tracking subsystem and optimizing the disturbance estimation performance by adjusting the auxiliary parameters, it was also observed that the anti-swing subsystem (10) and the trajectory tracking subsystem (19) have the same state-space expression, so only one of them is proven; the swing angle error and angular velocity error are defined as:

[0067]

[0068] Combining formulas (10), (11), (12), (13), and (25), we obtain the oscillation error system:

[0069]

[0070] To minimize the swing angle error and angular velocity error For a system to converge to boundedness, system (33) must satisfy:

[0071] Matrix A is the Herwitz matrix, i.e., the gain k p >0,kd >0; estimation error Bounded convergence; the solution for the oscillation error system is:

[0072]

[0073] As t→∞, the zero-input response e At →0, Substituting formula (31) into the zero-state response, we get:

[0074]

[0075]

[0076] Obviously, as t→∞, (||I||2-||e At ||2)→||I||2, and when ζ→0, ε(ζ)→0. Therefore, it is proved that the oscillation error system satisfies: Thus, it can be proven that the anti-slip system and the trajectory tracking system are exponentially stable, and the performance of perturbation estimation can be improved by reducing the auxiliary parameter ζ.

[0077] The beneficial effects of this invention are:

[0078] 1. This invention provides a quadcopter sling transport control method based on a dual time-varying disturbance estimator (TVUDE). By introducing a dual time-varying uncertainty and disturbance estimator (TVUDE), this invention significantly improves the anti-interference capability of quadcopter UAVs in the face of time-varying multi-interference environments. TVUDE can dynamically estimate and compensate for time-varying disturbances caused by wind disturbances, load swaying, etc., thereby reducing the impact of these disturbances on the flight stability of the UAV. This invention achieves high-precision tracking of a predetermined trajectory by integrating TVUDE into the trajectory tracking controller. TVUDE helps to correct trajectory deviations caused by external disturbances in real time, ensuring that the UAV can fly stably and accurately along the predetermined trajectory. Traditional disturbance estimators have a trade-off between transient response and steady-state performance. This invention effectively solves this trade-off by introducing time-varying parameters and optimizing the disturbance estimation bandwidth. The introduction of time-varying parameters allows TVUDE to achieve a better trade-off between transient response and steady-state performance, thereby improving overall control performance.

[0079] 2. The present invention has the following advantages: By introducing time-varying parameters, the present invention can adapt to the diversity of interference and effectively improve the anti-interference performance of the system; the dual-loop TVUDE control framework of the present invention has significant advantages over the single estimator framework in simultaneously improving the performance of anti-sway and trajectory tracking; the present invention sets out corresponding control schemes and controller parameters for the anti-sway and trajectory tracking scenarios of quadcopter load hoisting.

[0080] 3. This invention employs payload suspension technology, which connects the load to the quadcopter's center of gravity via cables. Compared to a rigid connection with a robotic arm, this payload suspension technology effectively preserves the quadcopter's inherent agility because it does not directly introduce inertia. To ensure safe and efficient completion of rescue missions, this invention considers the anti-interference issue of the quadcopter load lifting system. Attached Figure Description

[0081] Figure 1 This is a coordinate system model diagram of the quadrotor transport system targeted by the quadrotor sling transport control method based on dual time-varying disturbance estimators of the present invention;

[0082] Figure 2 This is a control framework diagram of the present invention;

[0083] Figure 3 This is a top view of the simulated flight results of the present invention;

[0084] Figure 4 The figure shows the simulation results of the anti-sway control of the present invention;

[0085] Figure 5 The figure shows the simulation results of the trajectory tracking control of the present invention. Detailed Implementation

[0086] The present invention will be further described below with reference to the accompanying drawings and specific embodiments:

[0087] like Figures 1 to 5 As shown, the present invention provides a quadcopter sling transport control method based on a dual time-varying disturbance estimator, comprising the following steps:

[0088] S1. Establish a dynamic model of the quadcopter UAV and its suspended load, including the transformation relationships between the world coordinate system, the body coordinate system, the suspension coordinate system, and the load coordinate system.

[0089] To achieve robust control of the quadcopter's position loop, it is first necessary to establish a position loop model of the quadcopter UAV and a disturbance model for load oscillation. The quadcopter lifting system addressed in this invention is as follows: Figure 1 As shown, the rotation matrix of the quadcopter is given by the following formula:

[0090]

[0091] Where C(·) and S(·) represent the trigonometric functions sin(·) and cos(·) respectively; θ, φ, ψ represent the Euler angles of the quadcopter in the inertial frame; B refers to the machine system, W refers to the world system, P refers to the load system, and S refers to the suspension system. It is the rotation matrix from the slave system to the world system; other rotation matrices are defined in the same way.

[0092] To achieve anti-sway control of the cargo, it is first necessary to model the cargo's swaying. The dynamic model of the suspended load in S1 includes a model of the cargo's swaying, which is represented as follows:

[0093]

[0094] Where λ x ,λ y The two swing angles of the cargo are represented by L; L is the length of the rope used to hoist the cargo; m p It refers to the quality of the goods; This represents the acceleration vectors along the three axes of a quadcopter drone; the symbol "T" in the upper right corner represents the transpose of the vector or matrix; F pd =[F pdx ,F pdy ,F pdz ] T is the air resistance vector acting on the cargo; g is the acceleration due to gravity.

[0095] Based on the cargo swaying model, a comprehensive quadrotor load model is established. The quadrotor linear motion dynamics formula in the world coordinate system is given by the following equation:

[0096]

[0097] Where m q It is the mass of the quadcopter; F qt =[0,0,F qt ] T It is the quadrotor thrust command vector in the body coordinate system; F is the cable tension vector on the quadcopter in the UAV's world coordinate system; qd This is the disturbance signal vector of the quadcopter position loop. The disturbance signals include model uncertainty, wind interference, and load oscillation interference. The derivatives of the disturbances are all bounded, and g = [0, 0, g]. T It is gravitational acceleration.

[0098] S2. The model is simplified using feedback linearization techniques, decoupling the nonlinearly coupled system into two subsystems: anti-sway and trajectory tracking.

[0099] After obtaining the above models, the quadrotor sway angle model and the quadrotor kinematic model are nonlinear and strongly coupled, making them difficult to control and design. Therefore, feedback linearization technology was first applied to simplify the sway angle model and the kinematic model.

[0100] First, the pendulum angle model (2) is simplified. In S2, the model is simplified using feedback linearization. The simplified pendulum angle model is as follows:

[0101]

[0102] Where λ=[λ x ,λ y ] T It's the tilt angle of the quadcopter, u λ =[u λx ,u λy ] T It is a virtual control input, and has:

[0103]

[0104] f λ =[f λx ,f λy ] T It is a lumped interference, and has:

[0105]

[0106] To prevent the quadcopter from producing aggressive vertical maneuvers that could lead to collisions, the system input is... To sum up to the interference term f λ In this embodiment, model (2) refers to the model corresponding to formula (2).

[0107] The quadrotor load integrated model (3) refers to the model corresponding to formula (3). After using feedback linearization technology, the quadrotor load integrated model (3) is simplified as follows:

[0108]

[0109] Where u p It is virtual input:

[0110]

[0111] f p It is lumped interference:

[0112]

[0113] This completes the simplification of the model, facilitating subsequent control design. Model (3) refers to the model corresponding to formula (3).

[0114] S3. Based on the swaying model of the hoisted goods, a time-varying uncertainty and disturbance estimator (TVUDE) is constructed in the sway elimination controller to dynamically aggregate time-varying disturbances to eliminate the swaying of the heavy objects.

[0115] The overall control framework of this invention is as follows: Figure 2As shown, to ensure rapid convergence of the swing angle during S3 and to actively resist interference, this interference consists of model uncertainty and external disturbance. An outer loop and an inner loop are set up. The outer loop is the control loop and the inner loop is the attitude loop. The first component of the outer loop is designed: the swing cancellation controller, which rewrites the simplified swing angle model (4) into a state-space expression:

[0116]

[0117] Where x1 and x2 represent the sway angle and angular velocity of the quadcopter, respectively, and the sway angle model (4) refers to the model corresponding to formula (4). It is assumed that the required state is generated by the following reference model:

[0118]

[0119] The robust control algorithm based on uncertainty and disturbance estimator designed for the model of formula (10) is expressed as follows:

[0120]

[0121] u λ0 It is a nominal controller, designed as follows:

[0122]

[0123] Where u λd It is a feedforward term, which can be considered zero in this robust stabilization problem, kp = diag{kpx,kpy} and k d =diag{k dx ,k dy} is the feedback gain. It is the quadcopter tilt angle tracking error. It's an angular velocity tracking error, and in this problem, the swing angle x... 1d and x 2d The expected value can also be considered as zero; substituting formula (13) into formula (12), we get:

[0124]

[0125] The interference estimation signal generated by TVUDE is designed in the time domain as follows:

[0126]

[0127] Where T λ (t)=diag{T λx (t),T λy (t)} is a time-varying parameter matrix designed by humans. Combining formula (4) and formula (12), we get In formula (15) use Replace and multiply the matrix on the left side of both sides of formula (15). The derivative of the interference estimation signal is obtained:

[0128]

[0129] Integrating both sides of equation (16) simultaneously and applying integral by parts, we obtain the disturbance estimation signal for the anti-sway controller:

[0130]

[0131]

[0132] After completing the anti-sway control design, the equation can be solved inversely from formula (5). and Setting the anti-oscillation acceleration on the Z-axis to zero will give us anti-oscillation acceleration commands for the three axes:

[0133]

[0134] S4. Based on the quadrotor kinematic model, a time-varying uncertainty and disturbance estimator (TVUDE) is constructed in the trajectory tracking controller to achieve accurate trajectory tracking of the quadrotor UAV.

[0135] To balance the conflict between pendulum elimination and trajectory tracking control, the trajectory tracking controller of TVUDE in S4 is used as the second component of the outer loop. The simplified trajectory tracking model is represented by formula (7), and its state-space expression is:

[0136]

[0137] in and These represent the position and velocity of the quadcopter, respectively. Next, we will design the quadcopter trajectory tracking controller:

[0138]

[0139] Where is u pd Feedforward term, k pp =diag{k ppx ,k ppy} and k pd =diag{k pdx ,k pdy} represents the feedback gain, e p =x 3d -x3 is defined as the trajectory tracking error, x 3d This is the desired position signal for the quadcopter. The trajectory tracking TVUDE is designed as follows:

[0140]

[0141] Similarly, multiply matrix T on both sides of equation (21) simultaneously. p The inverse of (t), and will and x4=u p +f p Substituting the values:

[0142]

[0143] Integrating both sides of equation (22) simultaneously and applying integral by parts, we obtain the disturbance estimation signal for the trajectory tracking controller:

[0144]

[0145] At this point, the design of the anti-sway and trajectory tracking controllers is complete. The outer-loop control can be represented as the sum of these two controllers:

[0146]

[0147] S5. The stability of the estimation error system is proved using the theory of linear time-varying systems, and the performance of disturbance estimation is optimized by adjusting auxiliary parameters.

[0148] After completing the control design, the stability of the estimation error system is proved using linear time-varying system theory. Furthermore, when optimizing the disturbance estimation performance by adjusting auxiliary parameters, the anti-sway disturbance estimator is taken into account. And trajectory tracking interference estimator The system maintains formal consistency, and in the proof, it is only necessary to prove that one of the estimation error systems is stable. Therefore, all subscripts in the system will be removed in the subsequent proof. The estimation error is defined as:

[0149]

[0150] Integrating both sides of equation (25) with respect to time t, and then substituting it into equation (15), we obtain the estimation error subsystem:

[0151]

[0152] According to the theory of linear systems, the state transition matrix of the linear time-varying system (26) is:

[0153]

[0154] The solution to system (26) is:

[0155]

[0156] Time-varying parameter T = diag{T x (t),T y (t),T z (t)} is a bounded positive number that is artificially defined. For the x-axis, we define 0 < T. xmin ≤T x (t)≤T xmax And there are From formula (28), we know that the solution of the system consists of two components: the zero-input response and the zero-state response. As t→∞, the zero-input response... Then, scaling the zero-state response yields:

[0157]

[0158] in As t→∞, The zero-state response can be simplified to:

[0159]

[0160] It can be proven using the same method. and It is also convergent; as t→∞, we have:

[0161]

[0162] in T m =diag{T xmax ,T ymax ,T zmax};make The advantage of doing this is that the time-varying parameter matrix T can be changed by adjusting the magnitude of the scalar ζ. m The magnitudes of each component, thus proving that when the external disturbance f is bounded, the output... It is also bounded, which proves that system (26) is stable.

[0163] S6. Prove the stability of the anti-swing subsystem and the trajectory tracking subsystem, and optimize the disturbance estimation performance by adjusting the auxiliary parameters.

[0164] In S6, when proving the stability of the anti-swing subsystem and the trajectory tracking subsystem and optimizing the disturbance estimation performance by adjusting the auxiliary parameters, it was also observed that the anti-swing subsystem (10) and the trajectory tracking subsystem (19) have the same state-space expression, so only one of them is proven; the swing angle error and angular velocity error are defined as:

[0165]

[0166] Combining formulas (10), (11), (12), (13), and (25), we obtain the oscillation error system:

[0167]

[0168] To minimize the swing angle error and angular velocity error For a system to converge to boundedness, system (33) must satisfy:

[0169] (1) Matrix A is the Herwitz matrix, i.e., gain k p >0,k d >0;

[0170] (2) Estimation error Bounded convergence; the solution for the oscillation error system is:

[0171]

[0172] As t→∞, the zero-input response e At →0, Substituting formula (31) into the zero-state response, we get:

[0173]

[0174] Obviously, as t→∞, (||I||2-||e At ||2)→||I||2, and when ζ→0, ε(ζ)→0. Therefore, it is proved that the oscillation error system satisfies: Thus, it can be proven that the anti-slip system and the trajectory tracking system are exponentially stable, and the performance of perturbation estimation can be improved by reducing the auxiliary parameter ζ.

[0175] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A quadcopter sling load transport control method based on dual time-varying jammer, characterized in that, The method comprises the following steps: S1, a dynamic model of a quadrotor unmanned aerial vehicle and a suspended load is established, including conversion relationships of a world coordinate system, a body coordinate system, a suspension coordinate system and a load coordinate system; S2, a feedback linearization technique is used to simplify the model, and the nonlinear coupled system is decoupled into two subsystems of anti-swing and trajectory tracking; S3, according to a swing model of the suspended load, a time-varying uncertainty and disturbance estimator is constructed in the anti-swing controller, and time-varying disturbances are dynamically aggregated to eliminate the swing of the heavy object; S4, according to a kinematics model of the quadrotor, a time-varying uncertainty and disturbance estimator is constructed in the trajectory tracking controller, and accurate trajectory tracking of the quadrotor unmanned aerial vehicle is realized; S5, the stability of the estimation error system is proved by using linear time-varying system theory, and the disturbance estimation performance is optimized by adjusting auxiliary parameters; S6, the stability of the anti-swing subsystem and the trajectory tracking subsystem is proved, and the disturbance estimation performance is optimized by adjusting auxiliary parameters; The model is simplified by using the feedback linearization technique in the step S2, and the simplified swing angle model is: where λ = [λ x , λ y ] T is the pitch angle of the quadrotor, u λ = [u λx , u λy ] T is the virtual control input, and has: f λ = [f λx ,f λy ] T is a lumped interference, and has: To make the quadrotor not to produce the aggressive vertical maneuver which leads to collision, the system input is thus reduced to the disturbance term f λ ; In the S3, the swing angle converges quickly and actively resists disturbances, which are uncertainties and external disturbances of the model, and an outer loop and an inner loop are provided, the outer loop is a control loop, and the inner loop is an attitude loop, a first component of the outer loop, the anti-swing controller, is designed, and the simplified swing angle model (4) is rewritten as a state space expression: Wherein, x1 and x2 represent the swing angle and the swing angle velocity of the quadrotor respectively, and it is assumed that the required state is generated by the following reference model: The robust control algorithm based on the uncertainty and disturbance estimator designed for the model of formula (10) is represented as: u λ0 is a nominal controller designed to: where u λd is the feedforward term, which is considered to be zero in this robust stabilization problem, k p = diag{k px , k py} and k d = diag{k dx , k dy} are feedback gains, is the roll angle tracking error, is the angular velocity tracking error, and in this problem, the desired values of the roll angle x 1d and x 2d are considered to be zero; substituting equation (13) into equation (12) gives: (14) is the interference estimation signal generated by TVUDE, designed in time domain as: where T λ (t) = diag{T λx (t), T λy (t)} is a man-made time-varying parameter matrix, combining equation (4) and equation (12), we get 2. The quadrotor sling load transport control method based on dual time-varying interference estimator according to claim 1, wherein, The rotation matrix of the quadrotor unmanned aerial vehicle in the S1 is given by the following formula: where the symbols C(·) and S(·) represent the sin(·) and cos(·) trigonometric functions, respectively; θ, φ, ψ represent the Euler angles of the quadrotor in the inertial frame; B refers to the body frame, W refers to the world frame, P refers to the payload frame, and S refers to the sling frame; is the rotation matrix from the body frame to the world frame; Other rotation matrices are defined in the same way.

3. The quadrotor sling load transport control method based on dual time-varying jammer estimator of claim 1, wherein, The dynamic model of the suspended load in the S1 includes a model of the cargo swing, and the model of the cargo swing is represented as: where λ x ,λ y represent the two swing angles of the cargo; L is the length of the rope hoisting the cargo; m p is the mass of the cargo; represent the acceleration vectors in the three axes of the quadcopter, the symbol "T" in the upper right corner represents the transpose of the vector or matrix; F pd =[F pdx ,F pdy ,F pdz ] T is the air resistance vector suffered by the cargo; g is the acceleration of gravity.

4. The quadrotor sling load transport control method based on dual time-varying jammer estimator of claim 3, wherein, Based on the model of the cargo swing, a quadrotor load comprehensive model is established, and the linear motion dynamics formula of the quadrotor in the world coordinate system is given by the following formula: where m q is the mass of the quadrotor; F qt = [0, 0, F qt ] T is the thrust command vector of the quadrotor in the body frame; is the tether tension vector on the quadrotor in the world frame of the UAV; F qd is the disturbance signal vector of the quadrotor position loop, which includes model uncertainty, wind disturbance, and load swing disturbance, the derivatives of which are all bounded, g = [0, 0, g T is the gravitational acceleration.

5. The quadrotor sling load transport control method based on dual time-varying jammer estimator of claim 4, wherein, After using the feedback linearization technique, the quadrotor load comprehensive model obtains a simplified model: where u p is a dummy input: f p is the lumped interference: At this point, the model is simplified to facilitate subsequent control design.

6. The quadrotor sling load transport control method based on dual time-varying jammer estimator of claim 1, wherein, The trajectory tracking controller of the TVUDE in the S4 is the second component of the outer loop, and the simplified trajectory tracking model is represented by formula (7), and the state space expression thereof is: where and denote the position and velocity of the quadrotor, respectively.

7. The quadcopter sling load transport control method based on dual time-varying jammer according to claim 1, characterized in that, The stability of the estimation error system is proved by using the linear time-varying system theory in S5, and the disturbance estimation performance is optimized by adjusting the auxiliary parameters when considering the swing suppression disturbance estimator and the trajectory tracking disturbance estimator There is consistency in form, and only one of the estimation error systems needs to be proved to be stable when proving, so all the subscripts in the system are removed in the subsequent proof; the estimation error is defined as: Integrate formula (25) on both sides with respect to time t, and then substitute it into formula (15) to obtain an estimation error subsystem: According to the linear system theory, it is known that the state transition matrix of the linear time-varying system (26) is: The solution of the linear time-varying system (26) is: time-varying parameter T = diag{T x (t), T y (t), T z (t)} are positive numbers set artificially and bounded, for the x-axis, set 0 < T xmin ≤ T x (t) ≤ T xmax , and have It is known from equation (28) that the solution of the system is composed of two components, zero input response and zero state response. At t→∞, the zero input response Then scale the zero state response to get: wherein At t→∞, The zero state response is simplified to: In the same way we prove that and is also convergent; at t→∞ we have: where Let The advantage of doing this is that by adjusting the size of the scalar ζ m the size of each component, it is shown that the output is also bounded, and hence that the linear time-varying system (26) is stable. 8.The quadrotor sling load transport control method based on dual time-varying jammer according to claim 1, wherein: In the S6, when the stability of the anti-swing subsystem and the trajectory tracking subsystem is proved, and the disturbance estimation performance is optimized by adjusting auxiliary parameters, it is also observed that the anti-swing subsystem (10) and the trajectory tracking subsystem (19) have the same state space expression, so only one of them is proved; Define the swing angle error and the angular velocity error as: The swing error system is obtained by combining formula (10), formula (11), formula (12), formula (13) and formula (25): To make the swing angle error and the angular velocity error bounded convergent, the swing error system (33) needs to satisfy: The matrix A is a Hurwitz matrix, i.e. the gain k p > 0, k d > 0; estimation error bounded convergence; the solution of the swing error system is: When t→∞, the zero-input response e At →0, substituting equation (31) into the zero-state response gives: Obviously, when t→∞, (||I||2-||e At ||2)→‖I||2, and when ζ→0, ε(ζ)→0, thus, it is proved that the swing error system satisfies: So far, it is proved that the swing elimination system and the trajectory tracking system are exponentially stable, and the performance of disturbance estimation is improved by reducing the auxiliary parameter ζ.

Images